How good was 'The Four Horsemen' deck before it got axed?

[nevilshute]

I wasn't playing much magic at the time the Four Horsemen deck was a thing. I know it got "banned" and I know why, but can anyone tell me whether or not it was actually very good or was it closer to mediocre when compared to other legacy combo decks?

[Julian23]

It's never been aaaaaaaaanywhere close to Tier1. Probably not even Tier2 but "not Tier2" is a pretty bold statement in Legacy.

[death]

If there was a Worst legacy decks poll, it'll be in the top ranks.

[PirateKing]

But it was silly, thus it's endearing history. See also: Nourishing Lich.

[Serbitar]

Can someone expand on in what sense it was 'banned'? Did judges come to agree that it's kill is not progressing the game state, or what?

[Julian23]

It was less "banned" and more just clarified that it constitutes behavior that will not be accepted. The fun part is that it still Top16'ed a SCG.

[DragoFireheart]

The deck is so amazing that it got banned without resulting in a single card being banned.

[JPoJohnson]

Can someone expand on in what sense it was 'banned'? Did judges come to agree that it's kill is not progressing the game state, or what?

It's kill doesn't always "not" progress the game state, but often enough it wouldn't and yes that is what caused it to be not allowed in tournament play. It typically takes far too long to do what it needs to do (and have to repeat it's kill method often enough before it works sometimes that it required a lot of time).

As Julian said in another thread (I believe it was him) Since it has a 99.9% chance of eventually doing the combo, it's not 100% and that constituted not advancing the game state.

[PirateKing]

Can someone expand on in what sense it was 'banned'? Did judges come to agree that it's kill is not progressing the game state, or what?

It was banned in the sense that the combo could meet the condition of slow play in the IPG. If your opponent is aware of this, they call a judge and you get warned for slow play. Eventually, enough players, enough warnings and you get game loss, on to DQ.

THEORETICALLY, the deck could meet it's kill condition every time without fizzling ever, in which case nobody would have merit to call a judge.

[JPoJohnson]

It was banned in the sense that the combo could meet the condition of slow play in the IPG. If your opponent is aware of this, they call a judge and you get warned for slow play. Eventually, enough players, enough warnings and you get game loss, on to DQ.

THEORETICALLY, the deck could meet it's kill condition every time without fizzling ever, in which case nobody would have merit to call a judge.

The guy that top 16 a SCG event only had the judge called on him once. Either he had patient opponents or he got lucky!

[Koby]

The guy that top 16 a SCG event only had the judge called on him once. Either he had patient opponents or he got lucky!

Or likely, his opponent did not understand that this was demonstrably not possible to kill someone without Slow Play; and prematurely scooped.

Mathematically, it's a feasible combo, as N approaches large numbers. However, as N approaches large numbers, the minutes wasted iterating the setup pushes into "do nothing, repeat" and grinds time off the clock. Once the setup is complete, the combo is executed with a simple loop that takes 20 seconds to explain and resolve.

[rufus]

The guy that top 16 a SCG event only had the judge called on him once. Either he had patient opponents or he got lucky!

By itself, the four horsemen combo is not actually that slow:

You're guaranteed to be able to pull all Narcomoebas out of the deck in one trip through. (1 trip through the deck, advances the game state)
Then you need to get Sharuum the Hegemon, Blasting Station and Dread Return all in the graveyard at the same time without an Emrakul trigger. (Assuming 1 of each, and 1 Emrakul, this takes 4 trips through the deck on average.)
From there, you do 3 damage per trip through the library, assuming 3 Narcomoebas. (~7 trips through the deck, pseudo-deterministic)

Though it has the potential to get stupid if it's trying to play around graveyard hate cards, or if there are multiple Emrakuls in the deck.

[Koby]

Then you need to get Sharuum the Hegemon, Blasting Station and Dread Return all in the graveyard at the same time without an Emrakul trigger. (Assuming 1 of each, and 1 Emrakul, this takes 4 trips through the deck on average.)

This is incorrect. It is mathematically indeterminate, because you shuffle the deck to a random state. Each time you shuffle, the probability event is reset.

Assuming you have a 50 card deck (say, turn 3):
Each "mill" has a 2% chance of being an Emrakul, and increases as you miss it. On top of that, you are adding conditional requirements for 3 other separate cards, which reduces the likelihood of the exit event from occurring.

I'm not an expert on calculating probabilities, but this number is much smaller than 25%.

[Julian23]

The only time I ever faced Four Horsemen, my opponent timed out in game1 on Magic Online :eek::cry:....:laugh:

[Tammit67]

I'm not an expert on calculating probabilities, but this number is much smaller than 25%.

If the rest of the deck doesn't matter, you can simply consider it a 4 card deck really. Consider the relative order of the important pieces.

There are 4! = 24 different ways the Sharuum (S), Dread return (D), Blasting station (B), and emrakyl(E) can be ordered relative to each other. Out of those possibilities, here are the ones with Emrakyl last:

SDBE
SBDE
BSDE
BDSE
DSBE
DBSE

For a total of 6. 6/24 = .25 -> there is a 25% chance of that individual shuffle having those 4 cards in the order you want. If I'm making a mistake, please correct me.

With a 25% chance of having the deck in the ascribed order, the expected number of shuffles you'd have to perform before reaching the relative order is only 4. However, distribution on expected number is heavily tailed and indeed there exists outcomes where you do not reach the conclusion in X shuffles, where X is whatever whole number you please.

[rufus]

This is incorrect
...
I'm not an expert on calculating probabilities, but this number is much smaller than 25%.

There are a bunch of cards in the library, but in that step we only care about four of them:
Sharuum
Dread Return
Blasting Station
Emrakul

Within those four cards, we can go off when Emrakul is the bottom card.
Since Emrakul is equally likely to be each position, there's a 25% chance.

[Koby]

I stand corrected on the probability.

However, this is a theoretical term. Succeeding in the event may happen in as few as 1, or as many as 30 events. After the second attempt failing, a table judge is within his authority to declare Slow Play. That is the ultimate nail in the coffin for this deck.

[TheInfamousBearAssassin]

There is a .75 chance that Emrakul will not be the fourth of the four cards to hit your yard in each iteration, and the deck is reset in each case, so the combo will be set up somewhere between the second and third shuffle on average.


The problem is that Magic does not allow infinite numbers, and because there is a random element to each repetition of the main action in the combo it is necessary to perform it by rote. This was ruled to constitute not advancing the game state, probably just because of judge-ire.

If you could declare an infinite repetition there would always be a 100% chance of any possible outcome occurring, so performing the combination each time would not be necessary, but again, mechanically Magic doesn't recognize infinite numbers; so you can say, "I'll target Daru Spiritualist a number of times equal to Graham's Number," (http://en.wikipedia.org/wiki/Graham%27s_number) but you can't, say, "I'll untap Basalt Monolith a number of times equal to Zilla's mom's mass in kilograms."

[TheInfamousBearAssassin]

If the rest of the deck doesn't matter, you can simply consider it a 4 card deck really. Consider the relative order of the important pieces.

There are 4! = 24 different ways the Sharuum (S), Dread return (D), Blasting station (B), and emrakyl(E) can be ordered relative to each other. Out of those possibilities, here are the ones with Emrakyl last:

SDBE
SBDE
BSDE
BDSE
DSBE
DBSE

For a total of 6. 6/24 = .25 -> there is a 25% chance of that individual shuffle having those 4 cards in the order you want. If I'm making a mistake, please correct me.

With a 25% chance of having the deck in the ascribed order, the expected number of shuffles you'd have to perform before reaching the relative order is only 4. However, distribution on expected number is heavily tailed and indeed there exists outcomes where you do not reach the conclusion in X shuffles, where X is whatever whole number you please.

The probability of Emrakul not being the last card in that order is .75. We have replacement so the probability of Emrakul not being the last card twice in a row is .75^2 which is .5625. The probability of Emrakul not being the last card three times in a row is .75^3 which is .42188, so you are slightly unlikely to achieve the correct order with two tries and slightly favored to with three.

[rufus]

The probability of Emrakul not being the last card in that order is .75. We have replacement so the probability of Emrakul not being the last card twice in a row is .75^2 which is .5625. The probability of Emrakul not being the last card three times in a row is .75^3 which is .42188, so you are slightly unlikely to achieve the correct order with two tries and slightly favored to with three.

Technically, there's a 1 in 7 chance to get all 6 pieces before Emrakul on the first trip, which pushes the average total shuffle count a bit lower.


THEORETICALLY, the deck could meet it's kill condition every time without fizzling ever, in which case nobody would have merit to call a judge.

I'm too lazy to do the math or work out the optimizations, but assuming that every casting of Cabal Therapy and every Narcomoeba that comes into play 'advances the game state', a version with 4x Cabal Therapy and 4x Narcomoeba should have decent chance of going off.

[MaximumC]

This deck, and the ruling that makes it inoperable, is a big pet peeve of mine. As it stands, you cannot shortcut -- or even manually perform any loop where you cannot specify, in advance, exactly how many times you will be repeating the loop and exactly what the resulting board state will be. In effect, this means if you have a loop involving any element of randomness whatsoever (coin flips or shuffling) then, according to the rules, you're done.

The IPG defines "Slow Play" infraction, in part as follows:

"It is also slow play if a player continues to execute a loop without being able to provide an exact number of iterations and the expected resulting game state." IPG 3.3.

This means that, for Four Horseman, unless your library is perfectly stacked the first time through the loop, you can't "continue" to do it a second time. There are logistical reasons that a tournament organize might want this kind of ruling for this kind of deck.

This rule annoys the bejeesus out of me, even so. Consider what else it bars. Let's say you've got yourself a Mijae Djinn in play enchanted with Bear Umbra, Aggravated Assault. and five untapped lands. Assume no blockers or instants from opponent.

Start by paying for Aggravated Assault. You can attack with the Djinn, maybe dealing damage to your opponent and maybe not. It's random. Either way, your lands all untapped on the attack. You can now spend mana for Aggravated Assault again. This is a loop that you can continue to execute until your opponent is dead.

Except that the Rules prohibit this. Huh? Well, you're trying to execute a loop, because you are arriving back at an identical game state if Djinn decides not to damage the opponent. You cannot "continue" to execute this loop because you cannot identify precisely how many iterations will be necessary to actually finish the game.

I would really prefer that the current IPG on loops and slow play be replaced with something like this:

"It is also slow play if a player continues to execute a loop without being able to proof that, for any given probability N less than 100%, there is a number of iterations that will increase the probability of an expected resulting game state above N." IPG 3.3.

This might be too difficult for judges and players to calculate, but it would allow infinite loops that WILL win, just not necessarily in a short period of time, to function. Would revive Four Horseman.

[Koby]

Let's say you've got yourself a Mijae Djinn in play enchanted with Bear Umbra, Aggravated Assault. and five untapped lands. Assume no blockers or instants from opponent.

This is covered under Tournament Shortcuts. You are correct in that you cannot loop this sequence, but you can easily demonstrate a shortcut, and resolve all the coinflips.

The chief complaint about the Four Horsemen combo is that it requires shuffling, then revealing cards at random. This eats up far too much time during a tournament.

[rufus]

This is covered under Tournament Shortcuts. You are correct in that you cannot loop this sequence, but you can easily demonstrate a shortcut, and resolve all the coinflips...

Since you don't know what your opponent's life will be at the end of the loop, you can't specify the end state before flipping the coins.

Edit:


716.2a At any point in the game, the player with priority may suggest a shortcut by describing a
sequence of game choices, for all players, that may be legally taken based on the current game
state and the predictable results of the sequence of choices. ...


Since the loop is game-ending, your opponent would be well-served to insist on it getting played out to force the penalty.

[Zombie]

Since you don't know what your opponent's life will be at the end of the loop, you can't specify the end state before flipping the coins.

Which is why you have to flip the coins. But the state between the coin flips is deterministic, and deterministic sequences can be shortcut. So you flip a coin, say shortcut, flip another coin and so on.

[MaximumC]

This is covered under Tournament Shortcuts. You are correct in that you cannot loop this sequence, but you can easily demonstrate a shortcut, and resolve all the coinflips.

The chief complaint about the Four Horsemen combo is that it requires shuffling, then revealing cards at random. This eats up far too much time during a tournament.

That's not what the loop rule says, though. It has nothing to do with how long the random event takes to evaluate. It only has to do with indeterminacy. If you can shortcut your attacks with Mijae Djinn, why can't you shortcut your win with Four Horsemen?

The Comprehensive Rules say:

"3.1b Occasionally the game gets into a state in which a set of actions could be repeated
indefinitely (thus creating a “loop”). In that case, the shortcut rules can be used to determine
how many times those actions are repeated without having to actually perform them, and how
the loop is broken.

716.2. Taking a shortcut follows the following procedure.

716.2a At any point in the game, the player with priority may suggest a shortcut by describing a sequence of game choices, for all players, that may be legally taken based on the current game state and the predictable results of the sequence of choices. This sequence may be a non-repetitive series of choices, a loop that repeats a specified number of times, multiple loops, or nested loops, and may even cross multiple turns. It can’t include conditional actions, where the outcome of a game event determines the next action a player takes. The ending point of this sequence must be a place where a player has priority, though it need not be the player proposing the shortcut.

716.2b Each other player, in turn order starting after the player who suggested the shortcut, may either accept the proposed sequence, or shorten it by naming a place where he or she will make a game choice that’s different than what’s been proposed. (The player doesn’t need to specify at this time what the new choice will be.) This place becomes the new ending point of the proposed sequence.

716.2c Once the last player has either accepted or shortened the shortcut proposal, the shortcut is taken. The game advances to the last proposed ending point, with all game choices contained in the shortcut proposal having been taken. If the shortcut was shortened from the original proposal, the player who now has priority must make a different game choice than what was originally proposed for that player."

Whew!

So, applying that to the Four Horseman:

Player A: Let me show you the combo. See, I can keep doing this infinitely. Eventually Emrakul will be at the bottom.
Player B: Okay. You can't tell how many times you have to do that in order to reach that result, though.
Player A: Let me propose a shortcut by describing a sequence of game choices, for all players, that may be legally taken based on the current game state and the predictable results of the sequence of choices. I will keep milling myself and shuffling my library until I Emrakul is the last card in my library and the rest are in a random order on top of it.
Player B: Alright, I have no responses until we reach that game state.

This should work just fine under the shortcut rules, right? It is the same as the Mijae example, which you had no trouble with, and would go down like this:

Player A: Let me show you the combo. See, I can keep doing this infinitely. Eventually, Mijae Djinn will deal 20 damage to you.
Player B: Okay. You can't tell how many times you have to do that in order to reach that result, though.
Player A: Let me propose a shortcut by describing a sequence of game choices, for all players, that may be legally taken based on the current game state and the predictable results of the sequence of choices. I will keep attacking you until you are dead.
Player B: Alright, I have no responses until we reach that game state.

See what I'm saying?

Perhaps you are saying that Mijae could not be looped this way, but instead could be looped by specifying that you will attack X times, and then just flipping a coin X times to ensure determine damage. In other words, make the number of iterations a constant and not conditional. I don't see how the Rules treat that any differently from the above example. Indeed, your shortcut may turn out to be invalid; if Mijae hits the first four flips, then those additional flips you shortcutted will never occur because you would never have obtained those additional combat steps!

Even worse, doing it that way leads to the same problem you get with Frenetic Efreet and Chance Encounter. Efreet allows you to stack activations, so I think there's no question that you can loop that effect X times if you want. (Unlike the Mijae example, you don't start flipping until all your stacked activations resolve, so there's never a paradox where the flip could never have occurred). But, as the Efreet player, I realize there is a nonzero chance I will whiff even with many flips. If I want to reduce my chances of losing down below an arbitrary probability, I need to flip LOTS of coins. So, to be safe, I specify 1,000,000 flips. We then end up bogging down the whole tournament and getting judges called over while we resolve the flips. Same deal with Mijae; if we really want to maximize the win, we need to specify we're attacking 1,000,000 times.

As a practical matter, your opponent would PROBABLY let you get away with stopping the flips once you reach the magic number. In the case of Djinn, you'd have to, since the rest of the flips "never occurred." But it's not a given than (a) you will nessarily hit the required number of flips quickly; or (b) your opponent will not insist on you finishing the flips.


Since you don't know what your opponent's life will be at the end of the loop, you can't specify the end state before flipping the coins.

Edit:


Since the loop is game-ending, your opponent would be well-served to insist on it getting played out to force the penalty.

I missed your post. This is, I think, precisely the problem. The rules do not let you shortcut "until something is true." They let you shortcut "X times" and ONLY if you can specifically identify the outcome when you're done -- even if the loop is mathematically certain to produce a certain result in infinite time.

Which is complete bunk and should be changed.

[rufus]

Which is why you have to flip the coins. But the state between the coin flips is deterministic, and deterministic sequences can be shortcut. So you flip a coin, say shortcut, flip another coin and so on.

"That word. I do not think it means what you think it means."
-- Fezzig, The Princess Bride

The opponet's life total depends on the result of the coin flip. Ergo the state is not deterministic.

[MaximumC]

"That word. I do not think it means what you think it means."
-- Fezzig, The Princess Bride

The opponet's life total depends on the result of the coin flip. Ergo the state is not deterministic.

Right, but see my post above: I think he might be suggesting that the player propose a Djinn loop by saying:

"I propose to repeat this 1,000,000 times. Let's now flip a coin that many times."

as opposed to

"I propose to repeat this until you are dead."

I think both are problematic for different reasons. (But the latter should be a perfectly acceptable thing to do in Magic)

[Koby]

Right, but see my post above: I think he might be suggesting that the player propose a Djinn loop by saying:

"I propose to repeat this 1,000,000 times. Let's now flip a coin that many times."

as opposed to

"I propose to repeat this until you are dead."

I think both are problematic for different reasons. (But the latter should be a perfectly acceptable thing to do in Magic)

I specifically made sure to not call it a loop. It cannot be looped.

The following action is an acceptable shortcut:
"Attack with Mijae Djinn enchanted with Boar Armor, trigger Boar Armor, untap all my lands, activate Aggrivated Assault."
And trail Mijae's trigger on the end, which will need to resolve independent of the shortcut.

Here, you're advancing the game state because you are declaring attackers and moving phases.

You are correct in that you cannot loop this because it is a dependent action (coin-flip). Some judges may accept this as a loop provided you give a realistic number, say 2^N-1, where N is <opp life total> MOD 6; but that's asking a lot from your judge and for them to understand math. Short answer is: not an acceptable loop.

[rufus]

I specifically made sure to not call it a loop. It cannot be looped.

The following action is an acceptable shortcut:
"Attack with Mijae Djinn enchanted with Boar Armor, trigger Boar Armor, untap all my lands, activate Aggrivated Assault."
And trail Mijae's trigger on the end, which will need to resolve independent of the shortcut.

Here, you're advancing the game state because you are declaring attackers and moving phases.

....

If you're flipping coins one at a time, you've got a repeated state as soon as one of the flips goes against you. AFAICT your opponent could then demand you do something different to break the cycle.

[Dzra]

Back to the original question... Four Horsemen is and will always remain a horrible deck. It is basically a worse version of Breakfast, which has probably been outclassed by the new Oops All Spells deck (which isn't even that good either for that matter).

As to the rules question, no you cannot shortcut the loop because you cannot say when you will get the result you want. Practically speaking, you could literally sit there for 50 minutes and never actually get the combo and both players would receive a draw. Alternatively, the Four Horsemen player could be up game 1, and loop game 2 for the remainder of the round. Neither of these things promote a good tournament environment.

[Julian23]

"I propose to repeat this until you are dead."

I think both are problematic for different reasons. (But the latter should be a perfectly acceptable thing to do in Magic)

I disagree. I see that your "should"-statement is a subjective evaluation of what you would like to see though.

What you (and others) fail to adress or even see is why allowing 99% to be 100% is very problematic: where do you draw the line? If a player is playing a combo that doesn't always win, strictly speaking, we can not just hand him that last few %s for free. It doesn't matter whether we are talking about 0,00..1% or any other number as that would be an arbitrary choice.

[MaximumC]

I disagree. I see that your "should"-statement is a subjective evaluation of what you would like to see though.

What you (and others) fail to adress or even see is why allowing 99% to be 100% is very problematic: where do you draw the line? If a player is playing a combo that doesn't always win, strictly speaking, we can not just hand him that last few %s for free. It doesn't matter whether we are talking about 0,00..1% or any other number as that would be an arbitrary choice.

Yes, I'm saying that I understand how the rules currently work (I think) but that I would prefer they worked differently.

I'm NOT saying that a combo that has a flat 90% chance of winning should be called a win. I don't even know what that kind of win would look like -- not flipping your singleton Blightsteel off of Bob in a 100 card deck? I'm not even saying a 99.9999999999% possibility to win should count as a win. I'm saying that 99.9 (9 repeating) should be. I'm saying that I'd prefer if the rule was this:

"It is also slow play if a player continues to execute a loop without being able to proof that, for any given probability N, there is a number of iterations, or the limit of an infinite series of iterations, that will increase the probability of an expected resulting game state above N." IPG 3.3.

Even here I am probably phrasing the limit concept incorrectly, but the point here is that Magic should let you touch the infinite. If you can get arbitrarily close to 100% simply by taking more iterations, then the limit is 100%. If you truly can set up a situation where you get to roll the dice as many times as you like, and you win if it ever rolls a 6, you should be allowed to call that a win. (IMHO)

[rufus]

...

Even here I am probably phrasing the limit concept incorrectly, but the point here is that Magic should let you touch the infinite. If you can get arbitrarily close to 100% simply by taking more iterations, then the limit is 100%. If you truly can set up a situation where you get to roll the dice as many times as you like, and you win if it ever rolls a 6, you should be allowed to call that a win.

If you want to make rules, keep in mind they have to handle scenarios like this:

Hero: No cards in hand, no cards in library, no cards in graveyard, 1 life, the following cards in play:

Panoptic Mirror (imprinted with Stitch in Time)
Panoptic Mirror (imprinted with Lava Spike)

Intruder Alarm
Obstinate Familiar(tapped)

Villain: No cards in hand, no cards in library, no cards in graveyard, 20 life, the following cards in play:

Panoptic Mirror (imprinted with Angel's Mercy)
Obstinate Familiar(tapped)

Hero's turn is starting.

... I'm pretty sure that Hero's chance to win here - in the limit - is more than 0, but less than 1. If you use Beacon of Immortality instead of Angel's Mercy, it's certainly the case, but it also becomes futile very quickly.

Edit: I'm not entirely sure how the existing rules would handle that scenario.

[MaximumC]

If you want to make rules, keep in mind they have to handle scenarios like this:

Hero: No cards in hand, no cards in library, no cards in graveyard, 1 life, the following cards in play:

Panoptic Mirror (imprinted with Stitch in Time)
Panoptic Mirror (imprinted with Lava Spike)

Intruder Alarm
Obstinate Familiar(tapped)

Villain: No cards in hand, no cards in library, no cards in graveyard, 20 life, the following cards in play:

Panoptic Mirror (imprinted with Angel's Mercy)
Obstinate Familiar(tapped)

Hero's turn is starting.

... I'm pretty sure that Hero's chance to win here - in the limit - is more than 0, but less than 1. If you use Beacon of Immortality instead of Angel's Mercy, it's certainly the case, but it also becomes futile very quickly.

Edit: I'm not entirely sure how the existing rules would handle that scenario.

That's a fascinating situation. Let's see if I can work out how that works under the current rules, and then under the limit-based rule I prefer.

Okay, so on Hero's first turn, he deals 3 damage to Villain (-3).

Next, 50% of the time, Hero will take another turn and deal another 3 damage (-3). The other half of the time, the Villain will get a turn and THEN the Hero gets a turn, meaning the Villain gains a net 4 life (7 - 3 = 4).

So you have a series where you start at 17 (Hero gets the first turn) and then randomly apply (-3) or (+4) in sequence. I don't have the math in front of me -- it's been awhile since I did series sums -- but I think that, given infinite time, the Hero wins for the same reason that the House always wins at a Casino; over a long enough period of time, the Hero can win enough consecutive rolls to deplete the Villain's limited life total, but the Villain can NEVER end the game by paying life. The Hero can simply repeat the process through infinity, assured in the knowledge that at some point the infinite series will include enough -3's in a row to deplete however much life the Villain has at that point. Again, I haven't done the math here ( I wish someone would! ) but my suspicion is that for any arbitrarily long string of -3s, if you extend the series far enough, you will hit a string of -3s that long.

So, under the current rules, what's going on here? First off, I don't think it's a loop. Unlike the Mijae Djinn situation, where a missed roll results in NO life being lost, and the Four Horsemen, where a whiff results in another randomized library, here a whiff by Hero actually does change the board state. Life totals change. So, it's not a loop and you can't be nailed for slow play under IPG for repeating it. I think. As a result, I think you CAN propose a shortcut in this situation. The only trouble is that, without a mathematical proof under your belt, the opponent and judge probably would not be certain that you would eventually win this way, and make you play it out. Even if you sat there and just flipped coins until time was called, there's no guarantee you'd be done. Finally, unlike the Djinn loop, this loop crosses turns, so eventually you will go to time and tie on turns. In that sense, it's less of a tournament problem.

Under my rules, IF there was a mathematical proof that the limit of your chances of winning was arbitrarily close to 100%, given infinite time, then you could propose a loop and just win.

EDIT: Upon some research, it appears that the sum of a series of randomly increasing and decreasing elements might not exist in the normal sense; there may be no convergence to find. Something something Cesaro sum something something. That may not necessarily mean that the series does not, at some point, produce a sum of 0 or less.

EDIT2: This appears to be how you might start calculating the average value for the infinite sum of random numbers (-3 and +4) but it doesn't address the range of sums you might expect to get along the way to that average: http://en.wikipedia.org/wiki/Wald's_equation

[rufus]

That's a fascinating situation. Let's see if I can work out how that works under the current rules, and then under the limit-based rule I prefer.
...
So you have a series where you start at 17 (Hero gets the first turn) and then randomly apply (-3) or (+4) in sequence. I don't have the math in front of me -- it's been awhile since I did series sums -- but I think that, given infinite time, the Hero wins for the same reason that the House always wins at a Casino; over a long enough period of time, the Hero can win enough consecutive rolls to deplete the Villain's limited life total, but the Villain can NEVER end the game by paying life.
...

It's a random walk, not a series sum. Unless the game is over. Hero's chance to win will be greater than 0 and less than 1.


So, it's not a loop and you can't be nailed for slow play under IPG for repeating it. I think.
There are still repeated states.


Under my rules, IF there was a mathematical proof that the limit of your chances of winning was arbitrarily close to 100%, given infinite time, then you could propose a loop and just win.
You really want judges to handle the arguments or decipher the proofs?

[pandaman]

Back to the original question... Four Horsemen is and will always remain a horrible deck. It is basically a worse version of Breakfast, which has probably been outclassed by the new Oops All Spells deck (which isn't even that good either for that matter).

As to the rules question, no you cannot shortcut the loop because you cannot say when you will get the result you want. Practically speaking, you could literally sit there for 50 minutes and never actually get the combo and both players would receive a draw. Alternatively, the Four Horsemen player could be up game 1, and loop game 2 for the remainder of the round. Neither of these things promote a good tournament environment.

I agree that is now a worse version of Breakfast. However, I think it used to be better than Breakfast, because: (1) Breakfast combo pieces (Nomad's en-Kor and Cephalid Illusionist) could be hit by spot removal while Basalt Monolith and Mesmeric Orb could not; (2) Mesmeric Orb also had the handy side-effect of stuffing up people's Ponder, Brainstorm, and Sensei's Divining Top ordering; and (3) the combo pieces were colourless, so ou could keep the deck in UB, rather than having to go Uwg (usually) for Breakfast. I think the offset in speed used to be worth it for these advantages.

However, we now have Abrupt Decay, which hits both Horsemen combo pieces, and Cavern of Souls, which lets Breakfast cast uncounterable combo pieces, so even if the rules hadn't changed, I don't think it would be worth playing it over Breakfast now.

[BadLegacyPlayer]

I do not think this deck is good. It isn't broken but i think they did the right thing. I do not care if you broken combo me and kill me on turn 2. But the fact that you have to sit forever while your opponent repeats the same monotonous thing just sucks. There is nothing worse than a 30 minute turn. Although I am a huge fan of believing magic should be an interactive game that involves turning creatures sideways, giving your opponent a chance, and having fun, I would much rather be flash hulked out of a game than sit through this monotonous crap.

[nevilshute]

Thanks for all the replies. Though a lot of them seem to center on whether or not it's key interaction should have been deemed as slow play. There are already other threads debating that topic I believe.

Obviously, once having the Basalt Monolith and Mesmeric Orb in play against an opponent unable to interact with the stack or the graveyard, the deck would theoretically "win on the spot". What I was interested in learning more about was how smooth the deck ran as compared to other (2-card) combos? It seems strong to be able to run both library manipulation (Brainstorms, Ponders, Lim-Dûl's Vaults) and permission (Force of Will, Pact of Negation etc) to protect a two card combo involving a 2CMC and a 3CMC card.

[lordofthepit]

It was so good that the only significant result I've seen with the deck came after the deck was axed by the slow play rules change. It's that resistant to everything.

[rufus]

Thanks for all the replies. Though a lot of them seem to center on whether or not it's key interaction should have been deemed as slow play. There are already other threads debating that topic I believe....

Sorry about the digression into undergraduate math.

The Mesmeric Orb/Basalt Monolith combo has some pretty potent fudamentals - both elements can be found with the same tutors and being able go off for :5: is exceptional. On the downside, the graveyard dependence makes you pretty vulnerable to hate aimed at other decks. If you can handle running two more 'dead' cards and marginally less resilience there are also less loopy ways to finish.

I'm not sure it's seen much refinement or play.

[Zilla]

If you could declare an infinite repetition there would always be a 100% chance of any possible outcome occurring, so performing the combination each time would not be necessary, but again, mechanically Magic doesn't recognize infinite numbers; so you can say, "I'll target Daru Spiritualist a number of times equal to Graham's Number," (http://en.wikipedia.org/wiki/Graham%27s_number) but you can't, say, "I'll untap Basalt Monolith a number of times equal to Zilla's mom's mass in kilograms."
My mom weighs 98 pounds. A more apt statement would be "I'll untap Basalt Monolith a number of times equal to the cubic footage contained within the gaping maw that is Zilla's mom's vagina."

[TheInfamousBearAssassin]

If your mom's vagina has an infinite circumference, then it follows that it must also have infinite mass and ergo so too your mom as a whole.

[MaximumC]

Sorry about the digression into undergraduate math.

The Mesmeric Orb/Basalt Monolith combo has some pretty potent fudamentals - both elements can be found with the same tutors and being able go off for :5: is exceptional.


Exceptional? Cephallid Illusionist + Shuko is only 2U for the same effect.

[rufus]

Exceptional? Cephallid Illusionist + Shuko is only 2U for the same effect.

Huh. It seems like most of the combos that people do play end up costing 6 mana.

[MaximumC]

Huh. It seems like most of the combos that people do play end up costing 6 mana.

Yeah, I never understood that. I think Breakfast top 8'd a SCG Open last year once and lost in the finals to Hive Mind, but other than that, you never see this dirt-cheap combo do anything. It seems really bizarre since it's so easy to assemble.

I mean, if you want a Four Horseman-style deck, why not do something like this:

Combos(9)
4 Cephalid Illusionist
4 Nomads en Kor
1 Shaman en Kor

Tutor (5)
3 Worldly Tutor
2 Eldrami's Call

Win Package (8)
3 Narcomebia
1 Bridge from Below
1 Dread Return
1 Angel of Bringing Back da Humans
1 Laboratory Maniac
1 Azami

Disruption (20)
4 Force of Will
4 Spell Pierce (with Diverts in the board against BUG)
4 Daze
4 Cabal Therapy
4 Thoughtseize

(18)
4 Lotus Petal
4 Chrome Mox
4 City of Brass
4 Gemstone Mine
2 Tundra

[TorpidNinja]

Exceptional? Cephallid Illusionist + Shuko is only 2U for the same effect.

2U plus the cost of another creature - you can't unequip equipment after it's attached, only move it around.

[MaximumC]

2U plus the cost of another creature - you can't unequip equipment after it's attached, only move it around.

No, that's not correct. There's nothing in the equip rules that prohibit you from paying the cost to re-equip to the same creature.

[Darkenslight]

No, that's not correct. There's nothing in the equip rules that prohibit you from paying the cost to re-equip to the same creature.

Yes, yes there is. I actually got a GRV playing Life in Extended. You can't re-attach a permanent to the permanent it's already attached to.

[MaximumC]

Yes, yes there is. I actually got a GRV playing Life in Extended. You can't re-attach a permanent to the permanent it's already attached to.

And what was the rule that you supposedly violated?

The Comprehensive Rules say:

702.6a Equip is an activated ability of Equipment cards. "Equip [cost]" means "[Cost]: Attach this permanent to target creature you control. Activate this ability only any time you could cast a sorcery."

Nothing in that definition requires you to attach it to a different creature.

EDIT: Note, however, that if the particular equipment gives shroud (hello Lightning Greaves) then you can't re-equip because the Greeves can no longer target the creature it is equipping. That's a function of the shroud ability, NOT of the equip ability, and it's a weakness that Shuko does not share.

EDIT2: Also note that, if you activate an equip ability targeting the creature already equipped, it won't do anything when it resolves. That's because of this Rule:

701.3b If an effect tries to attach an Aura, Equipment, or Fortification to an object it can’t be attached to, the Aura, Equipment, or Fortification doesn’t move. If an effect tries to attach an Aura, Equipment, or Fortification to the object it’s already attached to, the effect does nothing. If an effect tries to attach an object that isn’t an Aura, Equipment, or Fortification to another object or player, the effect does nothing and the first object doesn’t move

However, that does not stop you from putting the ability on the stack, and so you can trigger Cephalid just peachy. It just means the equipment doesn't actually get "re-attached" when the ability resolves. So, no, you cannot "re-attach" a piece of equipment - which might be what you meant in your post - but you certainly can "re-equip" it.

[PirateKing]

Comprehensive rules don't seem to bear on that. To equip it seems the limitations are:


It's your turn
The target is a creature
You control the target
The equipment isn't a creature
None of the above change during resolution



I await to be publicly humiliated for my poor understanding of this game.

Also hooray for staying on topic!

[MaximumC]

Also hooray for staying on topic!

This is all related to the Four Horseman, and it's a fascinating discussion of math and rules minutiae without anyone flaming. It's The Best Thread!



Comprehensive rules don't seem to bear on that. To equip it seems the limitations are:
....
I await to be publicly humiliated for my poor understanding of this game.

If someone's got a cite, I'm going down with you.

But I strong suspect we are both correct because we, you know, read the rules.

[DragoFireheart]

If your mom's vagina has an infinite circumference, then it follows that it must also have infinite mass and ergo so too your mom as a whole.

But infinite values aren't allowed at tournaments. If she ever tried to go to one she'd get a Game Loss just for existing and would be escorted outside by security.

"Sorry ma'am, your vagina has infinite circumference. Your kind isn't allowed here."

[anomie-p]

What you (and others) fail to adress or even see is why allowing 99% to be 100% is very problematic

This isn't 99% vs. 100%, though. If you are sufficiently randomizing, you *will* eventually reach the win, you just can't ever say how many manually executed loops you would need to get there. (Executing, say, two loops a minute or something, it might take you a minute, it might take you five minutes, it might take you 10^24 years, if you're unlucky ... but you can't fail to get there given an unlimited number of loops unless there's a bias in your shuffle - in which case you're not sufficiently randomizing)

As for the OP question, there's a guy at my LGS who built it a year or so back, we played some test games, he never combo'd off becuase I'd always counter a peice and kill him before he could recover. Since I'm a terrible player, and I never lost to it, I don't think it is very good, but this is of course a small data sample.

[Julian23]

This isn't 99% vs. 100%, though.

It is. Your point about "infinite" repetitions is pointless as others have explained before. Therefore we are looking at players that want their %s upgrades for free, which is not okay because of fairness and stuff.

[anomie-p]

It is. Your point about "infinite" repetitions is pointless as others have explained before.

Mathematics is not pointless. The probability of not winning is (3/4)^n where n is the number of loops you execute. Take the limit of that as N approaches infinity, you get 0. The probability of winning is therefore 1 - 0 == 1, not 0.99

I am not saying the ruling is wrong, or anything like that, as the rules require stating an exact number of loops, which you obviously cannot do.

I am saying that this is not a case of 99 being treated as 100, as the chance of reaching the required graveyard configuration is *actually* 100%. It is not 99, it is not 98, it is not 99.9999999999999999999999999998, etc, it is 100%.

If it were a matter of someone saying, say, 'I draw this number of cards, the odds I get the cards I need are 99%, you should scoop' -> then someone is suggesting treating 99 as 100.

[Julian23]

If it were a matter of someone saying, say, 'I draw this number of cards, the odds I get the cards I need are 99%, you should scoop' -> then someone is suggesting treating 99 as 100.

That's exactly how the Four Horsemen combo operates in tournament practice.

My point about your mathematics being pointless was with regards to the fact that they are based on something that doesn't exist in Magic.

[anomie-p]

That's exactly how the Four Horsemen combo operates in tournament practice.

My point about your mathematics being pointless was with regards to the fact that they are based on something that doesn't exist in Magic.

When someone is suggesting you change the rule to allow something that actually has a 100% probability, but cannot be done due to the current rules, how is it valid to object to this on the basis of the current rules and current practice?

I don't think allowing 'until you are dead' is right, because now you're allowing a player to attempt to make a mathematical argument to a judge who may or may not understand the necessary math to explain it (so I am ok with the current ruling and having to name an exact number of loops, etc).

I just don't see how you can claim that the probability is not actually 100%, and the fact that the horseman combo cannot actually get to the 100% probability in practice under the current rules does *not* mean that someone suggesting an alternate method allowing the combo is treating 99% as 100%, because any suggestion to allow the combo requires a rules change to begin with.

[Julian23]

I just don't see how you can claim that the probability is not actually 100%

Because it requires the existance of this number called "infinity", that doesn't exist in the Magic spectrum of numbers.

If you want to introduce infinity to Magic, there are a whole lot of problems this would bring about and I'm glad we don't have that.

[anomie-p]

Because it requires the existance of this number called "infinity", that doesn't exist in the Magic spectrum of numbers.

If you want to introduce infinity to Magic, there are a whole lot of problems this would bring about and I'm glad we don't have that.

I agree with your second statement - but this does not mean the actual probability is not 100%.

If the actual probability is 100%, and someone suggests a rule change where you can say 'I repeat this until you are dead', it is not valid to object by saying 'that's comparing 99% to 100%', because the *actual* probability under the proposed rule change *is* 100%.

It is *entirely* valid to object saying 'well, this causes problems such as expecting every single judge to be able to follow this or that mathematical statement of the probability, etc', 'allowing this would cause a bunch of problems with this interaction over here', etc.

Your statement that 'it requires the existance of this number called "infinity", that doesn't exist in the Magic spectrum of numbers' is true, but it is not an argument against allowing infinity in the magic spectrum of numbers. Claiming it is, to me, is saying 'Changing the rule is not possible because that would change the rule', it's just circular.

I am not advocating a rule change, I am trying to point out how the claim that this is a matter of 99% vs. 100% is not actually true.

[Bed Decks Palyer]

Player A: I will keep attacking you until you are dead.
Player B: Alright, I have no responses until we reach that game state.


EPIC PLAINSWALKERS BATTLE at its best! I laughed so loud you might even heard it. :laugh:

[Julian23]

Your statement that 'it requires the existance of this number called "infinity", that doesn't exist in the Magic spectrum of numbers' is true, but it is not an argument against allowing infinity in the magic spectrum of numbers. Claiming it is, to me, is saying 'Changing the rule is not possible because that would change the rule', it's just circular.

To me this while conversation has never been about changing the rules in the first place. The purpose is to show people that feel entitled to be allowed to win with Four Horsemen that their claim is invalid.

And as you said yourself, changing the rules isn't even an option because of the precedence it creates: while the probabilities of this combo are very basic and easy to understand, this might open the door to way more complicated interactions and calculations that we don't want unfamiliar judges to make, especially in the tense situation of a big tournament setting where time is an issue.

[anomie-p]

*Shrug*. You appear to not want to catch my point so, there's not really much I can do, eh? :)

[Julian23]

You have yet failed to make an actual statement, besides challenging my observations of 99,..9% != 100% in Magic context.

[anomie-p]

You have yet failed to make an actual statement, besides challenging my observations of 99,..9% != 100% in Magic context.

Only if you ignore the fact that the post you made that observation in was in response to a statement saying something should be allowed, that would require a rule change in the first place, making the current 'Magic context' irrelevant (because it requires the 'Magic context' change to begin with)

Which really only supports my statement that you've decided not to (or perhaps, just didn't) catch my point.

[Julian23]

I just went back to double check what you mentioned. So do you have an actual opinion on how to handle on this or was there no other undlying purpose here?


Which really only supports my statement that you've decided not to (or perhaps, just didn't) catch my point.
You should also probably not invovle emotional statements like this as they serve no other apparant purpose than provocing people, also known as trolling.

[rufus]

To me this while conversation has never been about changing the rules in the first place. The purpose is to show people that feel entitled to be allowed to win with Four Horsemen that their claim is invalid....

Considering it has a resonable expectation to be finished quickly, I tend to think that the Four Horsemen combo should be allowed to play out. That said, I'm not sure there's a bright line to draw on the spectrum of progressively slower (or more desperate) possible finishers.

[anomie-p]

It can be played out, the player just has to be careful that they do something to advance the game state (which puts a cap on the number of possible iterations, making it possible to fizzle).

If you could shortcut it, it wouldn't be slow, it'd just be a matter of the player stacking the GY the way they want (which in practice wouldn't even be stacking the full graveyard, just making sure four cards were in the correct order relative to each other)

[Koby]

Did you just evoke Mathematics, then pronounce the limit of an infinite fractional exponential series to EQUAL zero?

(X/Y)^n, where Y > X, and N tends to infinity is approximated as Zero. It never reaches the value.

Therefore, there is still a 0.0000000...000001% chance of failing to reach the desired state. In a practical sense, this is Zero. If you're considering infinity as a quantity, then you can't also approximate the limit of an infinite series.

Thus the probability of reaching the state is 1-0.0000....0001 = 99.999....9999%; this value is not 100%.

Ultimately, the issue is with the execution of the loop. Math proofs are not considered an accepted explanation of a loop. Moreover, there is no concept of infinity in Magic terms; only Really Large Numbers. The Four Horsemen loop iterated a 'really large number' of times, is most certainly not Zero, even if it appears to be heading to that value.

[MaximumC]

Did you just evoke Mathematics, then pronounce the limit of an infinite fractional exponential series to EQUAL zero?

(X/Y)^n, where Y > X, and N tends to infinity is approximated as Zero. It never reaches the value.

Therefore, there is still a 0.0000000...000001% chance of failing to reach the desired state. In a practical sense, this is Zero. If you're considering infinity as a quantity, then you can't also approximate the limit of an infinite series.

Thus the probability of reaching the state is 1-0.0000....0001 = 99.999....9999%; this value is not 100%.

Ultimately, the issue is with the execution of the loop. Math proofs are not considered an accepted explanation of a loop. Moreover, there is no concept of infinity in Magic terms; only Really Large Numbers. The Four Horsemen loop iterated a 'really large number' of times, is most certainly not Zero, even if it appears to be heading to that value.

I believe you're quibbling with infinitesimals and the concept of the limit, an issue that Newton and Leibniz overcame to invent calculus in the first place. Say you have an infinite series tending towards X. If you can prove that, for any arbitrarily small number Y, you can get within Y of X by simply continuing the series long enough, then the sum of the series is in fact X. This is the whole "delta" analysis in the fundamental theorem of calculus, I recall.

It is also why - and this is gnarly - 0.999... where 9 repeats to infinity is actually EQUAL to 1. It's not "close" to 1. It's EQUAL to one. In our number system, every number can be written in two ways; a finite expression and an infinite expression as a series in this way.

Now, mind you, not every infinite series actually converges. Some do not. But if you can prove that a series converges to a number, the sum of the series IS that number.

There is no expression of this concept in Magic rules right now, that's true. I think there should be because infinite series math is awesome and because it really isn't that hard to evaluate the infinite series that might actually pop up in Magic. Plus, it wouldn't hurt to require judges to pass Calculus exams as well as rules exams to get accredited! :D

At a fundamental level, it seems unfair to me that decks like the Four Horseman, which have locked up the game in a way that can only ever result in a win -- with mathematical certainty -- end up not winning simply because of time constraints. That's bogus.

[TheInfamousBearAssassin]

.999...=1 is correct.

It is also pretty pedantic to rule that since there's a non-zero chance of the combo fizzling as long as the iterations are not infinite, it's stalling. We could set the number arbitrarily high, say we go with a Graham's number of iterations, and the odds of the combo fizzling are significantly less than the odds that the entire tournament and all of the reality surrounding it are holographic projections that just spontaneously came into existence in media res, meaning that in real terms the previous plays never took place and thus the current board state is illegal; or the odds that all of your opponents will spontaneously turn into miniature black holes leaving you the default winner of the tournament, or that an apparently illegal decklist is a result of the molecules in a given card spontaneously rearranging on a molecular level to turn a mountain into a fifth Goblin Lackey. But judges seem to feel enough certainty about these events despite the incredibly tiny odds of quantum shenanigans to treat them as givens.

[anomie-p]

I believe you're quibbling with infinitesimals and the concept of the limit, an issue that Newton and Leibniz overcame to invent calculus in the first place. Say you have an infinite series tending towards X. If you can prove that, for any arbitrarily small number Y, you can get within Y of X by simply continuing the series long enough, then the sum of the series is in fact X. This is the whole "delta" analysis in the fundamental theorem of calculus, I recall.
...

At a fundamental level, it seems unfair to me that decks like the Four Horseman, which have locked up the game in a way that can only ever result in a win -- with mathematical certainty -- end up not winning simply because of time constraints. That's bogus.

Although, the post you are responding to illustrates why it's not necessarily a good idea to allow the ruling to go the way mathematical certainty demands, no?

I mean, I was looking at the facebook comments on an article on a different site and someone claimed that 'no respectable mathematician believes' that 0.999... == 1

[MaximumC]

Although, the post you are responding to illustrates why it's not necessarily a good idea to allow the ruling to go the way mathematical certainty demands, no?

I mean, I was looking at the facebook comments on an article on a different site and someone claimed that 'no respectable mathematician believes' that 0.999... == 1

There no respectable mathematicians who would claim 0.999... does not equal 1. You can prove it in a variety of ways. Here's a cite explaining a few of them. http://www.purplemath.com/modules/howcan1.htm

If your point is that non-mathematicians may not be familiar with this concept, then I get that. On the one hand, people who doubt a correct mathematical proof are probably worse than climate change deniers in terms of lack of understanding, but on the other hand it might be easier for Magic rules to recognize they won't be dealing with the cream of the crop all of the time, and make rules that are less likely to confuse the "rabble."

I dunno, I guess I have more faith in the fact that you can educate people on this point. Heck, look at how judges have to keep explaining how Oblivion Ring tricks work. Just because there's a good chunk of the player base that doesn't "get it" doesn't stop judges from learning otherwise and then educating them.

[rufus]

There no respectable mathematicians who would claim 0.999... does not equal 1. You can prove it in a variety of ways. Here's a cite explaining a few of them. http://www.purplemath.com/modules/howcan1.htm

Being able to approach something arbitrarily closely isn't the same a reaching it. (Electrons don't go the speed of light.)

Moreover, it's not always going to be easy to determine whether the win probability approaches 1 over repeated iterations. Consider, for example the Stitch in Time scenario I posted above.

In a scenario where the win probability does not approach 1, the win probability can get arbitrarily small - do we want to let players repeat actions 'on a wing and a prayer' until time is called?

[anomie-p]

On the one hand, people who doubt a correct mathematical proof are probably worse than climate change deniers in terms of lack of understanding, but on the other hand it might be easier for Magic rules to recognize they won't be dealing with the cream of the crop all of the time, and make rules that are less likely to confuse the "rabble."

I dunno, I guess I have more faith in the fact that you can educate people on this point. Heck, look at how judges have to keep explaining how Oblivion Ring tricks work. Just because there's a good chunk of the player base that doesn't "get it" doesn't stop judges from learning otherwise and then educating them.

I do think that people can be educated, I just think that perhaps a tournament floor is not necessarily the best place for that to happen with respect to a mathematical point. Plus, I don't think judges would be immune, unless you want them ruling on rote in arbitrary situations (this player over here is slinging some mathematical B.S. that sounds nice, they get the win when it's actually not mathematically certain, etc). In some respects it's just cleaner to require that the loop count be definite.

Not that the people I'm considering here are rabble, just that there's evidence there are a fair number of people who might not understand the mathematical truth. (they may be judges, they may be players, etc), and at some point you have to balance the value of being mathematically correct against 'how well would that work in practice'.

On the other hand, it could also be fine to allow indefinite loop shortcuts, but avoiding the issue entirely as the current rules do is a workable solution at the least. (I'm a little ambivalent here, to be honest - maybe preclearing a deck with a win con that requires an indefinite loop with the head judge might be sufficient, depending on, say, the odds that he or someone else helping him can reliably go/no-go such a loop)

[MaximumC]

@anomie: Yea, that's all probably true. I just wish there was some way for a deck which is mathematically certain to win, just not in a finite amount of time, would be allowed to win. Calling it slow play just seems like ascribing bad faith to someone without cause.


Being able to approach something arbitrarily closely isn't the same a reaching it. (Electrons don't go the speed of light.)

Moreover, it's not always going to be easy to determine whether the win probability approaches 1 over repeated iterations. Consider, for example the Stitch in Time scenario I posted above.

In a scenario where the win probability does not approach 1, the win probability can get arbitrarily small - do we want to let players repeat actions 'on a wing and a prayer' until time is called?

Rufus, my man, you're just wrong about this. Read the proof(s). An infinite series that converges to a number is in fact equal to that number. Your electron example is not an infinite series. A proposed repeated behavior in Magic may well be. Magic is an artificial system of rules that can be modeled mathematically and where the laws of physics do not apply to restrain infinite behavior.

The trick is whether, for a given scenario, you can actually come up with a mathematical proof that your win change converges to 100%. That calculation is unique for each potential combo, and the whole point to the rules change I like is that you can't play that deck unless you've done that calculation.

[anomie-p]

In a scenario where the win probability does not approach 1, the win probability can get arbitrarily small - do we want to let players repeat actions 'on a wing and a prayer' until time is called?

In something like modo it should be fairly easy to let someone set a condition like 'these three cards above this card in the graveyard' and have the software loop it for them as fast as it can until they either hit or *their* time runs out ;) (although I doubt anything like that will be implemented in actual modo)

Edit: also, if you're not approaching, I don't think you can get 'arbitrarily small', as not approaching means the variable either becomes the value at some point, and then passes it (in the way you can't say x = y + 1 'approaches 1'), or the function will reverse away at an exact quantity (which is not arbitrary).

[rufus]

...
The trick is whether, for a given scenario, you can actually come up with a mathematical proof that your win change converges to 100%. That calculation is unique for each potential combo, and the whole point to the rules change I like is that you can't play that deck unless you've done that calculation.

And then you want the judges to decide whether the proof is valid? :laugh:

[MaximumC]

And then you want the judges to decide whether the proof is valid? :laugh:

Judges should all have math degrees, that's my position. :tongue:

[TsumiBand]

I grok the notion that .999… == 1 and I'm familiar in a passing way with the notion that infinity can be expressed in unique ways in mathematics, probably the more infamous of which is the old 1 + 2 + 3 + 4… = -1/12 idea. I think this is all very cool, actually, though I know I do not understand it.

Whether or not I can follow this math and understand or explain the nitty-gritty is one thing. Whether or not the concepts are being applied correctly is another. I don't think this is a situation where it is worthwhile to point out that X iterations of a combo yield an eventual win; as Magic players we scoop to things we can't answer but we don't hardcode it into the rules of the game that if your opponent plays something which you cannot answer satisfactorily that you are required to pick up your cards and mark your scoresheets.

Take the theoretical case of a deck which can not demonstrate a win, but rather can only demonstrate a gamestate where it cannot lose. "Not losing" does not translate to "winning"; not because there are multiple paths to victory in Magic, but because the two phrases mean different things.

I am not entirely sure why one would argue for any infinities in Magic -- the rules of the game do not allow for them, but yet some would use equivocations to 'backdoor' infinity into the game. The fact is that one does not simply approach infinity; to suggest its application wherein one "approaches" infinity, or in a situation where there's a ceiling on the number of iterations we engage in as players -- even if it is declared to be immensely high -- still represents a termination. .999… == 1, that's fine, but .999… != .999999999999999999999999999999999999999 != 1, nor does any number where you end it by discontinuing to say 'nine'. The number stopped. It did not 'get close enough' to infinity, precisely because it ended.

Likewise, 'positive infinity' is not 'approached' when someone says "I'll demonstrate the combo once, and then I'll declare that I've done it 42.7 quintillion times." There's a gamestate 42.7 quintillion iterations in the future which can be arrived at, and your opponent may not be demonstrably dead. You cannot declare "I do this until you are dead" because unless you can assign a hard number to it, you're declaring an infinity in a system that doesn't acknowledge them.*

Again, I'm totally fine with the math involved in stating that the sum of all numbers is -1/12 and that .999 == 1 because those are situations that deal with infinite values. But really I'm reminded of the paradox of concept of expressing distance as a ratio between two objects, which therefore says they can never meet if moving in the same direction. It's not Zeno's arrow, it's something else, I don't remember right now. But the point is, it's a misapplication of the math. The proofs and concepts that support it aren't really acknowledged by Magic in the first place, because Magic says, "An infinite number, eh? Welp, let me get my clippers; I say it's actually *snip* only this big." I would submit that it applies to concepts such as non-terminating decimals as well, which means that pointing out where 99% != 100% is entirely an accurate statement in the vein of the math as it applies within Magic.

That doesn't mean I'm a proponent of the deck or anything, it's not Guys and Burn or White Weenie, so I probably hate it. vOv

EDIT: …unless you can guarantee that X loops results in N damage with no chance for variance. As in, one loop = 2 damage to the face, or three loops = 1 damage, or some other convoluted method of actually proving what the damage/loop actually is, such that "until you are dead" is just a variable you can solve for.

[anomie-p]

I would submit that it applies to concepts such as non-terminating decimals as well, which means that pointing out where 99% != 100% is entirely an accurate statement in the vein of the math as it applies within Magic.

I don't think anyone has stated that this would not be the case with respect to the current rules of Magic. If you have to name a definite number, you are stuck short.

The point I was making is that responding to someone who is suggesting a means whereby you might shortcut, which by itself would require you not name a definite number, but a condition, with 'but that's treating 99% as 100%', is basically just completely ignoring that this implies a rules change to begin with, and fundamentally simplifies to 'we can't change the rule because the rule is what it is'.

[MaximumC]

I grok the notion that .999… == 1 and I'm familiar in a passing way with the notion that infinity can be expressed in unique ways in mathematics, probably the more infamous of which is the old 1 + 2 + 3 + 4… = -1/12 idea. I think this is all very cool, actually, though I know I do not understand it.

Gooby plz. The -1/12 "proof" you suggest is fallacious. They bury an error in the "proof" that is not obvious to laypeople the same way that all those algebraic proofs that 1=2 usually sneak in a division by zero.

Anyway, the rest of your post is totally correct. Current rules do not allow you to actually repeat an action an infinite number of times, only an arbitrary number of times. What I'm saying is that if you prove that you WILL WIN with 100% certainty if you were actually allowed to repeat an infinite number of times - or "until you die" - then that'd be a good thing to add to the rules. As others point out, proving the certainty of the win condition is itself problematic enough to keep the rules the way they are.

[Finn]

Hi folks. This has come up a few times in the past. This is the information I have given in the past to try to quell this discussion.

Since infinity is not a number, the likelihood of getting the cards stacked in the graveyard in the proper order can not be agreed to happen after x number of iterations. That is, it is not quite a certainty even at a very high number. Wotc was right to get rid of the interaction not because of this deck that I made that just happened to function like that. It is because of what could be done with it. If they permitted infinite, but indefinite combos, you could upset deck design big time. Once you rule that you can shortcut to get Sharuum, Station, and Return in the graveyard before Emrakul, you open up a big can of worms. As a simple example to make this clear, I will change the numbers. I could have responded to Disenchant aimed at Basalt Monolith by saying "OK, since I can run through the deck infinite times, I will set up the library such that the only cards left in my library are the remaining three copies of Monolith, plus the one Emrakul." Clearly, they can not permit this sort of thing.

And as said, the deck could have had a place in the meta if given the chance (especially when people started to realize that you could do THIS sort of shenanigan), but it was never amazing. It is moot anyway due to Abrupt Decay.

[MaximumC]

Hi folks. This has come up a few times in the past. This is the information I have given in the past to try to quell this discussion.

Since infinity is not a number, the likelihood of getting the cards stacked in the graveyard in the proper order can not be agreed to happen after x number of iterations. That is, it is not quite a certainty even at a very high number. Wotc was right to get rid of the interaction not because of this deck that I made that just happened to function like that. It is because of what could be done with it. If they permitted infinite, but indefinite combos, you could upset deck design big time.

That doesn't "quell" much because it's basically what we're talking about. Everyone recognizes that the current rules do not permit you to execute a loop (or shortcut it) an indefinite or infinite number of times. You have to be able to specify the number of times you will execute the loop and precisely what the board state will be when you are done. We all get that.

The discussion we're having is whether or not that's a good rule. The situation you suggest doesn't convince me:



Once you rule that you can shortcut to get Sharuum, Station, and Return in the graveyard before Emrakul, you open up a big can of worms. As a simple example to make this clear, I will change the numbers. I could have responded to Disenchant aimed at Basalt Monolith by saying "OK, since I can run through the deck infinite times, I will set up the library such that the only cards left in my library are the remaining three copies of Monolith, plus the one Emrakul." Clearly, they can not permit this sort of thing.


What's the problem here? Say you allowed someone to define a loop until a condition is met at random, or prove that the limit of the probability of the event tends to 100% at infinity, or whatever. Say you let them shortcut Emrakul to the bottom.

In that case, yea, you're right, they could also shortcut to having Monolith plus Emrakul. So what? Why is it "clear" that they cannot permit "this kind of thing?" Says who? Sounds like a robust combo, nothing more. What part of Magic Design does this thwart?

In my mind, the better argument against changing the indefinite loop rules has more to do with expecting floor judges to have the math background to evaluate proofs in bizarre loop situations than it does with strengthening a combo. For example, a few pages back I brought up a Mijae Djinn scenario that was, technically speaking, just as illegal as the Four Horseman was, but since it was very easy to see that it would play out quickly and reach its result without wasting time, people shrugged and said, "Eh, judge can let you do that." Why? Because it's easy to understand.

[TsumiBand]

Gooby plz. The -1/12 "proof" you suggest is fallacious. They bury an error in the "proof" that is not obvious to laypeople the same way that all those algebraic proofs that 1=2 usually sneak in a division by zero.

I thought dude was just keeping it simple for nubs like me. :P anyway I've seen a handful of alternative suggested proofs that point to the same answer from different authors and of course other people have weighed in and disagreed with the entire concept and showcased the different ways to derive different answers. I just find it interesting conceptually vOv


What I'm saying is that if you prove that you WILL WIN with 100% certainty if you were actually allowed to repeat an infinite number of times - or "until you die" - then that'd be a good thing to add to the rules. As others point out, proving the certainty of the win condition is itself problematic enough to keep the rules the way they are.

I can appreciate this to an extent; the Djinn/Assault explain from earlier was a fair example of how that should apply, but doesn't. I think the reasoning is just slightly recursive though, right -- saying "it's true there are no infinities in Magic; however, let us assume there were. I would surely win with this game-state, so while I cannot name X iterations, I can tell you it is less than infinity." Whether or not something is verifiably "less than infinity" isn't really any better than just leaving it indeterminate. I'm kind of a "fuck you, kill me" type, but if I were sitting across from Bear/Djinn/Assault with no answer, I think I would see the inevitability and scoop, but that's a judgment call based on capriciousness.

I guess as an irrational human, in the BearDjinnAssault combo, a 50% whiff rate resulting in an average "1 damage/loop" is enough to convince me to scoop, whereas Four Horses long process and whiff rate give me pause as to whether or not going to time will save my ass. Aggro decks will submit to the clock in this way; I've used this in Limited to prevent myself from flatly losing to a much better deck than mine. I have no idea how to quantify this, but whereas Bear Umbra + M-Djinn + A-Assault can be demonstrated fairly quickly, a deck that deign to show up with a mathematical proof at its side to prove its viability had better, by my estimation, be prepared to demonstrate -- and if it cannot, then what makes it any different from the aggro 50 minutes + 5 turns draw? True-Name Nemesis now creates a board state which could lead to inevitability on the 6th or 7th turn; but if they don't get there in 50 + 5 turns, they are still getting a 1 - 1 - 1 on their scoresheet. I think if we're going to invite the capacity to demonstrate a kill that requires 10^32 turns to kill someone in the blink of an eye, but force other decks to play to time, there is something uncool about that approach as well.

I guess I see this as a situation where we either accept the evil of allowing infinity across the board, or the evil of dismissing it across the board. If we dismiss it, we have to accept that it can not be applied under any circumstance; if we allow for it in limited circumstances, then others necessarily start creeping up. Life.dec with Nomads en-Kor 'should' be able to grant me literally infinite life, for example (or, if you prefer, -1/12 life? ha)

[Finn]

Lemme tell you that in testing The four horsemen, I was constantly asking rules questions regarding the limits if looping until event X. It felt powerful as hell to be able to partially stack your library like that. But looping until an event, rather than a finite number creates problems. What if the opponent has a response that he does not want to reveal, but is best used after some number of iterations and which depends on how many iterations happen? It's not a perfect example but think Children if Korliss used to save you from Tendrils. You have to play it just right or it fails to save you. Before the ruling, I was just taking sick advantage of people this way. Everyone thought it was unfair, and could see why. But I got away with it.

[anomie-p]

I would surely win with this game-state, so while I cannot name X iterations, I can tell you it is less than infinity." Whether or not something is verifiably "less than infinity" isn't really any better than just leaving it indeterminate.

I think there is a misunderstanding here, though. Your probability of hitting a win is 1 due to the your loop being infinite - and since your loop is infinite, *you cannot name a number of iterations*. If you name a number of iterations, your loop is *finite*, not infinite.

Edit: I'm not sure I got at what I meant to say here. You're not saying that the iterations are less than infinity, because if you do you make the loop finite. make sense? Looping to the condition lets you leave the number of loops you will actually execute indeterminate, which is how you can say that the probability is 1 (the loop is infinite).

In practice, the number of actually executed loops *usually* won't be too large (because you don't need a specific graveyard order, just four cards within it). I wrote a small simulation of the card flipping and over 100k runs the highest number of emrakul bins before having the other three cards in the yard was 43 or so.

[TsumiBand]

I think there is a misunderstanding here, though. Your probability of hitting a win is 1 due to the your loop being infinite - and since your loop is infinite, *you cannot name a number of iterations*. If you name a number of iterations, your loop is *finite*, not infinite.

...sure, but you have to solve for X without introducing infinity, or you're just cat-assing.

sent from phone, don't be a dick

[MaximumC]

@Tsumi - What exactly is cat-assing? Do I want to know?


Lemme tell you that in testing The four horsemen, I was constantly asking rules questions regarding the limits if looping until event X. It felt powerful as hell to be able to partially stack your library like that. But looping until an event, rather than a finite number creates problems. What if the opponent has a response that he does not want to reveal, but is best used after some number of iterations and which depends on how many iterations happen? It's not a perfect example but think Children if Korliss used to save you from Tendrils. You have to play it just right or it fails to save you. Before the ruling, I was just taking sick advantage of people this way. Everyone thought it was unfair, and could see why. But I got away with it.

Hmmmm... I don't really follow your explanation, but let me see if I can tease out what you're saying.

Lets' say you are playing the Four Horseman, we have Field of Dreams in play.

You announce you want to execute the loop until Emrakul is on the bottom and propose a shortcut. I tell you, under the shortcut rules, that I want to interrupt you and respond when Emrakul ends up on the very top of your library. Maybe I've got a split-second way of exiling your top card or something, I dunno. This situation leads to an impossible decision, because you cannot know in advance whether Emrakul is gonna hit the top or the bottom of your library first.

If this is what you mean, then THIS truly is the nail in the coffin concerning my idea about a changed rule. This makes it totally impossible to shortcut these kind of loops, and since your alternative is then to play it out, means games will go on forever, potentially, while you shuffle repeatedly.

I'm convinced.

[TheInfamousBearAssassin]

I am not entirely sure why one would argue for any infinities in Magic -- the rules of the game do not allow for them, but yet some would use equivocations to 'backdoor' infinity into the game. The fact is that one does not simply approach infinity; to suggest its application wherein one "approaches" infinity, or in a situation where there's a ceiling on the number of iterations we engage in as players -- even if it is declared to be immensely high -- still represents a termination. .999… == 1, that's fine, but .999… != .999999999999999999999999999999999999999 != 1, nor does any number where you end it by discontinuing to say 'nine'. The number stopped. It did not 'get close enough' to infinity, precisely because it ended.

1) Why would you argue against infinity in Magic? The fact that you can name arbitrarily large numbers but not infinite numbers is a completely arbitrary rule of the game that is in no way necessary to allow the game to function. In fact there is a case in which there is an infinite combo in Magic; players can't choose to do anything an infinite number of times, but an infinite loop can be created by the game state itself (say, nothing but lands and three O-Rings.) This will end the game but it's precisely because it's an infinite sequence that the game ends.

2) Again, if your odds of having an out to a given combo are one over, say, a Googolplex, then it is much less likely that you will win than it is that your opponent was killed by an alien doppelganger made entirely out of cheese who is playing for him, and that the match is thus fraudulent. But if judges don't worry about the latter possibility, why get bent out of shape over the former? Numbers that are very, very big are equivalent to infinity for all functional intents and purposes that mere human beings can care about.


I could have responded to Disenchant aimed at Basalt Monolith by saying "OK, since I can run through the deck infinite times, I will set up the library such that the only cards left in my library are the remaining three copies of Monolith, plus the one Emrakul." Clearly, they can not permit this sort of thing.

Why not? Is there an actual reason or is this, like most of Wizards' supporters rationales, a "clearly" that presupposes total agreement and thus doesn't have any actual underlying basis.

[anomie-p]

You announce you want to execute the loop until Emrakul is on the bottom and propose a shortcut. I tell you, under the shortcut rules, that I want to interrupt you and respond when Emrakul ends up on the very top of your library. Maybe I've got a split-second way of exiling your top card or something, I dunno. This situation leads to an impossible decision, because you cannot know in advance whether Emrakul is gonna hit the top or the bottom of your library first.

If this is what you mean, then THIS truly is the nail in the coffin concerning my idea about a changed rule. This makes it totally impossible to shortcut these kind of loops, and since your alternative is then to play it out, means games will go on forever, potentially, while you shuffle repeatedly.

I'm convinced.

You could just call it a draw along the lines of O-ring, but that seems unsatisfactory.

So yeah, I concur (to the position I was ambivalently holding to begin with ;) )

[TsumiBand]

1) Why would you argue against infinity in Magic? The fact that you can name arbitrarily large numbers but not infinite numbers is a completely arbitrary rule of the game that is in no way necessary to allow the game to function. In fact there is a case in which there is an infinite combo in Magic; players can't choose to do anything an infinite number of times, but an infinite loop can be created by the game state itself (say, nothing but lands and three O-Rings.) This will end the game but it's precisely because it's an infinite sequence that the game ends.

The O-Ring Orgy breaks the game, though. It ends because the game state cannot advance in a legal fashion; you can't apply the "no infinities" rule and achieve a legal game-state. If I elect to repeat the O-Ring loop 2^15 times and then see where the game is, it's right back where it started; a freshly entered-the-battlefield Oblivion Ring with a requisite target. If we 'allow' infinity to persist, this game cannot end. If we force the loop to break, then the rules have effectively countered the effect's resolution. Maybe the latter is preferable because the game-state can advance, but the former is just untenable.

There's no violation here; the game becomes a draw when an infinite loop *cannot* be changed by any player, and therefore stalls the game - which ends the game, which in essence breaks the infinitely choiceless loop. This, as opposed to choosing to maintain an infinite loop with optional actions, wherein a player is forced to break the stall by selecting an arbitrary number. Would you suggest that it's better to let a player choose to execute an infinite loop to its logical conclusion? How many durdly non-game-winning infinite loops are even in Magic? If I control two Kiora's Followers, they can untap each other into infinity. One could argue that one never has to give the opponent priority ever again, thus drawing the game.


2) Again, if your odds of having an out to a given combo are one over, say, a Googolplex, then it is much less likely that you will win than it is that your opponent was killed by an alien doppelganger made entirely out of cheese who is playing for him, and that the match is thus fraudulent. But if judges don't worry about the latter possibility, why get bent out of shape over the former? Numbers that are very, very big are equivalent to infinity for all functional intents and purposes that mere human beings can care about.

Clearly that's untrue. We've known about the cheese aliens for some time now.

But logistically, I will still point back to the Running Of The Clock example. Iterations of combat phases or card drawing can be proven to end a player; yet we still require a definitive win to occur over the course of 50 minutes and 5 turns, because we have to draw a line in the sand somehow in order to keep games/events/tournaments moving forward. Whereas a demonstrable win condition such as "my opponent is at 7, and I control a 1/1 which my opponent cannot block or interact with in any way. We are only allowed 5 turns to play it out; yet I submit that, if we were allowed more, I would definitively win this game. Hence I should be recognized as the winner of this game." What makes the combo deck that can go to time to demonstrate its win any different? If you can't define a number of turns it takes to kill me, and you can't play it out, you're trying to declare victory through indeterminable game-states vis-a-vis "if you'd just let me *try* to count to infinity, you'd see that I can win this game". I don't see a compelling reason to allow this kind of a win in the face of the 50 minutes + 5 turns rule having been used to allow a player to play to a draw.


@Tsumi - What exactly is cat-assing? Do I want to know?

Basically 'just wasting time'. Can be used to refer to an inane in-game activity which is less efficient than proper grinding to achieve XP but is endlessly repeatable, or can also mean locking oneself away just to play MMOs yet not actually attaining any in-game goals.

[TheInfamousBearAssassin]

If it is wasting time or taking an unacceptable short cut to kill someone with a series of unpredictable steps towards an absolutely certain outcome, and they must be forced to play out each step without a shrotcut, I don't see why that's different from killing someone with a series of predictable steps towards an absolute outcome, and in that case the player can take shortcuts.

[MaximumC]

If it is wasting time or taking an unacceptable short cut to kill someone with a series of unpredictable steps towards an absolutely certain outcome, and they must be forced to play out each step without a shrotcut, I don't see why that's different from killing someone with a series of predictable steps towards an absolute outcome, and in that case the player can take shortcuts.

Because then the tournament would never end.

See, it's like this. If you want to be able to win with an infinite combo that will win with 100%, just not in finite time, you can only do that if you can shortcut the process. The rules do not currently allow shortcutting of such processes. I initially didn't see why they couldn't be changed to allow that, but then a few posts up I realized that it is impossible. If you want to shortcut until condition A occurs, but your opponent wants to interrupt the shortcut when B occurs, and you can't tell if A or B will happen first, the whole shortcut process craps in the bathtub.

You can't play it out because the judges cant ever get the next round going. You're at turns, but so what? You never stop taking your turn. It's High Tide and Egg taken to their horrible, terrible extreme.

[MaximumC]

(forum gave me an error and then multiposted)

[TsumiBand]

If it is wasting time or taking an unacceptable short cut to kill someone with a series of unpredictable steps towards an absolutely certain outcome, and they must be forced to play out each step without a shrotcut, I don't see why that's different from killing someone with a series of predictable steps towards an absolute outcome, and in that case the player can take shortcuts.

Tkaen to a logical extreme, this is tantamount to just wearing a shirt with "Deck Name Here" printed on it, along with the over-under percentages of all the possible matchups, and then you just sit at your assigned seat and see who sits down across from you, plug the numbers into a computer, and go "Hm, yeah you probably would have won if we played it out." or "Oh shit son, looks like I beat you in game 3 despite the odds" and so on.

If the goal here is to point out the intangibles of playing it out vs. 'playing it out', that's an absurdity I don't think adds anything. Max-C hit the nail on the head with the 4Horses + Field of Dreams scenario; there's no way to do this reliably without actually playing the game, because the opponent could be holding A Card That Wreck You but only if it is played at the maximal time. If that happens to be iteration #12 or 2^(1+√5/2) turns into the future, who can say?

[anomie-p]

Tkaen to a logical extreme, this is tantamount to just wearing a shirt with "Deck Name Here" printed on it, along with the over-under percentages of all the possible matchups, and then you just sit at your assigned seat and see who sits down across from you, plug the numbers into a computer, and go "Hm, yeah you probably would have won if we played it out." or "Oh shit son, looks like I beat you in game 3 despite the odds" and so on.

Why do you consider this the logical extreme? To me it doesn't really look at all logical, on the one side you've got a mathematical certainty, on the other, just different outcomes with probabilities adding up to one, there's an easy dividing split where you just say 'this outcome is certain, those are not'. (When you hit MaximumC's stated situation, you would logically end up having to play out, as he noted, or call it a draw, both of which are reasons to not allow such loops, but neither of which logically lead to what you are suggesting)

[TheInfamousBearAssassin]

Yeah, I mean they're very clearly and materially distinct circumstances, one is mathematical certainty of a specific outcome, and one isn't. It's not that hard to figure out.

[TsumiBand]

I just don't see any merit in finding ways around actually demonstrably killing your opponent. If they want to scoop they can scoop whenever, but not everyone will, and honestly you shouldn't expect a judge to back up your assertion that if you could shuffle/kind of stack/draw/reshuffle/etc your deck an arbitrarily high number of times that eventually some magic pile of cards is guaranteed to emerge from all that masturbation so your opponent's resistance to the process can be transmuted into a autowin for you, because instead of playing cards you can do maths. Just lay your cards on the table in the right fucking order and quit looking for easy outs. Apparently that's harder than the math required to find the right number of pile shuffles it takes to (maybe) kill your opponent.

[Dice_Box]

You can demonstrate how your combo works, I will watch and I will comprehended. If there is even a 1% chance that you can not make it work in the time we have available I will make you do it. Because you can say that given all the time in the world you can fire it off but you don't have all the time in the world, so that's not a valid argument.

Bring on all the mathematics you like, but if given a limited time can you guarantee it will work? No? Well in the end that's what matters is it not?

I can not load Page 6.

[TheInfamousBearAssassin]

You can demonstrate how your combo works, I will watch and I will comprehended. If there is even a 1% chance that you can not make it work in the time we have available I will make you do it. Because you can say that given all the time in the world you can fire it off but you don't have all the time in the world, so that's not a valid argument.

Bring on all the mathematics you like, but if given a limited time can you guarantee it will work? No? Well in the end that's what matters is it not?

I feel like the problem here is certain people not appreciating how statistics works and how big numbers are big.

We're not talking about things in the 1% range. We don't even have to talk about things in the one quadrillionth of a percent range.

I want you to understand that if I say, "I will execute this 1/4 chance of success combo a number of times equal to Graham's number," your odds of not dying are vastly, vastly less than the odds that you, Dice_Box, will have your experience of reading this sentence that I am currently typing and you are currently reading punctuated by the obliteration of the Earth in a super nova.


VAAAASTLY less.

Like if you took the number of elementary particles that exist in the Universe, and multiplied them by those odds, it would still be more likely, the you being obliterated right now, literally right now this very second, along with everyone you love, by a super nova.

[TheInfamousBearAssassin]

Or more for our purposes, it is far more likely that a card in your deck will spontaneously rearrange on a molecular level to be a different card, thus giving you a game loss; but if I tell a judge that quantum card-swapping is the reason for my illegal decklist they are comfortable treating it as a certainty that I am lying.

This is because we are human being, nothing is certain, we have to get on with life, and for our purposes there is no distinction between infinite numbers and arbitrarily large numbers that actually matters.

[Dice_Box]

You are not reading what I write, your reading what you want to read. I said "If there is even a 1% chance that you can not make it work in the time we have available I will make you do it." That's not the same as saying in this situation you have only a 1% chance. It's me saying that even if in a given situation 1 time out of 100 it will fail, I will still ask you to do it.

In this situation you may make it work 1 time in 3 but I will still ask you to do it. Because I don't give a flying what your numbers say, I am playing to win and there is a chance you will brick. If I have no other outs I am taking that chance.

You don't have infinite time at an event, you are limited. So you can say, given X many goes and there is a high probability I can make this work. My answer will be, "Don't tell me, show me."

[TheInfamousBearAssassin]

No, it's you who aren't listening. I am saying that the odds are not 1%, or 1 in 3. The odds are 1, or as close to it as to be indistinguishable from the perspective of any organic or electric brain capable of doing calculations we have access to or are likely to over the next millenium.

[TheInfamousBearAssassin]

The finite number set, as it turns out, contains numbers that are really really big. So big, in fact, as to replicate infinite for all observable purposes.

[Dice_Box]

There is an error with this thread. I can not load past Page 5.

[Finn]

I could have responded to Disenchant aimed at Basalt Monolith by saying "OK, since I can run through the deck infinite times, I will set up the library such that the only cards left in my library are the remaining three copies of Monolith, plus the one Emrakul." Clearly, they can not permit this sort of thing.


Why not? Is there an actual reason or is this, like most of Wizards' supporters rationales, a "clearly" that presupposes total agreement and thus doesn't have any actual underlying basis.It is the kind of "clearly" that is irrefutably and demonstrably apparent. You can say "I loop ten quintillion times", but you can not say "I loop until event 'x'" because looping until an event (such as until cards A, B, and C are in your graveyard in a particular sequence) presupposes that there are no external factors (such as the opponent having a fast effect) that can affect the outcome. Wizards "clearly" can not permit this sort of shortcut because it removes the potential for the opponent to use this fast effect.

There. Hopefully, it is now even clear to you.

[TheInfamousBearAssassin]

It is the kind of "clearly" that is irrefutably and demonstrably apparent. You can say "I loop ten quintillion times", but you can not say "I loop until event 'x'" because looping until an event (such as until cards A, B, and C are in your graveyard in a particular sequence) presupposes that there are no external factors (such as the opponent having a fast effect) that can affect the outcome. Wizards "clearly" can not permit this sort of shortcut because it removes the potential for the opponent to use this fast effect.

There. Hopefully, it is now even clear to you.

MaximumC already brought this up and anomie-p already addressed it:


(When you hit MaximumC's stated situation, you would logically end up having to play out, as he noted, or call it a draw, both of which are reasons to not allow such loops, but neither of which logically lead to what you are suggesting)

[MaximumC]

It is the kind of "clearly" that is irrefutably and demonstrably apparent.

Okay, now we've gone from "clearly" to "irrefutably and demonstrably apparent." Well I guess you're right then!

Look, the problem is that your post doesn't actually distinguish "looping ten quntillion times" from "loop until event x." You complain that proposing a loop until an event occurs "presupposes that there are no external factors (such as the opponent having a fast effect*." Sure. And so does proposing to execute a loop 10000 times. The shortcut rules specifically anticipate this and specifically give your opponent a chance to react no matter what kind of shortcut you are proposing. When you propose a shortcut, your opponent gets a chance to say, "Okay, begin your loop. However, when I get priority at exactly this point in this iteration, I will retain priority and act. Let's go to that point." This is true whether you're looping a definite or indefinite number of times.

The problem for the shortcut rules does not lurk in trying to "trick" your opponent. It lurks in the fact that each iteration of the loop is indeterminate. So, in the example I came up with (to prove YOUR point!) the opponent wants to retain priority to act once a certain state is reached during your indeterminate loop - Emrakul on top after shuffling. You want to execute until Emrakul is on the bottom after shuffling. Since it's impossible to know in advance whether Emrakul reaches the top or the bottom first during the shortcut, the shortcut procedure implodes. You can't determine whether the opponent's requested stop point will ever occur. So, it would seem like you can't use a shortcut at all.

So what do you do with your indeterminate loop? Well, you could try to just play it out. Except, here's where the IPG steps in with the slow play rule regarding loops. Allowing you to play out your indeterminate loop would stop the whole tournament. Time expires, but turns have not ended because turns are not passing. Opponent would be foolish to concede because you have not yet won and might not win until you are sick of trying and stop. Chaos ensues, cars burn, judges go insane. To plug the gap, we have a slow play rule to just prevent this from happening.

Do you see what I'm saying, my man? It's not "clear," it's not "irrefutably and demonstrably apparent." It is, however, perfectly logical once we break down each part and actually read the Rules... instead of just explaining how your opponents were not familiar enough with shortcut procedure to stop you when you played the deck. And I have to thank you for showing up in the thread, because it was your comment that got us thinking along the right lines to noodle through the rules and figure this out!

* = This term hasn't been used in, like, a decade or more. It's more correct simply to say that the opponent may decline to pass priority at some point and take any actions that he or she may wish at that point.

[rufus]

The finite number set, as it turns out, contains numbers that are really really big. So big, in fact, as to replicate infinite for all observable purposes.

If that's true, then what is the smallest finite number that is big enough to replicate infinite for all observable purposes?

Let's say that I have indefinite Goblin Bangchuckers activations available against a Life.dec player who's got a win on me and is at googolplex life, so I put some stupidly large number of bangchucker activations on the stack (that is to say, some finite number that is "infinite for all observable purposes"). Assuming that my opponent has no outs other than luck, should he be forced to concede, or can he insist I flip coins until time expires? (Since the chance he dies is 'effectively' 1, I'm guessing your answer is yes.)

Now, let's say that, because I'm silly I put googolplex / 2 + googol activations on the stack instead. Now, my opponent is almost certain to survive. Do you think he should be forced to concede in this scenario? He's almost certain to survive, so it would be silly to force a concession.

Now, assuming you want to be able to force a concession in the first scenario, and not in the second, how do you decide what the maximum survival probability that forces a concession is?

[TheInfamousBearAssassin]

@MaximumC: Again, there's no reason we can't use a shortcut in this scenario in all events where the opponent does not have responses, and if they want to force the other player to play it out without such, rule that they are the ones guilty of stalling.

@rufus: Actually in a video I linked to recently, Richard Garfield himself describes a theoretical game where two players simultaneously reveal any number of fingers, where the goal is to have the most fingers. Generally this would end in a 5 v 5 tie, unless I guess someone is Count Rugen. The point is that this is not an interesting game.

In all events where someone has the ability to generate any sized number, if it butts up against an existing static number of problem of odds, rather than asking the person to do calculations it makes sense to simply shortcut the process to the presumption that the player will generate a number much larger than the opposing one, where "much larger" in this case means, "Of a size to be proportionately infinite for all human purposes."

[TheInfamousBearAssassin]

I feel like this point has to be hammered home, do people understand that for all statistical and observable purposes a human being could have, Googolplex is an infinite number? If your odds of something are one in Googolplex then it is basically you wasting everyone's time to pretend you have an actual chance, because you don't.

Big numbers are big. Really, really big. You just won't believe how vastly, mind-bogglingly big they are. I mean, you may think a million of something is a lot, but that's just peanuts to Graham's Number.

[rufus]

...
In all events where someone has the ability to generate any sized number, if it butts up against an existing static number of problem of odds, rather than asking the person to do calculations it makes sense to simply shortcut the process to the presumption that the player will generate a number much larger than the opposing one, where "much larger" in this case means, "Of a size to be proportionately infinite for all human purposes."

So you believe that there's a sufficiently large number of bangchucker activations to force a concession from a life.dec player with, say, 10^100 life?

[TheInfamousBearAssassin]

So you believe that there's a sufficiently large number of bangchucker activations to force a concession from a life.dec player with, say, 10^100 life?

I mean, pretty easily. We could standardize a shortcut if you want, and say, in any case where some large number n has to be overcome, the number generated will be a power tower of n to the expontent of n to the exponent of n... to the exponent of n, where the number of stacked exponents is itself equal to n. Only the pettiest of pedants could possibly be bent out of shape out of presuming certainty in that scenario, unless they just didn't grok the whole concept of numbers-that-are-big, or statistics, to start with.

[TheInfamousBearAssassin]

I mean, let's be clear:

You do not spend your life worrying about spontaneously turning into solid gold. I'm not going to ask if you do, you just don't. No one does.

Therefore, if you are claiming to be worried about something that is vastly less probable than your turning spontaneously at this second into solid gold, you either don't understand the premise, or you are putting on a pretense for some random or petty, pedantic reason.

[rufus]

I mean, pretty easily. We could standardize a shortcut if you want, and say, in any case where some large number n has to be overcome, ...

I don't understand what you mean.
Let's try a simpler case: How many Goblin Bangchucker activations do I need to put on the stack to be 'certain' that I win (at least) 1 of the coin flips?

[DragoFireheart]

I mean, let's be clear:

You do not spend your life worrying about spontaneously turning into solid gold. I'm not going to ask if you do, you just don't. No one does.


Such a bigot. I have to take supplements for my Midas genetic defect. Have you ever had your dick turn into solid gold? It's not pleasant. How about being more considerate next time instead of being a giant jackass?

[TheInfamousBearAssassin]

I don't understand what you mean.
Let's try a simpler case: How many Goblin Bangchucker activations do I need to put on the stack to be 'certain' that I win (at least) 1 of the coin flips?

Your focus on a specific marginal number that serves as such a boundary leads me to believe that you do not really grok the concepts we are talking about here.

What is such a number?

Something well under 9↑↑↑9.

[TheInfamousBearAssassin]

Such a bigot. I have to take supplements for my Midas genetic defect. Have you ever had your dick turn into solid gold? It's not pleasant. How about being more considerate next time instead of being a giant jackass?

I apologize for my lack of tact.

[Dice_Box]

No, it's you who aren't listening. I am saying that the odds are not 1%, or 1 in 3. The odds are 1, or as close to it as to be indistinguishable from the perspective of any organic or electric brain capable of doing calculations we have access to or are likely to over the next millenium.

I understand this, but you don't have any more than 40 min to do it. So I will ask you to play it out. Given all the time you need your going to be fine, but in a comp you don't have it and that's what matters here.

You can't say "Given 3 trys and I 'Should' have made this work". You need to be able to say "I can do this first time, every time" or I will make you play it out. Simple really.

[TheInfamousBearAssassin]

I understand this, but you don't have any more than 40 min to do it. So I will ask you to play it out.

Two scenarios here:

1) You actually don't understand this and that's why you're asking to play it out.

or,

2) You are just taking advantage of the current rulings which you acknowledge to be stupid and pedantic.


Given all the time you need your going to be fine, but in a comp you don't have it and that's what matters here.

Only because of an arbitrary declaration. You could say the same thing about any infinite combo.

[TsumiBand]

It doesn't matter if something is 'functionally infinite' or 'actually infinite' because Magic does not deal with infinities. A googolplex is a googolplex. 1 == 1. 1 != .999, and .999... cannot be introduced.

I understand that the argument is for corner-cases where Magic should allow for it, but understand that I do think it is (a) a distraction from the game (at best), and (b) not a great precedent to establish for cases when approaching or achieving infinity in the game lead to a breakdown or a need for higher math to enter into the picture.

If you allow for 'functionally infinite' in certain areas, such as calculating the likelihood of the 4H deck getting there against the opponent, then there are other situations where it can be arguably applied. There's a large difference between a 2^(10^100)/2^(10^100) Omnath and an ∞/∞ Omnath, precisely because it can lead to affect many things. Why worry about life totals after infinite-Omnath is StPed? What happens when it is blocked by an opponent's ∞/∞ Serra Avatar? These are concepts which shouldn't be introduced to the game, because it turns the game into infinitesimal calculus or set theory, or anything but players playing Magic.

In addition to this, like I already stated -- it breaks the 50 mins + 5 turns rule on a technicality, because the combo player never needs to pass priority in this instance. Wherein many other decks are beholden to this rule and it is a legitimate method for forcing a draw, the combo deck need only demonstrate its potential to win on a theoretically Very Large Number of iterations, where as the fundamental mechanics of the other decks cannot even say "You're unable to prevent my victory. I can do this in 6 turns, but we only have 5 -- I should be the winner anyway."

There's also no opportunity for a post-mortem examination of any of the iterations. That is frustrating, not just because you're only saying "At step 0, you're alive, but at step N, you're dead", but there is no opportunity for a player to review steps 1 through N and say, "I really should have taken [this action] when the game-state was verifiably [such]." Perhaps they require a specific pile of cards as well, or they want to respond uniquely at the time of their choosing, and how do you even begin to quantify whether or not that happens prior to your proposed shortcut to the win?

The rules rely on predictable game-states at all points during the loop, and you can only declare one -- the one where you're the winner. The inability to establish any step in the loop besides that one is the problem. We can't guarantee how many 'wrong stacks' occur before the 'right one', and so we cannot determine if or when the opponent could have responded at all. Or for that matter, given a different kind of combo, what if there's a whiff rate? Forget about this deck for a minute - we can't just say "only Four Horsemen gets to pull off this trick". How many shady combos decks are allowed to exist if we use hazy concepts like "infinity-ish" to get to a win?

[Dice_Box]

Can you tell me you can, given 40 min, kill me with it? No. Then you have your answer o the if I will concede right there.

I will concede given the right situation. High tide had just Glimpsed for 15 cards? I will give you that one. DnT has the board locked? Yep, you can have that too. Horseman cannot guarantee a kill? Sorry, that's not my fault. Play it out.

[MaximumC]

Two scenarios here:

1) You actually don't understand this and that's why you're asking to play it out.

or,

2) You are just taking advantage of the current rulings which you acknowledge to be stupid and pedantic.



Only because of an arbitrary declaration. You could say the same thing about any infinite combo.

I've been staying out of this because the last thing I need is a lecture on quantum probability by a dude who has a cartoon horse as his avatar. Even so...

He's being neither thick nor arbitrary and is not taking advantage of anything. The Comprehensive Rules have a section on shortcuts. It permits you to shortcut loops you can repeat an infinite number of times using a certain procedure. Your opponent cannot stop you from executing such a loop, but can specify a place in the loop to pause while he retains priority to do something.

The shortcut rules do not work for indefinite loops, even those that are certain to win given unlimited time, because if your opponent tries to stop the loop the game implodes.

Since you can't shortcut, you have to do it. You have no choice in the matter. (And the IPG says no you cant do that)

[TheInfamousBearAssassin]

It doesn't matter if something is 'functionally infinite' or 'actually infinite' because Magic does not deal with infinities. A googolplex is a googolplex. 1 == 1. 1 != .999, and .999... cannot be introduced.

I understand that the argument is for corner-cases where Magic should allow for it, but understand that I do think it is (a) a distraction from the game (at best), and (b) not a great precedent to establish for cases when approaching or achieving infinity in the game lead to a breakdown or a need for higher math to enter into the picture.

If you allow for 'functionally infinite' in certain areas, such as calculating the likelihood of the 4H deck getting there against the opponent, then there are other situations where it can be arguably applied. There's a large difference between a 2^(10^100)/2^(10^100) Omnath and an ∞/∞ Omnath, precisely because it can lead to affect many things. Why worry about life totals after infinite-Omnath is StPed? What happens when it is blocked by an opponent's ∞/∞ Serra Avatar? These are concepts which shouldn't be introduced to the game, because it turns the game into infinitesimal calculus or set theory, or anything but players playing Magic.

In addition to this, like I already stated -- it breaks the 50 mins + 5 turns rule on a technicality, because the combo player never needs to pass priority in this instance. Wherein many other decks are beholden to this rule and it is a legitimate method for forcing a draw, the combo deck need only demonstrate its potential to win on a theoretically Very Large Number of iterations, where as the fundamental mechanics of the other decks cannot even say "You're unable to prevent my victory. I can do this in 6 turns, but we only have 5 -- I should be the winner anyway."

There's also no opportunity for a post-mortem examination of any of the iterations. That is frustrating, not just because you're only saying "At step 0, you're alive, but at step N, you're dead", but there is no opportunity for a player to review steps 1 through N and say, "I really should have taken [this action] when the game-state was verifiably [such]." Perhaps they require a specific pile of cards as well, or they want to respond uniquely at the time of their choosing, and how do you even begin to quantify whether or not that happens prior to your proposed shortcut to the win?

The rules rely on predictable game-states at all points during the loop, and you can only declare one -- the one where you're the winner. The inability to establish any step in the loop besides that one is the problem. We can't guarantee how many 'wrong stacks' occur before the 'right one', and so we cannot determine if or when the opponent could have responded at all. Or for that matter, given a different kind of combo, what if there's a whiff rate? Forget about this deck for a minute - we can't just say "only Four Horsemen gets to pull off this trick". How many shady combos decks are allowed to exist if we use hazy concepts like "infinity-ish" to get to a win?


Okay, I'm a fucking history major, isn't Magic mostly stem nerds, why do I have to keep explaining this shit.

The numbers that I am saying are functionally infinite are not actually infinite, hence the modifier. Like, the rules of the game don't prohibit you naming something extremely large like a Googolplex, or 9↑↑↑9, or even Graham's Number. All of these numbers basically replicate infinity for all practical purposes, but are all legal in the game of Magic.

Nor do they require higher math particularly. In fact part of the point of them being functionally infinite is that you can't actually build a computer large enough to calculate them, it would require more matter than exists in the Universe.

I suppose it's possible that someone could face an opponent with say, Graham's Number of life and try to deal a Googolplex of damage and not understand that that doesn't came anywhere close to impacting their opponent's life total. But that's easy to deal with, as I've already said: just say that in such a scenario (overcoming another large number,) the arbitrarily large number named in response is arbitrarily larger, where that is defined as the first number, n, let's say n↑↑↑n↑↑↑n. There, we are now essentially infinite relative to the number n, assuming n was greater than 1.

Saying that the game of Magic allows any finite but no infinite numbers is basically dick-wagging semantics.

[TheInfamousBearAssassin]

Can you tell me you can, given 40 min, kill me with it? No.

Yes.

That's what you're not understanding:

The answer is yes.

The odds against it are in this scenario vastly less than the odds that reality will still exist in 40 minutes (let alone you,) and you have no problem taking that as a premise in your question.

So yes.

[TheInfamousBearAssassin]

I've been staying out of this because the last thing I need is a lecture on quantum probability by a dude who has a cartoon horse as his avatar. Even so...

He's being neither thick nor arbitrary and is not taking advantage of anything. The Comprehensive Rules have a section on shortcuts. It permits you to shortcut loops you can repeat an infinite number of times using a certain procedure. Your opponent cannot stop you from executing such a loop, but can specify a place in the loop to pause while he retains priority to do something.

The shortcut rules do not work for indefinite loops, even those that are certain to win given unlimited time, because if your opponent tries to stop the loop the game implodes.

Since you can't shortcut, you have to do it. You have no choice in the matter. (And the IPG says no you cant do that)

But they do work if your opponent has no response to the loop, so it's nit-picking to argue that in the case where no response exists we can't proceed to shortcut it.

The game doesn't implode, it's an incredibly simple patch, you're just being obtuse. Yes, that's because the rules are semantic and obtuse to start with, but there's no need to compound that.

[Dzra]

Like so many threads, I don't understand the point of this one. It does not matter if you will combo when given infinite iterations. It does not matter that as a finite number of iterations grows larger, it approaches statistical infinity. Rounds are finite. When two players sit there staring at each other for fifteen minutes with no difference in the board state then that seems like the very definition of slow play; I don't care that you have a statistically high chance to do something. We've sat here for upwards of a quarter of the round not doing something. Players only have so long to play and if one player is consuming time by durdling around, even if he ends up winning the game in the end, then that is not fair to the other player.

[TheInfamousBearAssassin]

Like so many threads, I don't understand the point of this one. It does not matter if you will combo when given infinite iterations. It does not matter that as a finite number of iterations grows larger, it approaches statistical infinity. Rounds are finite. When two players sit there staring at each other for fifteen minutes with no difference in the board state then that seems like the very definition of slow play; I don't care that you have a statistically high chance to do something. We've sat here for upwards of a quarter of the round not doing something. Players only have so long to play and if one player is consuming time by durdling around, even if he ends up winning the game in the end, then that is not fair to the other player.

This is the plainest jackassery since the entire point is that it makes no sense to not shortcut such a combo if the opponent has no ability to respond.

[Dice_Box]

Yes.

That's what you're not understanding:

The answer is yes.

The odds against it are in this scenario vastly less than the odds that reality will still exist in 40 minutes (let alone you,) and you have no problem taking that as a premise in your question.

So yes.

No, you can't. You can not say 100% of the time given an hour you can make it work. You can say your have a good chance, but you can't say you can guarantee it. You can't say given 100 goes you can guarantee it. You can say the likelihood is extremely high, but you can not guarantee it. That's why you have to play it out.

I am done here.

[TheInfamousBearAssassin]

No, you can't. You can not say 100% of the time given an hour you can make it work. You can say your have a good chance, but you can't say you can guarantee it. You can't say given 100 goes you can guarantee it. You can say the likelihood is extremely high, but you can not guarantee it. That's why you have to play it out.

I am done here.

You are probably struggling to imagine numbers greater than 100, which, again, peanuts, but it's not really necessary for you to do so. The very fact that you can't imagine a googol of flips, let alone a googoplex, is part of the point. The odds are functionally infinite and hence the probability is 1. So yes, I can guarantee it. If I can guarantee the continued existence of myself and a coin and reality for the space of a googolplex flips, even if we could shortcut so that it takes a fraction of a second somehow, then it would be a trifling reassurance to say that I can also guarantee at least 20 iterations of heads resulting; compared to the much, much more improbably events that I and the coin and reality and the laws therein will continue to exist in a stable form for another marginal fraction of a second.

[Dzra]

This is the plainest jackassery since the entire point is that it makes no sense to not shortcut such a combo if the opponent has no ability to respond.

Where in the rules does it say that I'm under any obligation to allow my opponent to shortcut a combo or loop?

[Bed Decks Palyer]

Ok, nice thread, now back on topic.

So how good was the Four Horsemen? (And please, no another shit about high numbers and probabilities ans Slow Play and such.) Did anybody won anything with it? There was athread buried somewhere on this site how someone went 4:1 with Lich, but I don't remember reading about the 4HM anywhere else than the "Legacy Deck Chart".

[TheInfamousBearAssassin]

Where in the rules does it say that I'm under any obligation to allow my opponent to shortcut a combo or loop?

Floor rule 4.2: Tournament Shortcuts


A tournament shortcut is an action taken by players to skip parts of the technical play sequence without explicitly
announcing them. Tournament shortcuts are essential for the smooth play of a game, as they allow players to play
in a clear fashion without getting bogged down in the minutia of the rules. Most tournament shortcuts involve
skipping one or more priority passes to the mutual understanding of all players; if a player wishes to demonstrate
or use a new tournament shortcut entailing any number of priority passes, he or she must be clear where the game
state will end up as part of the request.

A player may interrupt a tournament shortcut by explaining how he or she is deviating from it or at which point in
the middle he or she wishes to take an action. A player may interrupt their own shortcut in this manner. A player
is not allowed to use a previously undeclared tournament shortcut, or to modify an in-use tournament shortcut
without announcing the modification, in order to create ambiguity in the game.

A player may not request priority and take no action with it. If a player decides he or she does not wish to do
anything, the request is nullified and priority is returned to the player that originally had it.

[TheInfamousBearAssassin]

Ok, nice thread, now back on topic.

So how good was the Four Horsemen?

It wasn't.

[Dzra]

So how good was the Four Horsemen?

I lost a game against Four Horsemen once because I had no idea what their combo did. I was holding FoW + Pierce and let them resolve both combo pieces. I won the next game rather easily. I suppose it'd be a tough matchup if you don't have access to discard, counters, or graveyard hate.


Most tournament shortcuts involve skipping one or more priority passes to the mutual understanding of all players; if a player wishes to demonstrate or use a new tournament shortcut entailing any number of priority passes, he or she must be clear where the game state will end up as part of the request.

First off, I am still not under an obligation to allow your shortcut. Second, saying that this applies to the Four Horsemen is akin to saying "I cast Ad Nauseam and then use Tendrils to kill you. Don't worry about it, my deck is full of low costing rituals and tutors, I'm over 90% favored to kill you right now."

[Final Fortune]

90% favorite to win =/ 100% favorite to win as iterations approach infinity, if you didn't graduate from high school math yet then you really just need to stop embarassing yourself and accept that banning this deck, and Worldgorger Dragon in addition to that, is just a shitty ruling of convenience.

[MaximumC]

First off, I am still not under an obligation to allow your shortcut. Second, saying that this applies to the Four Horsemen is akin to saying "I cast Ad Nauseam and then use Tendrils to kill you. Don't worry about it, my deck is full of low costing rituals and tutors, I'm over 90% favored to kill you right now."

This isn't right. If you opponent is going off with kitchen finks in pod or making tokens in splintertwin, you cannot force them to move their cards around a million times. They will propose a shortcut of a definite number of repeditions and you get to stop the shortcut somewhere if you wish. You must allow proper shortcuts with loops.

Tendrils is not a loop.

Four horseman is a loop, but unlike splintertwin, it is not clear what will happen before it randomly produces the win. Because it is a loop, the horseman player cannot lose and will eventually win, he or she just can't tell you how long it will take or what states the library will go through along the way. Normal shortcut rules don't apply.

IBA is insisting that if horseman was permitted to play it out manually, horseman would win fairly quickly. Maybe? I suppose you could amend the slow play rule to kick in after so nany repetitions or at time or something, but that seems more confusing to players than the (admittedly hamhanded) way they handle it now.

IBA also suggests amending the shortcut rules to allow indefinite loop shortcuts provided the opponent has no response. That's no good either, because you can't know in advance of the game whether your shortcut will function. Worse, all your opponent has to do is ask for a definite stopping place; they don't have to show you that they have a card in hand to use.

Basically I'm thinking the current rules are the least complicated and least rewarding of rules lawyering that you could have.

[MaximumC]

90% favorite to win =/ 100% favorite to win as iterations approach infinity, if you didn't graduate from high school math yet then you really just need to stop embarassing yourself and accept that banning this deck, and Worldgorger Dragon in addition to that, is just a shitty ruling of convenience.

Nonono, dragon is not indefinite. Nothing random involved. You can shortcut to infinity mana all day long.

[TheInfamousBearAssassin]

First off, I am still not under an obligation to allow your shortcut.

Yes, you are:


716.2a At any point in the game, the player with priority may suggest a shortcut by describing a sequence of game choices, for all players, that may be legally taken based on the current game state and the predictable results of the sequence of choices. This sequence may be a non-repetitive series of choices, a loop that repeats a specified number of times, multiple loops, or nested loops, and may even cross multiple turns. It can’t include conditional actions, where the outcome of a game event determines the next action a player takes. The ending point of this sequence must be a place where a player has priority, though it need not be the player proposing the shortcut.
Example: A player controls a creature enchanted by Presence of Gond, which grants the creature the ability "{T}: Put a 1/1 green Elf Warrior creature token onto the battlefield," and another player controls Intruder Alarm, which reads, in part, "Whenever a creature enters the battlefield, untap all creatures." When the player has priority, he may suggest "I’ll create a million tokens," indicating the sequence of activating the creature’s ability, all players passing priority, letting the creature’s ability resolve and put a token onto the battlefield (which causes Intruder Alarm’s ability to trigger), Intruder Alarm’s controller putting that triggered ability on the stack, all players passing priority, Intruder Alarm’s triggered ability resolving, all players passing priority until the player proposing the shortcut has priority, and repeating that sequence 999,999 more times, ending just after the last token-creating ability resolves.

716.2b Each other player, in turn order starting after the player who suggested the shortcut, may either accept the proposed sequence, or shorten it by naming a place where he or she will make a game choice that’s different than what’s been proposed. (The player doesn’t need to specify at this time what the new choice will be.) This place becomes the new ending point of the proposed sequence.


If you can't do anything about a combo Second, saying that this applies to the Four Horsemen is akin to saying "I cast Ad Nauseam and then use Tendrils to kill you. Don't worry about it, my deck is full of low costing rituals and tutors, I'm over 90% favored to kill you right now."

I'm a little tired of this vapid and idiotic pseudo-intellectual conceit that a chance is a chance is a chance.

10% is wildly outside of the realm we are talkling about. If your odds of something are 1 over a googoplex, which is one of the smaller big numbers we could name here, then they are worse than the odds that hinge on any material interaction with the real world you ever have.

And yet judges make decisions all the time where the probability of error is vastly vastly larger, where there's maybe a .000001% chance that someone drew an extra card by accident, or a 1.03735 x 10^-987 chance that the extra land they registered spontaneously transformed into a fifth foil Japanese Brainstorm.

You are a human being who has to interact with the real world on a daily basis, so I refuse to dignify your pedantic nerd-whinge that there's a 1 x 10 ^-( 9↑↑↑9 9↑↑↑9 9↑↑↑9) chance of the combo fizzling any further.

[TheInfamousBearAssassin]

@MaximumC: You don't really have to reveal that you have a given way of interacting. Legalize the shortcut, and you can make an opponent play it out, who can then call a judge who can look at your hand to see whether you're just stalling.

This is already pretty much the case for how it works with similar sequences anyway. If there were no uncertainty involve in the combo and you just had to mill your entire library, and your opponent was making you activate the elaborate combo each time for each card in your deck, it would clearly be stalling unless they actually had a way to interrupt the combo.

[Dzra]

Tendrils is not a loop.

Four horseman is a loop, but unlike splintertwin, it is not clear what will happen before it randomly produces the win. Because it is a loop, the horseman player cannot lose and will eventually win, he or she just can't tell you how long it will take or what states the library will go through along the way. Normal shortcut rules don't apply.

While I may admit to the use of hyperbole, this is basically my point. In TES, your opponent casting Ad Nauseam is more than likely going to produce the result of you getting Tendrils'd to death. The point is that there is no way for either player to know what will happen between casting Ad Nauseam and you dying, therefore attempting to force the use of a shortcut is disingenuous and misleading. Similarly, from the moment Four Horsemen starts milling until the time that they can cast their Dread Return, no one can tell you what will happen. They can tell you what they'd like to happen, but that is not the same thing as telling you what will happen, making a shortcut equally disingenuous.

[Dice_Box]

a loop that repeats a specified number of times

And this is where the rules and the argument differ. You can not define a set number. If you could, no one would disagree with you.

[TheInfamousBearAssassin]

While I may admit to the use of hyperbole, this is basically my point. In TES, your opponent casting Ad Nauseam is more than likely going to produce the result of you getting Tendrils'd to death. The point is that there is no way for either player to know what will happen between casting Ad Nauseam and you dying, therefore attempting to force the use of a shortcut is disingenuous and misleading. Similarly, from the moment Four Horsemen starts milling until the time that they can cast their Dread Return, no one can tell you what will happen. They can tell you what they'd like to happen, but that is not the same thing as telling you what will happen, making a shortcut equally disingenuous.

This isn't hyperbole. This is simply wrong and your not understanding how math and probability work.

[TheInfamousBearAssassin]

And this is where the rules and the argument differ. You can not define a set number. If you could, no one would disagree with you.

We can define an upper bound with certainty so this is really just pedantic nerd dick-wrangling.

[Dzra]

716.2a At any point in the game, the player with priority may suggest a shortcut by describing a sequence of game choices... It can’t include conditional actions, where the outcome of a game event determines the next action a player takes.

I'm fairly certain that milling and then checking the GY to determine if you have the correct cards in there to combo qualifies as an outcome of a game event that determines the next action a player takes.

Come to think of it, why not just play the combo without the Emrakul? Are we simply concerned that we'll have too many combo pieces stranded in our hand? No one would have a problem with the deck if it didn't attempt to continually loop Emrakul shuffles.

[Dice_Box]

You can say given limitless time you can do it, but you can't say "I can do this one million times and guarantee I can kill you." You can say that the probably is so high that I have more chance of getting stuck by lightning than you have of bricking, but the rules don't care. It wants a sure fire guarantee and you can't say that you have one. Because regardless of how many times you say you do it, there is always, regardless of how small it may be, a chance you will brick. So the best way to fix the issue in that situation is to play it out.

Your not dealing in absolutes, your dealing with probably. Even if I do happen to get struck by lightning as I leave the building, your still not able to list any number where there is zero chance of a brick. Only a a miniscule chance. While that chance is not worth talking about because it's so low, the rules don't care and that is where the math, along with the debate, falls flat.

The rules are not going to be changed, and we are on a lovely little carousel here, so anyone got anything new to add or are we to keep going round and round while getting exactly nowhere?

Edit.
I understand that, given the conditions you will eventually make it work 100% of the time, I know this. That's not the point though. You can't say when. That's the issue

[Infinitium]

As others have mentioned, the problem with the Horsemen loop isn't that the outcome cannot be worded to be functionally deterministic ("I will continue to iterate this loop until such a time that Sharuum, Blasting Station and Dread Return, in that order, are the sole cards in my graveyard"), but that the intermediate gamestates cannot be predicted. I don't see how this can be alleviated in a competitive game with chance and hidden information as game elements.

EDIT: Dzra - Emrakul isn't a requirement, but was a definite answer to singular gravehate (bar Extirpate). Simply stack the offending effect, mill until Emrakul, vacate the graveyard, allow the card/ability in question to resolve and start over. There's nothing keeping you from playing the same deck with an Angel/Azami/Researcher kill or whatever.

[TheInfamousBearAssassin]

I'm fairly certain that milling and then checking the GY to determine if you have the correct cards in there to combo qualifies as an outcome of a game event that determines the next action a player takes.

No one has disputed that that rule exists; you claimed that you didn't have to accept shortcuts. I was demonstrating that you were wrong. Whether the existing ruling that neuters Four Horsemen and potentially other decks should be modified is the topic of discussion. Welcome to the conversation.


You can say given limitless time you can do it, but you can't say "I can do this one million times and guarantee I can kill you." You can say that the probably is so high that I have more chance of getting stuck by lightning than you have of bricking, but the rules don't care. It wants a sure fire guarantee and you can't say that you have one. Because regardless of how many times you say you so do it, there is always, regardless of how small it may be, a chance you will brick. So the best way to fix the issue in that situation is to play it out.

Even if the mechanics of the game allowed for no uncertainty in the abstract, I could still claim that I should make you play out an elaborate combo-loop because there is a chance that you could have a heart attack and die, leaving me the victor by default, so why should I scoop early?

This would, on the odds, be a much more reasonable position to take than saying, "Well, but you could fizzle after a googol of iterations," but it would still be ruled as stalling and I should be warned for it.

It turns out that Magic tournaments take place in the material universe and as such, everything about them is subject to much, much more uncertainty than what we are discussing here in terms of realistic probability of fizzling over very large numbers of loops.


Your not dealing in absolutes, your dealing with probably.

Only a Sith deals in absolutes.

But in seriousness, do you not see how stupid and pedantic this complaint is? We're not talking about an actual probability, we are talking about a certainty that is more functionally closer to certainty than, say, you actually existing as a person and not a temporary hallucination.


Even if I do happen to get struck by lightning

You're still not thinking big enough, nowhere near. Instead think; if you happen to spontaneously turn into a five hundred pound banana. That's a little bit closer to the order we're talking about, although still not really anywhere close.


as I leave the building, your still not able to list any number where there is zero chance of a brick. Only a a miniscule chance. While that chance is not worth talking about because it's so low, the rules don't care and that is where the math, along with the debate, falls flat.

The rules are not going to be changed, and we are on a lovely little carousel here, so anyone got anything new to add or are we to keep going round and round while getting exactly nowhere?

"The rules are right because that's what the rules are" is a pretty open abdication.

[TheInfamousBearAssassin]

As others have mentioned, the problem with the Horsemen loop isn't that the outcome cannot be worded to be functionally deterministic ("I will continue to iterate this loop until such a time that Sharuum, Blasting Station and Dread Return, in that order, are the sole cards in my graveyard"), but that the intermediate gamestates cannot be predicted. I don't see how this can be alleviated in a competitive game with chance and hidden information as game elements.

This is a much less stupid complaint than the above, but I'll repeat my address to it here:


You don't really have to reveal that you have a given way of interacting. Legalize the shortcut, and you can make an opponent play it out, who can then call a judge who can look at your hand to see whether you're just stalling.

This is already pretty much the case for how it works with similar sequences anyway. If there were no uncertainty involve in the combo and you just had to mill your entire library, and your opponent was making you activate the elaborate combo each time for each card in your deck, it would clearly be stalling unless they actually had a way to interrupt the combo.

[Dzra]

No one has disputed that that rule exists; you claimed that you didn't have to accept shortcuts. I was demonstrating that you were wrong. Whether the existing ruling that neuters Four Horsemen and potentially other decks should be modified is the topic of discussion. Welcome to the conversation.

The game defines what an acceptable short cut is. In addition to what the game defines as "shortcuts," players can shortcut things themselves by simply conceding. I can combo off with Show and Tell, put Omniscience into play, show my opponent a Cunning Wish and a Force of Will, and ask for a concession. My opponent is under no obligation to scoop, but he is going to lose. Four Horsemen mills itself a bit and says "I'm going to continue to do this until I have all my combo pieces in the graveyard, then I will Dread Return, and kill you. There is no way you can avoid this outcome, will you concede?" Likewise, no one is being forced to shortcut or forced to concede. It is still entirely valid to say "show me the win."

[Dice_Box]

Only a Sith deals in absolutes.
Ok, your an ass, but that was funny.


"The rules are right because that's what the rules are" is a pretty open abdication.
Yes it is, but at the end of the day, the judge makes a call based on those rules so they kind of matter somewhat.

[Infinitium]

@IBA: Maybe an example to illustrate potential problems with that line of thought. Your proposed loop shortcut rules apply. The opponent is tapped out, has a Sensei's Divining Top in play and Ravenous Trap somewhere in his library, but no information about his top cards. The Horsemen player starts comboing, but reveals secondary win conditions (say Tarmogoyfs) that the oppnent wants to remove by topdecking a potential trap with the SDT in response to a Emrakul trigger with all Tarmos and any combo piece in the gy.

Now what? It's not like a judge can waltz up and effectively give the players information about the (hidden) top card of the opponents library to determine whether the 4HM player can legitimately shortcut, and even if he would the 4HM player can't prove that the "iterate until only-3-combo-cards-in-specific-order" will occur before a gamestate where all 4 tarmos and a combo piece would be vulnerable in the gy with Emrakul on the stack, nor is it possible to predict what that gamestate would look like (other cards milled etc).

[Julian23]

Exactly what Infinitium said.

We're all wasting our time if we're not discussing the implications of the proposed rules changes as they are the sole reason why people like me oppose them.

[TheInfamousBearAssassin]

@IBA: Maybe an example to illustrate potential problems with that line of thought. Your proposed loop shortcut rules apply. The opponent is tapped out, has a Sensei's Divining Top in play and Ravenous Trap somewhere in his library, but no information about his top cards. The Horsemen player starts comboing, but reveals secondary win conditions (say Tarmogoyfs) that the oppnent wants to remove by topdecking a potential trap with the SDT in response to a Emrakul trigger with all Tarmos and any combo piece in the gy.

Now what? It's not like a judge can waltz up and effectively give the players information about the (hidden) top card of the opponents library to determine whether the 4HM player can legitimately shortcut, and even if he would the 4HM player can't prove that the "iterate until only-3-combo-cards-in-specific-order" will occur before a gamestate where all 4 tarmos and a combo piece would be vulnerable in the gy with Emrakul on the stack, nor is it possible to predict what that gamestate would look like (other cards milled etc).

Well, let me ask you a different scenario. The combo player, instead of attempting a kill with Sharuum etc., is running the on-demand mill combo with Narcomoebas, Emrakul, and a Goblin Bombardment in play.

Now there is no real variance and order doesn't matter; they can loop infinitely and sac at instant speed before the Emrakul trigger will resolve anyway. This would fall under the normal purview of predictable infinite combos, then.

However, the above scenario (Top, possible Ravenous Trap) still would apply, right? So, how do we resolve it?

Well, as near as I can tell, the combo player has abdicated interrupting their own combo; it is the other player who is proposing truncating the combo. I'm actually not sure how this scenario would be resolved by a current judge, but I imagine that the player demanding that the loop be played out would now have to demonstrate that they could interrupt the combo- that is to say, that they are going to draw into Trap with the Top. If it's not the top card of their library the combo proceeds into its infinite arbitrarily large mechanism, while if it is the milling combo must play out until the player who controls Ravenous Trap wishes to play it or is dead, I suppose.

But when we're dealing with graveyard and library order it really doesn't matter in this case whether the loop is fully predictable or not, the exact same scenario can arise.

[Zombie]

A solution to Horseman-style loops, maybe? The problem is potential interaction, right? What happens before, and after, and so on, and the possible time taken for it? And people with such loops would want to win with them, given it really makes sense (googolplex Reveillark-Entity-Body Double iterations are valid, for example). So, what about (with less-than-perfect legalese):

So, say a loop meets certain conditions, like:
1. The player proposing the loop cannot lose from the tournament shortcut.
2. The loop must, given infinite random iteration, result in a condition that ends the game without a shred of doubt.
3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.
4. If the opponent can take no Meaningful Actions anymore, the original loop proceeds randomly, that is, towards the opponent's certain annihilation.

Or something of the sort. Basically, twist the randomness in the victim's favour so he is ensured the chance to interact in whatever manner he wishes to, and if he has no answers, kill him then and only then. Seems like a reasonable solution to me? Maybe force the initiating player to take a couple times through the loop first as the demonstration to ensure self-mill loops show the whole deck?

[cdr]

The "Four Horsemen" combo never worked under tournament rules and it wasn't "banned", because outside of uninformed judges it was never allowed. A clarification was issued at some point though that yes, indeed, it didn't work.

(This will be the first and last time I read or reply to this thread.)

[MaximumC]

The "Four Horsemen" combo never worked under tournament rules and it wasn't "banned", because outside of uninformed judges it was never allowed. A clarification was issued at some point though that yes, indeed, it didn't work.

(This will be the first and last time I read or reply to this thread.)

"Never worked?" I don't think this is correct. As I recall -- and I can't find a cite on this right now, so maybe I am wrong -- the slow play rule related to loops was invented specifically for Four Horsemen.

The combo "never worked" in the sense that you could never shortcut it because it implodes the shortcut rules, but I don't think it was always slow play to repeatedly try to win through random chance.

Also, I think you can still play Four Horseman today. You can't shortcut the process, of course, but you can keep repeating your loop as long as you meaningfully change the game state in each iteration. For example, by putting a narco into play. But the moment your loop comes around to an identical game state, you're done.

[Bed Decks Palyer]

"Never worked?" I don't think this is correct. As I recall -- and I can't find a cite on this right now, so maybe I am wrong -- the slow play rule related to loops was invented specifically for Four Horsemen.

The combo "never worked" in the sense that you could never shortcut it because it implodes the shortcut rules, but I don't think it was always slow play to repeatedly try to win through random chance.

Also, I think you can still play Four Horseman today. You can't shortcut the process, of course, but you can keep repeating your loop as long as you meaningfully change the game state in each iteration. For example, by putting a narco into play. But the moment your loop comes around to an identical game state, you're done.

Would something like Greater Gargadon help to keep the deck legal? No sane judge may argue "guy that's some slow play, you're not moving further" if you'll eat a Moeba to GGD every time Emrakul tries to reshuffle the grave.

[Julian23]

"Never worked?" I don't think this is correct. As I recall -- and I can't find a cite on this right now, so maybe I am wrong -- the slow play rule related to loops was invented specifically for Four Horsemen.

As you already suggested, you're actually wrong on this. There has never been a rules change that in any way affected Four Horsemen. As cdr mentioned, because there was some confusion on how the handle the combo ruleswise, an official clarification was released.

[Julian23]

Well, let me ask you a different scenario. The combo player, instead of attempting a kill with Sharuum etc., is running the on-demand mill combo with Narcomoebas, Emrakul, and a Goblin Bombardment in play.
[...]
However, the above scenario (Top, possible Ravenous Trap) still would apply, right? So, how do we resolve it?

That's the question "we" (the guys that think changing the rules in the way "you" are suggesting would hurt the game) need answered.

[TheInfamousBearAssassin]

That's the question "we" (the guys that think changing the rules in the way "you" are suggesting would hurt the game) need answered.

Why? Under current rules it's completely legal to shortcut that loop.


Also don't use quotes for emphasis, it just makes you look like you're being sarcastic.

[MaximumC]

Would something like Greater Gargadon help to keep the deck legal? No sane judge may argue "guy that's some slow play, you're not moving further" if you'll eat a Moeba to GGD every time Emrakul tries to reshuffle the grave.

Why not? This shoulkd work just fine, the same way blasting station hits their life total each shiffle.

But the moment you hit emrakul before a narco on a shuffle, you nust do something else that changes the board state before you can try again.

[TheInfamousBearAssassin]

Why not? This shoulkd work just fine, the same way blasting station hits their life total each shiffle.

But the moment you hit emrakul before a narco on a shuffle, you nust do something else that changes the board state before you can try again.

Why? You can just respond to Emrakul hitting the bin by milling the rest of your library, saccing the moebas, and then recycling the yard by letting Emrakul's trigger resolve.

There isn't any actual uncertainty in the loop unless an opponent is responding somehow.

[cherub_daemon]

Well, let me ask you a different scenario. The combo player, instead of attempting a kill with Sharuum etc., is running the on-demand mill combo with Narcomoebas, Emrakul, and a Goblin Bombardment in play.

Now there is no real variance and order doesn't matter; they can loop infinitely and sac at instant speed before the Emrakul trigger will resolve anyway. This would fall under the normal purview of predictable infinite combos, then.

However, the above scenario (Top, possible Ravenous Trap) still would apply, right? So, how do we resolve it?

Well, as near as I can tell, the combo player has abdicated interrupting their own combo; it is the other player who is proposing truncating the combo. I'm actually not sure how this scenario would be resolved by a current judge, but I imagine that the player demanding that the loop be played out would now have to demonstrate that they could interrupt the combo- that is to say, that they are going to draw into Trap with the Top. If it's not the top card of their library the combo proceeds into its infinite arbitrarily large mechanism, while if it is the milling combo must play out until the player who controls Ravenous Trap wishes to play it or is dead, I suppose.

But when we're dealing with graveyard and library order it really doesn't matter in this case whether the loop is fully predictable or not, the exact same scenario can arise.

The Bombardment and 4Horse versions are both problematic from a response standpoint. Let's say that in the 4Horse version, I say that I would like to respond only when Sharuum has hit the graveyard, but no other relevant cards. Without playing it out, we can't know for sure whether I'll die before my response condition is met.

Similarly (and more concretely), against the Bombardment version, Life.dec may want to Ravenous Trap only when all 4 Narcomoebas are in the yard before the reshuffle, rendering your deck "non-infinite". If Life hasn't gone off yet, you're going to need to play this out. (Technically, you need to play this out whether Life has gone off or not, but I'll accept that Life is so likely to see 4 Narcomoebas in the yard before dying in that case that I won't revisit the 0.999... vs 1 nonsense.) It is not certain that you will advance the game state on every trip around the library, since you could hit Emrakul before hitting any Narcomoebas; the slow play argument still applies to the Bombardment list. Admittedly, it's more of a jerk move to call slow play on a loop that will advance the state 80% of the time than one which advances 25% of the time, but we're not talking about "transmogrify into banana"-type probability, so you have to account for it.

The trouble is that these combos rely on a shortcut which may be permissible depending on the opponent; stating that shuffle loop combos are not okay seems fine to me.



3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.


I like this as a theoretical solution, but it doesn't really solve the slow play problem; you're proposing to let an opponent durdle around with an unfamiliar library looking for advantageous orderings. If the opponent takes too long, is the opponent guilty of slow play? What is a reasonable amount of time for a task like this?

[TsumiBand]

Yes, the arguments for "dying from a heart attack" or "Goblin Lackey transforming into solid gold" are rubbish when we're trying to talk about an indeterminate game state and nothing else.

The fact that there is uncertainty enough to have no idea whether or not an answer could ever be played to the deck, as its piles are random and you can't guarantee that Condition A is met before Condition B, no matter how small the odds for either one to actually occur, should be enough to show anyone that even the "so-called/functional infinities" part of the argument is a red herring, and honestly a mistake on my part for engaging in it*.

What matters most is that there are game states where a player could respond, and you cannot guarantee which one will arise first; the one where the opponent is dead, or the one where the opponent responds to your combo and prevents you from winning. I believe on the merit of that reason alone, being forced to play it out is the only option.

And there are also sub-optimal game states which the opponent may try to take advantage of. It might be the best play to wait until Condition A is true before casting Answer Card, but if Condition B arises and the player feels it is in their best interest to go for it at that point because they are unwilling to let the loop iterate again and risk it being THE winning game-state, then they should have that ability. You can't just math those things away, and you cannot always guarantee that you can preclude the opponent from having an answer.

* ...I have to say though, I still do not believe there is room in Magic for infinite concepts; the loop/shortcut rules were colloquially called "the Infinity Rules" for a reason. This is an old version of the Comp Rules, but I think it does underline what their intention truly was -- as recently as Lorwyn AFAICT, this was still part of the Comp Rules Glossary:


Infinity Rule (informal): The concept of infinity does not exist in Magic. If asked to pick a number, you must pick an actual number, regardless of how high you want that number to be. If the game enters a state where a set of actions could be repeated forever, the "Infinity Rule" handles how to break such loops. See Rule 421 of the Magic Comprehensive Rules for more detailed information.

So unless one wants to quibble over what 'actual number' means, which I interpret to mean "an explicit number that is known; an integer that is not also a variable, or is not expressed as a variable", or however you get away from the idea of framing it as "I move through X iterations where X is the non-infinite number required to smoke your ass", I believe that this does away with not just infinities but any situation where the player cannot explicitly state for a fact when the opponent is 100% dead. Not 99%, not 'functionally equivalent' to dead, fucking completely no longer in the game as a result of state-based actions giving you the shove.

[Julian23]

Why? Under current rules it's completely legal to shortcut that loop.

As it's an indermined loop, it is not. You keep sidestepping the question of how to handle the negative implications your proposed rules change brings about.

[TheInfamousBearAssassin]

As it's an indermined loop, it is not. You keep sidestepping the question of how to handle the negative implications your proposed rules change brings about.

Except it's not. It's no an indeterminate loop.

Also how to respond to the above was already outlined in one of my previous posts.


Similarly (and more concretely), against the Bombardment version, Life.dec may want to Ravenous Trap only when all 4 Narcomoebas are in the yard before the reshuffle, rendering your deck "non-infinite". If Life hasn't gone off yet, you're going to need to play this out. (Technically, you need to play this out whether Life has gone off or not, but I'll accept that Life is so likely to see 4 Narcomoebas in the yard before dying in that case that I won't revisit the 0.999... vs 1 nonsense.) It is not certain that you will advance the game state on every trip around the library, since you could hit Emrakul before hitting any Narcomoebas; the slow play argument still applies to the Bombardment list.

Holy shit no. Why do I have to keep... okay, let's make this easier.

IF YOU HIT EMRAKUL BEFORE HITTING ALL FOUR NARCOMOEBAS, YOU ALLOW THAT TRIGGER TO SIT ON THE STACK, RESPONDING TO IT BY MILLING THE REST OF YOUR LIBRARY. NARCOMOEBA TRIGGERS WILL THEN RESOLVE BEFORE THE SHUFFLE AND ALLOW YOU TO SAC THEM TO BOMBARDMENT, PUTTING THEM IN THE GRAVEYARD FOR THE EMRAKUL TRIGGER.

There is zero indeterminance here, the board state is advanced the exact same way each iteration. It is completely predictable, except insofar as responding with a possible Ravenous Trap is concerned.

We could make the sequence even easier for everyone here to understand: A player could have Vulturous Zombie attacking, and the same Mesmeric Orb + Emrakul combo. In this case their opponent is at a trillion life because of some previous combo. Now priority of other resolving abilities doesn't matter, all the player wants to do is dump a trillion-ish cards into their yard. Emrakul could even be the first card they hit every single time and it wouldn't matter, even if they couldn't keep milling in response.

This is a standard, short-cuttable combo for as clear a case as can be conceived.

Yet, the prior objection of Ravenous Trap + Top still applies, because the mechanic of the combo is still the library hitting the graveyard. So, how would you resolve that? Allow the Top player to essentially thwart a demonstrably predictable, demonstrably infinite combo on the basis that there may be an answer on top of their library?

[TheInfamousBearAssassin]

Yes, the arguments for "dying from a heart attack" or "Goblin Lackey transforming into solid gold" are rubbish when we're trying to talk about an indeterminate game state and nothing else.

The fact that there is uncertainty enough to have no idea whether or not an answer could ever be played to the deck, as its piles are random and you can't guarantee that Condition A is met before Condition B, no matter how small the odds for either one to actually occur, should be enough to show anyone that even the "so-called/functional infinities" part of the argument is a red herring, and honestly a mistake on my part for engaging in it*.

What matters most is that there are game states where a player could respond, and you cannot guarantee which one will arise first; the one where the opponent is dead, or the one where the opponent responds to your combo and prevents you from winning. I believe on the merit of that reason alone, being forced to play it out is the only option.

And there are also sub-optimal game states which the opponent may try to take advantage of. It might be the best play to wait until Condition A is true before casting Answer Card, but if Condition B arises and the player feels it is in their best interest to go for it at that point because they are unwilling to let the loop iterate again and risk it being THE winning game-state, then they should have that ability. You can't just math those things away, and you cannot always guarantee that you can preclude the opponent from having an answer.

So you admit you're wrong, but you think someone else has saved you with a different argument you want to jump on.


* ...I have to say though, I still do not believe there is room in Magic for infinite concepts; the loop/shortcut rules were colloquially called "the Infinity Rules" for a reason. This is an old version of the Comp Rules, but I think it does underline what their intention truly was -- as recently as Lorwyn AFAICT, this was still part of the Comp Rules Glossary:



So unless one wants to quibble over what 'actual number' means, which I interpret to mean "an explicit number that is known; an integer that is not also a variable, or is not expressed as a variable", or however you get away from the idea of framing it as "I move through X iterations where X is the non-infinite number required to smoke your ass", I believe that this does away with not just infinities but any situation where the player cannot explicitly state for a fact when the opponent is 100% dead. Not 99%, not 'functionally equivalent' to dead, fucking completely no longer in the game as a result of state-based actions giving you the shove.

Oh, no, you couldn't leave well enough alone.

And still seemingly incapable of grasping numbers smaller than .01 I guess.

[TsumiBand]

You're such a fucking bonehead.


So you admit you're wrong, but you think someone else has saved you with a different argument you want to jump on.



Oh, no, you couldn't leave well enough alone.

And still seemingly incapable of grasping numbers smaller than .01 I guess.

So it's really cool that in your mind, someone admitting that a given argument was a red herring is a fallacy. Well that's unfortunate, because that's what grown-ups do; they see their errors and they turn around.

I'm not incorrect about infinity in Magic, there's precedent going back for quite some time about how they attempt to keep it out of the game. That's less relevant in this case than the fact that you cannot put a quantifiable integer value on the number of cycles it takes to 4H-club your opponent. Moreover you cannot prove that an unfavorable game-state will occur between then and now, because of the randomization of critical game zones.

I've been demanding a provable game-state since the beginning, so my argument hasn't changed really, I've just de-emphasized the bit about whether or not 'infinity' can be applied in a game that tries to keep it out, in particular where loops are involved. You have to cite a number of iterations where you get to your win, and it has to be a non-variable, actual integer. You also have to submit to the idea that a "bad" game-state could occur BEFORE yours does, in which case the opponent has a window to respond. You can't do those things, and you can't play it out, so what's left? Like there is genuinely no way for you to lay out every possible iteration, and you do have to because your opponent is able to respond to things, AND you have to prove they are dead before the game-state (a) reaches X iterations (b) reaches the "bad" game-state where either the opponent can respond or your fizzle -- fizzle implying any given combo, not just 4H's madness.

You cannot do those things, and you cannot satisfactorily tell anyone here why those arguments are irrelevant or do not apply, so just what the fuck are you doing trying to make it personal, clown car? Cheesy Petes.

[cherub_daemon]

Except it's not. It's no an indeterminate loop.

(big font redacted)

We could make the sequence even easier for everyone here to understand: A player could have Vulturous Zombie attacking, and the same Mesmeric Orb + Emrakul combo. In this case their opponent is at a trillion life because of some previous combo. Now priority of other resolving abilities doesn't matter, all the player wants to do is dump a trillion-ish cards into their yard. Emrakul could even be the first card they hit every single time and it wouldn't matter, even if they couldn't keep milling in response.

This is a standard, short-cuttable combo for as clear a case as can be conceived.

Yet, the prior objection of Ravenous Trap + Top still applies, because the mechanic of the combo is still the library hitting the graveyard. So, how would you resolve that? Allow the Top player to essentially thwart a demonstrably predictable, demonstrably infinite combo on the basis that there may be an answer on top of their library?

Sorry about the (wrong) concrete example; you posted that timing while I was writing mine, and I forgot to check responses before I posted. Apologies.

And yes, I would allow a possible response to "thwart" the combo in the following way: if there is a game state which may arise as a result of an infinite loop that I have a potential response to, I should be allowed to ask you to demonstrate that it won't happen before you kill me. If you cannot, you need to play it out, and not shortcut.

However, the order-independent lists are different than the order-dependent lists in that there is no chance for you to make a lap around the library without advancing the game state, so there's no way to call them out for slow play. No such guarantee exists for the order-dependent lists, so the slow play issue stands for those lists.

(Edit: clarified language.)

[Dzra]

If anyone is actually interested in using a Mesmeric Orb combo then they can just play the Laboratory Maniac wincon. There's really not a good reason to argue for the Emrakul looping crap; the current rules are fine as is and I don't know of any worthwhile deck that is actually hurt by this ruling. For whatever reason, IBA takes personal interest in arguing for unpopular and unnecessary causes; I don't believe there is any changing his mind in this.

[TheInfamousBearAssassin]

Sorry about the (wrong) concrete example; you posted that timing while I was writing mine, and I forgot to check responses before I posted. Apologies.

And yes, I would allow a possible response to "thwart" the combo in the following way: if there is a game state which may arise as a result of an infinite loop that I have a potential response to, I should be allowed to ask you to demonstrate that it won't happen before you kill me. If you cannot, you need to play it out, and not shortcut.

However, the Vulturous Zombie version IS different than the others in that there is no chance for you to make a lap around the library without advancing the game state, so there's no way that they can call that version out for slow play. No such guarantee exists for the order-dependent lists, so the slow play issue stands for those lists.

Well, but that's really begging the question. The point is that that combination involves a sequence of plays that, even under the most fiddly of objections, are demonstrably infinite, predictable, and concrete, but it falls victim to the exact same complaint that was lodged as the reason Four Horsemen couldn't be allowed to be shortcutted, despite clearly being a shortcuttable combo under current rules.

[Zombie]

A solution to Horseman-style loops, maybe? The problem is potential interaction, right? What happens before, and after, and so on, and the possible time taken for it? And people with such loops would want to win with them, given it really makes sense (googolplex Reveillark-Entity-Body Double iterations are valid, for example). So, what about (with less-than-perfect legalese):

So, say a loop meets certain conditions, like:
1. The player proposing the loop cannot lose from the tournament shortcut.
2. The loop must, given infinite random iteration, result in a condition that ends the game without a shred of doubt.
3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.
4. If the opponent can take no Meaningful Actions anymore, the original loop proceeds randomly, that is, towards the opponent's certain annihilation.

Or something of the sort. Basically, twist the randomness in the victim's favour so he is ensured the chance to interact in whatever manner he wishes to, and if he has no answers, kill him then and only then. Seems like a reasonable solution to me? Maybe force the initiating player to take a couple times through the loop first as the demonstration to ensure self-mill loops show the whole deck?

[Dzra]

Well, but that's really begging the question. The point is that that combination involves a sequence of plays that, even under the most fiddly of objections, are demonstrably infinite, predictable, and concrete, but it falls victim to the exact same complaint that was lodged as the reason Four Horsemen couldn't be allowed to be shortcutted, despite clearly being a shortcuttable combo under current rules.

What do you mean? Shortcutting the act of milling your library into your graveyard is acceptable precisely because it is finite. Mesmeric Orb makes the whole library go to the yard, Narcomeobas come into play, Dread Return Angel of Glory's Rise, then win. There is no infinite loop; there is no "wait until condition X occurs after an indeterminable number of iterations". Or did I misunderstand what you meant?

[TheInfamousBearAssassin]

I was responding to the wrong post, fixed.

[Julian23]

There is zero indeterminance here, the board state is advanced the exact same way each iteration. It is completely predictable, except insofar as responding with a possible Ravenous Trap is concerned.

You keep sidestepping this issue. For several pages, this has been the sole issue of concern. This is the reason we oppose your proposed rules change.

Also, don't make a fool of yourself by explaining the technical details of a combo everyone clearly understands; especially after size=72'ing it.

You seem to be paranoid that people might not understand the combo and therefore oppose what you are suggesting , when in fact we are talking about something completly else.

[TheInfamousBearAssassin]

You keep sidestepping this issue. For several pages, this has been the sole issue of concern. This is the reason we oppose your proposed rules change.

Also, don't make a fool of yourself by explaining the technical details of a combo everyone clearly understands. You seem to be paranoid that people might oppose what you are suggesting because they might not be understanding the combo, when in fact we are talking about something completly else.

1) I addressed how I would deal with this situation several posts back.

2) As I've pointed out several times now, this same issue can arise with a purely determinant combo.


You clearly don't understand what we're talking about or basic English as I have to keep reiterating these things to you, so perhaps you should be less concerned about whether other people are making fools of themselves.

[Echelon]

Who cares. The deck isn't being played anymore nor is any other that could display similar behavior.

[Julian23]

1) I addressed how I would deal with this situation several posts back.

I'm going through all your answers in this thread right now because I'm genuinely interested in how we could solve this problem.
Ok, I'm at your SuperNova comparision right now. As you said, you feel the problem is people not understanding statistics/big numbers. As I am telling you, that's not the problem. Everyone understands how even a 0.01% chance approaches 100% when you repeat it Graham's number times. There might be people that actually don't understand this, so it makes sense you explained this but you really have move on now and stop beating a dead horse as you're really wasting your time explaining this with ever-new analogies.

Ok, you keep talking about big numbers and how people don't understand them. I read on and still you're talking about how people don't understand big numbers, 9↑↑↑9, Googoplex etc. As I said, it's nice to point this out but doesn't help with solving what we actually want to know: how to deal with the implications of your suggested rules change. Don't get me wrong, but you really need to work on your Cassandra complex with regards to people not understanding large numbers. It must be frustrating to see people draw a comparision to shortcutting "probably killing" with Ad Nauseam, but they are so far off, just don't waste your time there.

Finally, I found your first post on the real issue on page 7:
"You don't really have to reveal that you have a given way of interacting. Legalize the shortcut, and you can make an opponent play it out, who can then call a judge who can look at your hand to see whether you're just stalling.

This is already pretty much the case for how it works with similar sequences anyway. If there were no uncertainty involve in the combo and you just had to mill your entire library, and your opponent was making you activate the elaborate combo each time for each card in your deck, it would clearly be stalling unless they actually had a way to interrupt the combo. "

I recognize this as a suitable way of dealing with the implications I have questioned you about. While you (presumably) evaluate this as a fair solution though, you got to see that others like me can not accept this. The solution you are suggesting here makes one player, who is completly irresponsible for bringing the problematic combo to the table, reveal hidden information (=the fact that he actually has a way of interacting) to his opponent. This is pretty much an absoltue no-go in the way I look at fairness.


2) As I've pointed out several times now, this same issue can arise with a purely determinant combo.
And we still don't have an answer who such a situation would be treated under the current rules.



You clearly don't understand what we're talking about or basic English as I have to keep reiterating these things to you, so perhaps you should be less concerned about whether other people are making fools of themselves.

It's just you're pretty much known for discussing things way too emotionally. That makes it hard to actually talk to you as you always seem to totally freak out. :frown:

[Echelon]

Funny to see how a topic on the Four Horsemen-deck ends up being a sorts of Four Horsemen-pissing contest, iterating over and over without really advancing the discussion, waiting for Godwin's law to kick in but not knowing when it'll happen.

Not takings sides here, by the way.

[(nameless one)]

This deck must be really that good before it got "axed" for getting all this attention.

I mean 10 pages of argument.

[Julian23]

waiting for Godwin's law to kick in but not knowing when it'll happen.

Miller's Paradox: As a network evolves, the number of Nazi comparisons not forestalled by citation to Godwin's Law converges to zero.

[TsumiBand]

Does it still count as invoking Godwin if only doing so by reference? "I'm surprised no one on this thread has used the 'Nazi' word yet." Is just a pointer to Godwin, only now it's "out there".

sent from phone, don't be a dick

[TheInfamousBearAssassin]

I'm going through all your answers in this thread right now because I'm genuinely interested in how we could solve this problem.
Ok, I'm at your SuperNova comparision right now. As you said, you feel the problem is people not understanding statistics/big numbers. As I am telling you, that's not the problem. Everyone understands how even a 0.01% chance approaches 100% when you repeat it Graham's number times. There might be people that actually don't understand this, so it makes sense you explained this but you really have move on now and stop beating a dead horse as you're really wasting your time explaining this with ever-new analogies.

I'm sorry you're tired of my explaining something simple over and over to idiots that won't get it. Please, let me instead focus all my attention and energy on addressing your concerns, as I've been doing repeatedly.


Finally, I found your first post on the real issue on page 7:
"You don't really have to reveal that you have a given way of interacting. Legalize the shortcut, and you can make an opponent play it out, who can then call a judge who can look at your hand to see whether you're just stalling.

This is already pretty much the case for how it works with similar sequences anyway. If there were no uncertainty involve in the combo and you just had to mill your entire library, and your opponent was making you activate the elaborate combo each time for each card in your deck, it would clearly be stalling unless they actually had a way to interrupt the combo. "

I recognize this as a suitable way of dealing with the implications I have questioned you about. While you (presumably) evaluate this as a fair solution though, you got to see that others like me can not accept this. The solution you are suggesting here makes one player, who is completly irresponsible for bringing the problematic combo to the table, reveal hidden information (=the fact that he actually has a way of interacting) to his opponent. This is pretty much an absoltue no-go in the way I look at fairness.


Well, let me ask you a different scenario. The combo player, instead of attempting a kill with Sharuum etc., is running the on-demand mill combo with Narcomoebas, Emrakul, and a Goblin Bombardment in play.

Now there is no real variance and order doesn't matter; they can loop infinitely and sac at instant speed before the Emrakul trigger will resolve anyway. This would fall under the normal purview of predictable infinite combos, then.

However, the above scenario (Top, possible Ravenous Trap) still would apply, right? So, how do we resolve it?

Well, as near as I can tell, the combo player has abdicated interrupting their own combo; it is the other player who is proposing truncating the combo. I'm actually not sure how this scenario would be resolved by a current judge, but I imagine that the player demanding that the loop be played out would now have to demonstrate that they could interrupt the combo- that is to say, that they are going to draw into Trap with the Top. If it's not the top card of their library the combo proceeds into its [s]infinite arbitrarily large mechanism, while if it is the milling combo must play out until the player who controls Ravenous Trap wishes to play it or is dead, I suppose.

But when we're dealing with graveyard and library order it really doesn't matter in this case whether the loop is fully predictable or not, the exact same scenario can arise.



And we still don't have an answer who such a situation would be treated under the current rules.

As I've pointed out like, what, five times now? If you're correct, then this problem is not limited to indeterminate sequencing combos. I gave you an example, two examples actually, the second one (using Vulturous Zombie) even more clearly a case of an "infinite" combo that has clear and predictable iterations that affect the board state in every sequence.

If this is an issue we have to deal with it's not something bound up specifically in the subset of the rules that disallow indeterminately numbered infinite combos. And it doesn't apply to many that are, say, flipping a googol coins to kill an opponent.

This is clearly not an objection that is tied specifically to the sub-rule in question and something of a red herring ultimately.

And generally if someone clearly has a win I would think it pretty clear that the other person is stalling if they demand the combo be played out without demonstrating that they can interact with it.


It's just you're pretty much known for discussing things way too emotionally. That makes it hard to actually talk to you as you always seem to totally freak out. :frown:

Oh what gaslighting bullshit, if you're saying stupid shit I'm going to say it's stupid shit.

[Echelon]

Miller's Paradox: As a network evolves, the number of Nazi comparisons not forestalled by citation to Godwin's Law converges to zero.

I lol'd :laugh:. Alrighty, let me fix this:

Funny to see how a topic on the Four Horsemen-deck ends up being a sorts of Four Horsemen-pissing contest, iterating over and over without really advancing the discussion, waiting for Godwin's law to kick in but not knowing when it'll break through Miller's Paradox and finally produce a result.

[anomie-p]

2) As I've pointed out several times now, this same issue can arise with a purely determinant combo.

I'm not grokking how trap makes the deterministic loop into a case of 'well, now we actually have to play out all the loops or we can't determine who wins at all'. Can't the trap player decline the shortcut, let one loop play out, and top when getting priority at the emrakul trigger (after letting your opponent bring out the narcs and bombard), and either find it, or not find it? Making the comparison "In the nondeterministic loop, you have to play out all loops, in the deterministic loop, you have to play one and you either find the trap or don't"?

Plus, having to play out loops with an indeterminate number of loops is potentially much worse time-wise than having to play out a fixed number of loops.

Am I missing something here? is there a graveyard state in the deterministic loop where you wouldn't want to trap if you could, once they've brought in the narcs and bombarded?

[rufus]

Your focus on a specific marginal number that serves as such a boundary leads me to believe that you do not really grok the concepts we are talking about here.

"The fault, dear Brutus, lies not in our stars, but in ourselves, that we are underlings."
-- Julius Caesar (I, ii, 140-141)

Funny that. In my experience, people who can't address specific concerns or work through specific examples aren't thinking their proposal through clearly enough.



...
Something well under 9↑↑↑9.

Assuming - for the moment - that there's a finite number of activations large enough to guarantee at least one winning coin flip, let's call that number N. Then for any M>N, M activations must also guarantee a winning coin flip. That means there's some minimum number of activations necessary to guarantee at least one win.

That means that there's some n with the properties that for all m>n, m activations guarantees at least one win, and
n activations does not guarantee at least one win. Specifically, n activations does not guarantee at least one win, but n+1 activations does.

So, instead of activating bangchuckers 9↑↑↑9 times, I could just activate it n+1 times and still be certain of a win.
Now, what happens to my guaranteed win if I play out the first coin flip?
I have a 50% chance of not winning the first coin flip, and the remaining n activations aren't enough to guarantee a win on their own.
n+1 flips cannot simultaneously be enough to guarantee at least one win, and not guarantee at least one win.
Therefore the initial assumption - that such an N exists - is false.

[MaximumC]

"The fault, dear Brutus, lies not in our stars, but in ourselves, that we are underlings."
-- Julius Caesar (I, ii, 140-141)

Funny that. In my experience, people who can't address specific concerns or work through specific examples aren't thinking their proposal through clearly enough.



Assuming - for the moment - that there's a finite number of activations large enough to guarantee at least one winning coin flip, let's call that number N. Then for any M>N, M activations must also guarantee a winning coin flip. That means there's some minimum number of activations necessary to guarantee at least one win.

That means that there's some n with the properties that for all m>n, m activations guarantees at least one win, and
n activations does not guarantee at least one win. Specifically, n activations does not guarantee at least one win, but n+1 activations does.

So, instead of activating bangchuckers 9↑↑↑9 times, I could just activate it n+1 times and still be certain of a win.
Now, what happens to my guaranteed win if I play out the first coin flip?
I have a 50% chance of not winning the first coin flip, and the remaining n activations aren't enough to guarantee a win on their own.
n+1 flips cannot simultaneously be enough to guarantee at least one win, and not guarantee at least one win.
Therefore the initial assumption - that such an N exists - is false.

Why are we even having this debate? Is IBA really arguing that looping a random event X finite times should be considered a win if the probability of winning on one of those X terms is very close to, but not equal to, 100%? I thought we were talking about the fact that the limit of such looping at infinity was equal to 100%. I'd be comfortable with a rule allowing shortcuts to infinity, essentially, in the case of looping sequences that have 100% probability of success as their limit. I'm not comfortable with any claim that 99.99999% should be considered 100% for Magic purposes. There's still a chance you could lose, so hey, let's live the dream.

The real question, in my mind, is whether you can change the shortcut and slow play rules to handle non-deterministic loops. That is, loops that end up in random states each go around. Originally I thought, sure, just prove that the limit of the random sequence is a win -- but then we had the problem of how an opponent determines a stopping place in the shortcut. If that's not possible, then you need to figure out a way to let people play out non-deterministic loops without stopping the whole tournament.

Finally, if none of that is possible, then the question is how best to set up your deck so that you can change game states at will to prevent a judge from hitting you with slow play as you try to execute your indeterminate loop. Narcomebias and Greater Gargadon were great examples when it comes to Four Horseman. Any other ideas?

[rufus]

...
Finally, if none of that is possible, then the question is how best to set up your deck so that you can change game states at will to prevent a judge from hitting you with slow play as you try to execute your indeterminate loop. Narcomebias and Greater Gargadon were great examples when it comes to Four Horseman. Any other ideas?

As far as I am aware it's a judges discretion thing, which makes it a little tricky. (Otherwise, you could just put 10 mill triggers on top of the emrakul trigger, and you'd end up with a different GY state on each loop.)

Running 4xCabal Therapy (and a 4th Narco) seems like the first thing to do.
Salvage Titan can be activated from the graveyard reasonably reliably, and then recycled in with Cabal Therapy. In a pinch, he's a decent sized body.
Bridge from Below and Fatestitcher can produce extra bodies for Dread Return, and open up some extra lines of play.
Lava Dart could also be flashed back for no mana cost, and has some utility value otherwise.

Suitably refined lines of play could also help improve the odds.

[TsumiBand]

"That word. I do not think it means what you think it means."
-- Fezzig, The Princess Bride

As long as I'm wasting away my break time going back over this goofy kerfuffle (well that and looking at matrix transformations among other dumb stuff in my browser tabs), the correct line was "…you keep using that word. I do not think it means what you think it means." and it was uttered by Inigo Montoya, not the giant.

I ain't even mad, just sayin'. :D


http://www.youtube.com/watch?v=G2y8Sx4B2Sk

[The Doctor]

The other day, I was playing a board game with a friend. I got to the point that all I needed to do was roll a six to win the game. Luckily, I was on a spot on the board where if I rolled anything else, I got to pick up the die and roll again, and again and again until I got a six. I rolled and got a one. As I picked up the die to roll again, someone walked to the table and informed me that I didn’t get to roll again.
“But the spot I’m on lets me roll again.”
The person replies “but you’re on the same spot, so rolling again makes it slow play.”
“Okay. Can we just shortcut it to the point that I roll a six? Clearly you can see than I’m eventually going to roll a six.”
“Nope. You can’t tell me when exactly you’re going to roll a six. Therefore the probability isn’t exactly equal to one so you can’t shortcut it.”
“Yeah but I mean. Any reasonable person could see that I’m eventually going to get a six. If I don’t get a six, I could just roll again. The probability isn’t one, but it is infinitesimally close. The entire world is more likely to be destroyed at this exact second than for me to not eventually get a six.”
“Sorry, that’s the rules. You have to stop rolling or forfeit the game for slow play.”
I sadly passed the turn back to my opponent and he won the game on that very turn.

This is what I think of Four Horsemen being “slow play.”

[MaximumC]

The other day, I was playing a board game with a friend. I got to the point that all I needed to do was roll a six to win the game. Luckily, I was on a spot on the board where if I rolled anything else, I got to pick up the die and roll again, and again and again until I got a six. I rolled and got a one. As I picked up the die to roll again, someone walked to the table and informed me that I didn’t get to roll again.
“But the spot I’m on lets me roll again.”
The person replies “but you’re on the same spot, so rolling again makes it slow play.”
“Okay. Can we just shortcut it to the point that I roll a six? Clearly you can see than I’m eventually going to roll a six.”
“Nope. You can’t tell me when exactly you’re going to roll a six. Therefore the probability isn’t exactly equal to one so you can’t shortcut it.”
“Yeah but I mean. Any reasonable person could see that I’m eventually going to get a six. If I don’t get a six, I could just roll again. The probability isn’t one, but it is infinitesimally close. The entire world is more likely to be destroyed at this exact second than for me to not eventually get a six.”
“Sorry, that’s the rules. You have to stop rolling or forfeit the game for slow play.”
I sadly passed the turn back to my opponent and he won the game on that very turn.

This is what I think of Four Horsemen being “slow play.”

I agree completely with this sentiment, but as this thread has forced me to decide that the current Rules are at least logical and reasonable. Also, I can't think of better rules!

Here's why. Let's say you've got your dice ready to go. You can execute a loop as many times as you like, and the probability of your winning the game tends towards 100% as your loop repetitions increase towards infinity.

Now, in Magic if you have a loop, you can propose a "shortcut" instead of actually executing the loop. When you go off with Splinter Twin, for example, you can ask your opponent: "I propose to shotcut this loop until I have made 5,000,000 copies of Exarch. At that point the board will be just like it is now, but with 5,000,000 exarch copies in play." Your opponent gets the option to either accept this shortcut, or to tell you that he is retaining priority at a specific place in the shortcut. "Okay, but we are stopping the shortcut when your 20th tokens enters the battlefield. I will retainer priority at that point." You then execute the shortcut to that point, change the board state accordingly, and if you are very unlucky your opponent casts Golgari Charm and you lose.

This process fails utterly with indeterminate loops. Say, "I propose to roll this dice until I hit a 6." Your opponent says, "Okay, but we are stopping the shortcut when you roll a 2." Now what? You can't determine whether you will roll your 6 or the 2 first! There's no way of knowing what the board state will be at the time the shortcut ends. Shortcut rules do not work.

The alternative would be to play it out. But, imagine if rolling a dice in your example took five minutes to do. The time you spend trying to roll a 6 quickly adds up. The judges need some way of preventing you from holding up the entire tournament. They do this by setting an arbitrary cap on the number of times you can repeat the roll. In this case, they choose 1 time. Seems like a logical place to start.

Does that make sense?

[Zombie]

I agree completely with this sentiment, but as this thread has forced me to decide that the current Rules are at least logical and reasonable. Also, I can't think of better rules!

Here's why. Let's say you've got your dice ready to go. You can execute a loop as many times as you like, and the probability of your winning the game tends towards 100% as your loop repetitions increase towards infinity.

Now, in Magic if you have a loop, you can propose a "shortcut" instead of actually executing the loop. When you go off with Splinter Twin, for example, you can ask your opponent: "I propose to shotcut this loop until I have made 5,000,000 copies of Exarch. At that point the board will be just like it is now, but with 5,000,000 exarch copies in play." Your opponent gets the option to either accept this shortcut, or to tell you that he is retaining priority at a specific place in the shortcut. "Okay, but we are stopping the shortcut when your 20th tokens enters the battlefield. I will retainer priority at that point." You then execute the shortcut to that point, change the board state accordingly, and if you are very unlucky your opponent casts Golgari Charm and you lose.

This process fails utterly with indeterminate loops. Say, "I propose to roll this dice until I hit a 6." Your opponent says, "Okay, but we are stopping the shortcut when you roll a 2." Now what? You can't determine whether you will roll your 6 or the 2 first! There's no way of knowing what the board state will be at the time the shortcut ends. Shortcut rules do not work.

The alternative would be to play it out. But, imagine if rolling a dice in your example took five minutes to do. The time you spend trying to roll a 6 quickly adds up. The judges need some way of preventing you from holding up the entire tournament. They do this by setting an arbitrary cap on the number of times you can repeat the roll. In this case, they choose 1 time. Seems like a logical place to start.

Does that make sense?

Yeah I wish people would actually comment on the solution I've posted twice already. Even to say it's unworkable horseshit. But no, bickering is more important I guess.


A solution to Horseman-style loops, maybe? The problem is potential interaction, right? What happens before, and after, and so on, and the possible time taken for it? And people with such loops would want to win with them, given it really makes sense (googolplex Reveillark-Entity-Body Double iterations are valid, for example). So, what about (with less-than-perfect legalese):

So, say a loop meets certain conditions, like:
1. The player proposing the loop cannot lose from the tournament shortcut.
2. The loop must, given infinite random iteration, result in a condition that ends the game without a shred of doubt.
3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.
4. If the opponent can take no Meaningful Actions anymore, the original loop proceeds randomly, that is, towards the opponent's certain annihilation.

Or something of the sort. Basically, twist the randomness in the victim's favour so he is ensured the chance to interact in whatever manner he wishes to, and if he has no answers, kill him then and only then. Seems like a reasonable solution to me? Maybe force the initiating player to take a couple times through the loop first as the demonstration to ensure self-mill loops show the whole deck?

[rufus]

Yeah I wish people would actually comment on the solution I've posted twice already. Even to say it's unworkable horseshit. But no, bickering is more important I guess.



1. The player proposing the loop cannot lose from the tournament shortcut.
2. The loop must, given infinite random iteration, result in a condition that ends the game without a shred of doubt.
...


This isn't something that can be effectively adjudicated in the general case.
It also doesn't seem to apply to the four horsemen kill con, since the actual game ender will reliably deal damage.

I guess step 3 addresses the 'priority passing' issue reasonably well.



... it was uttered by Inigo Montoya, not the giant.

Inconceivable!

[cherub_daemon]

Yeah I wish people would actually comment on the solution I've posted twice already. Even to say it's unworkable horseshit. But no, bickering is more important I guess.
...
3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.


I did, once, but it got lost, because there was a bunch of other crap in my post. I'll repeat it:

I like [Step 3] as a theoretical solution, but it doesn't really solve the slow play problem because you're proposing to let an opponent durdle around with an unfamiliar library looking for advantageous orderings. If the opponent takes too long, is the opponent guilty of slow play? What is a reasonable amount of time for a task like this?

[ubernostrum]

The alternative would be to play it out. But, imagine if rolling a dice in your example took five minutes to do. The time you spend trying to roll a 6 quickly adds up. The judges need some way of preventing you from holding up the entire tournament. They do this by setting an arbitrary cap on the number of times you can repeat the roll. In this case, they choose 1 time. Seems like a logical place to start.

Mostly this, though there is no fixed number of iterations of an indeterminate loop mentioned in the IPG -- Slow Play is by definition a subjective infraction which exists in the eyes of the judge who issues it.

But yes, the way Slow Play works is entirely about ensuring the tournament progresses at a reasonable pace, plus not needing everyone present to be well-versed enough in statistics to conclude how likely it is that an indeterminate loop produces the desired result, or that it produces the desired result before an opponent's desired interruption point, etc. etc. The wording of Slow Play for loops ensures that either you can give it to us in exact terms, or you open yourself to Slow Play.

(also, if you want to know why Slow Play is subjective, again it boils down to practicalities: if there is a fixed number of iterations, or a fixed time limit for making a decision, then it is absolutely guaranteed, not by statistics but by human nature, that someone, and probably several someones, will show up with a deliberate plan of indeterminately looping N-1 times or of taking T-1 seconds to make every single one of their choices, or both)

(also also: reiterating that registering the exact 75 for Four Horsemen is not illegal, and playing those cards in a real match is not illegal, but the way you play with them may earn you Slow Play, which is why it's not a great idea. Same is true of people who can't use Top or other grindy/time-consuming/decision-heavy strategies in a timely manner)

[MaximumC]

Yeah I wish people would actually comment on the solution I've posted twice already. Even to say it's unworkable horseshit. But no, bickering is more important I guess.

I'm sorry we have not given you more attention :)

Your proposed rule:



1. The player proposing the loop cannot lose from the tournament shortcut.
2. The loop must, given infinite random iteration, result in a condition that ends the game without a shred of doubt.
3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.
4. If the opponent can take no Meaningful Actions anymore, the original loop proceeds randomly, that is, towards the opponent's certain annihilation.


Well, let's see. I do not share the other poster's complaint that step 2 is asking too much of people. It seems to me that most tournament-viable indeterminate loops make it pretty clear that you will win. For example, take the Four Horsemen. How much math do you need to know to see that: (1) It is a perfect loop, returning to an identical board state each go-around EXCEPT THAT (2) your opponent may have less life; and (3) your library will be in a different order. Since you can always just try again if you whiff, it seems pretty clear you will win.

It's possible that there are situations that are not so clear. The more complicated loop involved that strange combo someone proposed awhile back where you and your opponent are passing turns with empty libraries and not drawing cards, you're casting Bolt every turn, they are casting Hero's Reunion each turn, and you are also casting Stitch in Time each turn. Since you're adding and subtracting life total, that loop would require some heavy duty math to show whether you are certain to win or not. (I tend to think you are because your opponent has a finite life total and can lose if you hit a string of good flips, but can never stop the process by simply gaining life. It's like how a casino always wins because it can keep playing the game with you. But, that's based just on my intuition and I don't know how to prove it.)

How frequently are you going to encounter such bizarre loops in practice, though? How frequently will you see an indeterminate loop that does a "random walk" towards victory? Or, similarly, how frequently will a loop "randomly walk" between you winning and losing?

So I think applying Rule 2 in practice is totally viable, but if there were concerns, then maybe you could modify it this way:

2. The loop must be of such a nature that each iteration has some nonzero chance of advancing the board state in a linear fashion towards victory, and/or winning the game, and no chance of moving the board state away from victory and/or losing the game. (Example: Each time you repeat the loop, you have a nonzero chance of assembling a combo or reducing your opponent's life total).

This would invalidate complicated proofs dealing with "random walk" which would be fascinating but not intuitive for people involved.

As for your Rule 3, are you suggesting that the shortcut rules simply get amended to say, "The opponent may elect to stop the shortcut at any other configuration that is a possible outcome at any point during the loop. If he or she does, the game state is moved to that configuration with the opponent retaining priority." ?

If so, that's interesting. It cuts the Gordian knot by simply stating, by fiat, that your opponent's selected state WILL occur first if you want to shortcut an indeterminate loop. So in playing the Four Horseman, you'd basically hand your library over to your opponent who can then stack the deck anyway they want. You probably need some provision saying that once your opponent has done this, if they then pass priority, you get to complete the loop to your preferred configuration, though, to avoid them doing this repeatedly.

Those fixes should work, but it's doing an AWFUL lot of work on the rules for essentially one deck. And, I wonder if there would be unintended consequences...?

[HammerAndSickled]

Didn't read the whole damn thread, but in my personal opinion the definition of "slow play" should include "slow" (taking a long time temporally) and "play" (the act of making decisions or manipulating game pieces with a goal in mind). I think as long as the player plays quickly and physically moves the fucking cards around in a timely fashion instead of tanking in response to every Emrakul trigger it should be allowed. If you try it a dozen times and (variance working against you) fail to meet the conditions to kill you should be able to call a judge, explain the loop to your judge and your opponent and get the W. Or if you have a particularly sadistic rules philosophy about this, make him pass the turn or whatever. But the current rule that doesn't let him even try to do it manually is bogus.

If you want a rule on the books forbidding nondeterministic combos then write one, but calling it "slow play" is utter nonsense. Slow play should apply to the guy who takes 30 seconds per top activation every turn and the guy who tanks constantly when he's demonstrably locked into no actions and the player who always "has a response" in his decks without countermagic. It shouldn't apply to the guy who knows his deck and plays fast but ran afoul of a wrongly-interpreted rule.

[Mr. Froggy]

Didn't read the whole damn thread, but in my personal opinion the definition of "slow play" should include "slow" (taking a long time temporally) and "play" (the act of making decisions or manipulating game pieces with a goal in mind). I think as long as the player plays quickly and physically moves the fucking cards around in a timely fashion instead of tanking in response to every Emrakul trigger it should be allowed. If you try it a dozen times and (variance working against you) fail to meet the conditions to kill you should be able to call a judge, explain the loop to your judge and your opponent and get the W. Or if you have a particularly sadistic rules philosophy about this, make him pass the turn or whatever. But the current rule that doesn't let him even try to do it manually is bogus.

If you want a rule on the books forbidding nondeterministic combos then write one, but calling it "slow play" is utter nonsense. Slow play should apply to the guy who takes 30 seconds per top activation every turn and the guy who tanks constantly when he's demonstrably locked into no actions and the player who always "has a response" in his decks without countermagic. It shouldn't apply to the guy who knows his deck and plays fast but ran afoul of a wrongly-interpreted rule.

I agree with this.

The Top analogy is perfect in my opinion.

[MaximumC]

Hammer, Froggy, you really SHOULD read the whole thread. It's actually very interesting, and you'll find people like me who do agree wholeheartedly with your sentiment yet come to understand the logic behind why the Rules are the way they are.



If you want a rule on the books forbidding nondeterministic combos then write one, but calling it "slow play" is utter nonsense.

They did write such a rule. Two of them: the Rules governing Shortcuts and the IPG rule defining "Slow Play." Shortcut rules do not work in the case of nondeterministic combos, because you can't tell which condition will occur first: the win or the situation in which the opponent wants to stop and retain priority to respond (will you roll 6 or 2 first?). There's a Rule in the IPG then that specifically prohibits you from "continuing" a loop that does not change the board state, which prevents you from playing it out without the shortcut. (Absent game state shenanigans.)

[TsumiBand]

Topping is a totally different ball of wax, though.

Really, there are a dozen minor effects that just sort of accumulated over the years in Legacy that involve deck stacking, shuffling, and other actions which give the player the possibility to make a different decision at different points. Anyone recall when Rebels were regarded as "too shuffly" by Magic R&D to ever return? They gummed up the works by permitting multiple shuffles over many or all of a player's turns. That's not a tenth as bad as current Legacy decklists; fetchlands, Brainstorms, Tops, tutors, so on and so forth.

Just because both lines of play - "taking way too long to SDT" and "executing a loop over indeterminate turns with indeterminate game-states between the demonstrated loop and the desired endpoint" - can be viewed as Slow Play, I would not consider them to be the same manner of Slow Play. The problem with SDT is not that the game is left in an unpredictable grey area, it's that the opponent isn't doing anything but staring at their hand like it was written in Urdu. The issue with 4H is that only the first and last iteration are defined; the iterations in between are dismissed as something that can be mentally 'played around' or declared mathematically insignificant.

Take for example a lousy combo; I control Sedris the Traitor King and have Deadeye Navigator and Palinchron in my graveyard, along with five untapped lands that each miraculously tap for any combination of :r:, :b:, and :u:. I play the Unearth cost on Palinchron, which puts its ETB ability on the stack; I float :u::b: before passing. The ability resolves and my lands untap. I tap three land to Unearth Deadeye Navigator; its Soulbond triggers on entry, and I pair it with Palinchron. I use my floating :u::b: to blink Palinchron - Unearth's replacement effect doesn't apply, because it's being exiled instead of put anywhere else. So it blinks (Unearth can no longer trigger at end of turn in a meaningful way for this creature), and its ability stacks again, as does Deadeye Navigator's. I untap my lands. I soulbond DEN to Palinchron. Combo illustrated. I say I'll do it again 400,000 times. If successful, I end up with 1.2 million-ish mana, in equal distributions :u:, :b:, and :r:. At any point during this combo, my opponent can single out the game-state and act whenever they're able to; my opponent can definitively say, "I have a response to Deadeye Navigator's Soulbond trigger after the 22,915th iteration" and we can sit and know what that is -- more importantly, we can get to a point where we can play that out.

4H relies too heavily on generating random gamestates in order to get to an eventual win, so we don't have that granular control over where we end up when we're declaring responses. And we have no idea how many iterations it will even take to get to a particular gamestate, or if one will occur before the other. The opponent can't know if they can respond; or if they can respond, what they are responding to and what will actually be in the graveyard. Maybe they can force the opponent to exile their Emrakuls; maybe they can try to force them to draw a card when they can't before they get their graveyard back into the library; they can't know what the right response will be, because they can't know which cards are where and which zone to even attack, or if they will even have the chance to do so before the combo actually works.

[PirateKing]

It was a cute deck though.

[Mr. Froggy]

I did read most it, all I got was "if N is bigger than Y then Blastoise uses Water Cannon". I hate math.

Now when someone says using Top and takes 30 secs, now I understand what's going on.

[HammerAndSickled]

I couldn't care less about advanced mathematics applied to children's card games and I find that that guy with the pony avatar never says anything that's not inflammatory so I find your description of the thread as "interesting" doesn't really apply to me given a cursory examination.

Regardless, I think your point about the rules doesn't apply, given that neither of those rules forbid non-deterministic loops on their own, nor was that the intent. The shortcut rules don't allow you to shortcut them, that makes sense, because it's hard or impossible to write rules that would let you do that while being fair. The IPG slow play rule exists to stop players from stalling out of their own actions and acts in deference to tournament time concerns. Neither rule specifically stops the combo, and it's only the convergence of the two rules (plus the ruling of some judge) that really killed the deck.

i'm not arguing that we allow the loop to be shortcutted, because I have no idea how that would even be logically done. I also don't advocate abolishing slow-play rules, because a lot of dumb motherfuckers legitimately play this game slowly and deserve to be hurried up. I'm arguing that the convergence of these rules shouldn't stop someone who plays quickly and efficiently from manually attempting a combo. They're not attempting an illegal shortcut, they're certainly not "slow playing" by any meaningful definition of the term. We all agree that if you get lucky and hit the right combination of cards each loop, the deck is perfectly legal and no one can stop you, right? If you mill three Narcos, then Emrakul, then Sharuum+DR+Blasting Station, then your combo is fine and you win fair and square. For that matter, the whole deck works fine with a flash Leyline in play or a Goblin Bombardment in play, and never runs afoul of any rules. So the only issue is that a judge decided that if you miss on one iteration you have to stop because you're not demonstrably "progressing the game state" (a nebulous term in and of itself) and that's what I take issue with. I feel the line should be drawn somewhere, because it's obviously stalling if you keep trying and failing, but it's ludicrous and draconian to draw the line at "1 chance, 25% odds, go!"

[MaximumC]

that guy with the pony avatar never says anything that's not inflammatory

Well, yeah, but you just need to ignore the fedora-waving he does and skip to the relevant discussion. The math is really interesting, or so I think.



Regardless, I think your point about the rules doesn't apply, given that neither of those rules forbid non-deterministic loops on their own, nor was that the intent. The shortcut rules don't allow you to shortcut them, that makes sense, because it's hard or impossible to write rules that would let you do that while being fair. The IPG slow play rule exists to stop players from stalling out of their own actions and acts in deference to tournament time concerns. Neither rule specifically stops the combo, and it's only the convergence of the two rules (plus the ruling of some judge) that really killed the deck.

Well... sort of? As you say, the two rules cut off your two ways of comboing out with Four Horseman -- you cannot shortcut and you cannot just "do it." As to the second option, just doing it, the slow play rule very specifically does stop the deck from working.

You seem to suggest that the slow play rule is some vague statement that people should play quickly. That's not what it says at all. It says:

"It is also slow play if a player continues to execute a loop without being able to provide an exact number of
iterations and the expected resulting game state." IPG 3.3

I mean, it's black and white. Unless you carefully change the game state each time you execute your loop (which is possible, by the way, and is why this deck can still see play) you run smack dab into the black letter law of this rule.

The slow play sentiment you're explaining is a different part of the rule, which says:

"A player takes longer than is reasonably required to complete game actions. If a judge believes a player is
intentionally playing slowly to take advantage of a time limit, the infraction is Unsporting Conduct — Stalling." IPG 3.3

It's fine to say you agree with the "stalling" part of the rule but not the "loop" part, but they're both right there in the IPG.

[anomie-p]

The issue with 4H is that only the first and last iteration are defined; the iterations in between are dismissed as something that can be mentally 'played around' or declared mathematically insignificant.

There is nothing about mathematical insignificance here. You are not saying 'this will happen because the odds it won't are mathematically insignificant', you are saying 'This will happen because it is absolutely mathematically certain to happen' (under a ruleset allowing it). If I am allowed to flip my library into the yard, then shuffle, then flip again, for as long as I like, I just keep going until I hit it, it *will* happen (think about, there are a finite number of library orders and thus a finite number of graveyard states, but I can actually flip until whatever I need is there, and my requirements aren't more than 'these three cards in the yard without hitting Emrakul').

We should be well past a misunderstanding based on thinking it's a matter of dismissing a mathematically insignificant probability at this point.

As you stated later ->


The opponent can't know if they can respond; or if they can respond, what they are responding to and what will actually be in the graveyard. Maybe they can force the opponent to exile their Emrakuls; maybe they can try to force them to draw a card when they can't before they get their graveyard back into the library; they can't know what the right response will be, because they can't know which cards are where and which zone to even attack, or if they will even have the chance to do so

*That* is the problem. In MaximumC's example, under the proposed 'allow nondeterministic shortcuts when mathematically certain' rule, both the state where the player on 4H will win, and the state where exiling the top card breaks the combo, are absolutely guaranteed to happen mathematically. There is no dismissal based on 'mathematical insignificance'. The *order* they will occur in is absolutely indeterminate without playing it out, but it is a certainty that flipping the library enough times, *all* library (and thus graveyard) orders will occur.

It doesn't make sense to me to consider allowing such a shortcut in the first place unless we're talking about something that is mathematically certain.

[HammerAndSickled]

My argument is that the second part of the rule need not exist, because any situations of abuse are covered by the first part (and the related infraction of Stalling) and in any instance where it's not being abused to stall, the rule needlessly kills combos like this. Also, I disagree with your characterization that it's "black and white" because it's pretty evident to me that the game state changes each time we cycle through the library, but it's not demonstrable with physical evidence so it runs afoul of the rule?

Again, I know the deck doesn't currently work by the current rules and judging philosophies. That's never been in contention. I argue that that's not the way it SHOULD be.

[anomie-p]

Without 3.3, though, how do you stop someone from forcing a draw on time with a nonlethal infinite combo where they are performing the game actions fast enough to be considered 'reasonable'? You'd have to add *some* rule about 'if a loop is infinite but cannot lead to a win' and, I dunno, is that a box we want to open? (looping finks off viscera seer with Melira in play, you might want to loop 10 million times, and go, doesn't directly lead to a win necessarily - but if they had a game 1 win, or maybe you're in game three, and just sat there and looped it over and over and over to run time out on you *because they can* ... that would suck (maybe that's a bad example, but take a loop that has no real effect but can be done forever ...))

[Zombie]

I'm sorry we have not given you more attention :)

Your proposed rule:


A solution to Horseman-style loops, maybe? The problem is potential interaction, right? What happens before, and after, and so on, and the possible time taken for it? And people with such loops would want to win with them, given it really makes sense (googolplex Reveillark-Entity-Body Double iterations are valid, for example). So, what about (with less-than-perfect legalese):

So, say a loop meets certain conditions, like:
1. The player proposing the loop cannot lose from the tournament shortcut.
2. The loop must, given infinite random iteration, result in a condition that ends the game without a shred of doubt.
3. Now, here's the secret sauce. After this proposal is accepted, the opponent can propose any sequence of random events the combo could produce and take an action that advances the game state or prevents the opponent from winning or other way to say "take a Meaningful Action(tm)". He can do this as many times as he likes.
4. If the opponent can take no Meaningful Actions anymore, the original loop proceeds randomly, that is, towards the opponent's certain annihilation.

Or something of the sort. Basically, twist the randomness in the victim's favour so he is ensured the chance to interact in whatever manner he wishes to, and if he has no answers, kill him then and only then. Seems like a reasonable solution to me? Maybe force the initiating player to take a couple times through the loop first as the demonstration to ensure self-mill loops show the whole deck?

Well, let's see. I do not share the other poster's complaint that step 2 is asking too much of people. It seems to me that most tournament-viable indeterminate loops make it pretty clear that you will win. For example, take the Four Horsemen. How much math do you need to know to see that: (1) It is a perfect loop, returning to an identical board state each go-around EXCEPT THAT (2) your opponent may have less life; and (3) your library will be in a different order. Since you can always just try again if you whiff, it seems pretty clear you will win.

It's possible that there are situations that are not so clear. The more complicated loop involved that strange combo someone proposed awhile back where you and your opponent are passing turns with empty libraries and not drawing cards, you're casting Bolt every turn, they are casting Hero's Reunion each turn, and you are also casting Stitch in Time each turn. Since you're adding and subtracting life total, that loop would require some heavy duty math to show whether you are certain to win or not. (I tend to think you are because your opponent has a finite life total and can lose if you hit a string of good flips, but can never stop the process by simply gaining life. It's like how a casino always wins because it can keep playing the game with you. But, that's based just on my intuition and I don't know how to prove it.)

How frequently are you going to encounter such bizarre loops in practice, though? How frequently will you see an indeterminate loop that does a "random walk" towards victory? Or, similarly, how frequently will a loop "randomly walk" between you winning and losing?

So I think applying Rule 2 in practice is totally viable, but if there were concerns, then maybe you could modify it this way:

2. The loop must be of such a nature that each iteration has some nonzero chance of advancing the board state in a linear fashion towards victory, and/or winning the game, and no chance of moving the board state away from victory and/or losing the game. (Example: Each time you repeat the loop, you have a nonzero chance of assembling a combo or reducing your opponent's life total).

This would invalidate complicated proofs dealing with "random walk" which would be fascinating but not intuitive for people involved.

As for your Rule 3, are you suggesting that the shortcut rules simply get amended to say, "The opponent may elect to stop the shortcut at any other configuration that is a possible outcome at any point during the loop. If he or she does, the game state is moved to that configuration with the opponent retaining priority." ?

If so, that's interesting. It cuts the Gordian knot by simply stating, by fiat, that your opponent's selected state WILL occur first if you want to shortcut an indeterminate loop. So in playing the Four Horseman, you'd basically hand your library over to your opponent who can then stack the deck anyway they want. You probably need some provision saying that once your opponent has done this, if they then pass priority, you get to complete the loop to your preferred configuration, though, to avoid them doing this repeatedly.

Those fixes should work, but it's doing an AWFUL lot of work on the rules for essentially one deck. And, I wonder if there would be unintended consequences...?

I like your amendment of rule 2. And yes, Rule 3 means just that: The opponent is allowed to play out whatever answers he wants in an ideal istuation - if he's forced to F6, you win.

And yes, it's clunky for essentially one deck, but the current ban-fix is kind of the same, so amending things like this doesn't necessarily seem harmful to me? And I want to underline that the loop would necessarily have to end in the game being over (as in, draw from empty, 0 life, "player loses the game" or somesuch) so it doesn't get abused as a free library-stacking mechanism, only as a kill mechanism.

[cherub_daemon]

Other possible issues with Zombie's opponent-stack rule:


If the opponent can respond to each ideal situation (out to infinity) but not a non-ideal one, how do you handle it? That is, if an opponent can always respond to a given stack, can they keep stacking it that way?
The opponent will be able to stack the deck and respond to what would normally be non-public knowledge (library order).
A judge would be needed to ensure that the non-deterministic combo player can make all required decisions based on public information.


Also not clear, but maybe not a problem: how long is too long to spend digesting your opponent's library and stacking it?

(Edited: clarification of first bullet)

[TheInfamousBearAssassin]

It's worth noting that, as I already demonstrated and MaximumC simply seems to want to ignore, the problem doesn't actually arise out of indeterminacy but out of the mechanism of card order mattering (one could very pedantically argue that card order itself is a form of indeterminacy, even if the effect on the game state is clear, but one who does so should suck a donkey dick instead.) There's no issue such as the Ravenous Trap example with flipping a coin an arbitrarily large number of times, whereas it would come up with any combo that relied on milling and recycling the library into the graveyard and vice versa.

As far as Zombie's solution goes, it actually has a bit of precedent in 3CB, where any random event is supposed to turn out against the favor of the player who created it.

[Dzra]

Slow play should apply to the guy who takes 30 seconds per top activation every turn and the guy who tanks constantly when he's demonstrably locked into no actions and the player who always "has a response" in his decks without countermagic. It shouldn't apply to the guy who knows his deck and plays fast but ran afoul of a wrongly-interpreted rule.

If you think your opponent is playing slowly then ask him politely to play faster. Since I'm primarily a Miracles player (and thus take a while to win), I am constantly having to encourage players to play at a faster pace than they normally might be inclined. "Slow Play" is something different though. Slow Play isn't just playing slowly; it is intentionally stalling the game. If you think your opponent is slow playing then you have every right to call a judge and ask if he'll watch the opponent for Slow Play. They usually will, and most of the time your opponent will stop goofing off simply because there is a judge watching.

As far as how Slow Play applies to the Four Horsemen, it is totally valid to ask a judge to watch for Slow Play when you've been sitting around for 15 minutes while staring at the same board state with nothing happening but the graveyard being shuffled about. This rule is practical above all else; it saves judges and players both from having to sit around twiddling while the Four Horsemen player messes around.

[TheInfamousBearAssassin]

If you think your opponent is playing slowly then ask him politely to play faster. Since I'm primarily a Miracles player (and thus take a while to win), I am constantly having to encourage players to play at a faster pace than they normally might be inclined. "Slow Play" is something different though. Slow Play isn't just playing slowly; it is intentionally stalling the game. If you think your opponent is slow playing then you have every right to call a judge and ask if he'll watch the opponent for Slow Play. They usually will, and most of the time your opponent will stop goofing off simply because there is a judge watching.

As far as how Slow Play applies to the Four Horsemen, it is totally valid to ask a judge to watch for Slow Play when you've been sitting around for 15 minutes while staring at the same board state with nothing happening but the graveyard being shuffled about. This rule is practical above all else; it saves judges and players both from having to sit around twiddling while the Four Horsemen player messes around.

Is practicality why we're doing this instead of shortcutting to the statistically inevitable end.

[Julian23]

Slow Play isn't just playing slowly; it is intentionally stalling the game.

You are confusing Slow Play with Stalling here. They are two different infractions which very different penalties.

Slow Play = unintentional, will result in a warning.
Stalling = intentional, will result in a DQ for cheating.

[Dice_Box]

Slow play ruling stops someone just tapping and untapping Basalt Monolith to run out the clock. That's the point of it.

Stalling is to stop someone stilling on a Brainstorm for 35 min to stall out the game.

Different rules for different situations. Roughly made to do the same thing.

[MaximumC]

I don't think you guys are thinking about slow play or stalling properly. "Slow Play" is defined by 3.3 and includes BOTH taking too long to do something or repeating indeterminate loops. "Stalling" occurs when Slow Play is intentional, because then it's cheating.

So, stalling is slow play but not all slow play is stalling. (A square is a rectangle...)

[DragoFireheart]

I couldn't care less about advanced mathematics applied to children's card games and

Then leave instead of acting like a pretentious douche bag.

[Julian23]

I don't think you guys are thinking about slow play or stalling properly. "Slow Play" is defined by 3.3 and includes BOTH taking too long to do something or repeating indeterminate loops. "Stalling" occurs when Slow Play is intentional, because then it's cheating.

So, stalling is slow play but not all slow play is stalling. (A square is a rectangle...)

That's what I was trying to tell people. Also note that not all ways of stalling necessarily qualify for slow play, but most of the time will involve it.

[TsumiBand]

For the sake of argument, since apparently I can't let this go -- because honestly, conceptually, this combo really appeals to me. It does! It's very cool.

Let us assume - just briefly, for argument's sake - that it is okay to declare X iterations as "until condition is met". So how do you determine the game states when your opponent genuinely messes with your game plan? Let us assume that you are forced to go off or quickly lose to an aggressive deck, and you cannot exercise any control over the cards in their hand; so you demonstrate the combo one time and they allow you to show it off. You declare "until condition is met" iterations; they respond with a Krosan Grip on your Basalt Monolith.

So, among other things, this inexorably puts +1 card in your graveyard and removes your ability to respond to hate with 'going off'. Your shuffles are necessarily different than they would have been; there's an extra card in your yard. You can argue any point about declaring some vague yet real value about number of iterations -- and even if it cannot be precisely defined, one would have to assume it was based on the cards left in your library and the number of cards in your graveyard at the time. It *has* to be; the number of iterations through the loop is different if you have 6 cards in your library as opposed to 50.

If we assume that the game allows you to say "X" times, where X is "the right number of iterations until the condition is satisfied", it must surely be a number that you solve for according to the state of the game when you begin your iterations. Krosan Grip necessarily fucks with that. So it would not matter if we didn't have to solve for X when we declare it; but unless I'm missing something I'm fairly certain that we could no longer say that it plays out the same after that Krosan Grip. Unless I've missed something that makes it not matter that there's at least one extra card in the graveyard now. Indeed, you might be able to say there are four Narcomoebas in play, but you can't guarantee the content of your graveyard anymore, because Emmy shows up in different places and at different times. So now you *have* to be able to say what your board, graveyard, and library look like, and they may not have the ideal contents.

I believe that other split second spells would have this effect as well - though not so damning, as I'm fairly certain that the 4H player could recover/start over. The point here is that the 4H player cannot respond by "going off again" and just executing the combo while waiting for the other triggers to resolve. Extirpate would remove all of any instance of a nonland card, which would change the outcome. Sudden Shock would add a Narcomoeba back to the graveyard; that changes the endpoint. So let's forget whether or not the GS can be recovered favorably for the combo player - is it enough to say that even just touching the card count in certain zones with a ton of triggers on the stack yet to be resolved fundamentally changes the output at that gamestate - and if it does, can one move forward without actually determining what the state should be?

[anomie-p]

You seem to be stuck on the indeterminacy bit. It doesn't matter if the number of cards in the yard is different one time through vs. the same time through with an extra card in the yard due to opponent action, *if you can still flip*, you're guaranteed to get the graveyard state you want, for the exact same reason that you are guaranteed to get the graveyard order you want in the first place - which is that, if you don't hit, you just keep going.


it must surely be a number that you solve for according to the state of the game when you begin your iterations

No, it is not. It is a number that cannot be solved for, because if we could solve for it there would be uncertainty in whether or not we can hit the graveyard order we want. You are certain to hit it because if you loop it any number of times and *don't* hit, you just loop *more*. The very instant you try to resolve that number of loops to a definite number, in any way whatsoever, you are no longer talking about a certainty. This is why such an indeterminate loop requires a rules change in the first place, if it were a certainty with a determined n we wouldn't need a rules change to allow the loop.

[rufus]

For the sake of argument, since apparently I can't let this go -- because honestly, conceptually, this combo really appeals to me. It does! It's very cool.

Let us assume - just briefly, for argument's sake - that it is okay to declare X iterations as "until condition is met". ...

I was expecting reductio ad absurdum after that line. I think the idea of X iterations as 'until condition is met' doesn't make much sense when carefully inspected.

In the context of this thread, I had wondered whether a system where the player with a 'random loop' declares some finite number N of iterations that he won't exceed, but there's still the slow play / stall issue since the random move has to be played out each time.

[TsumiBand]

You seem to be stuck on the indeterminacy bit. It doesn't matter if the number of cards in the yard is different one time through vs. the same time through with an extra card in the yard due to opponent action, *if you can still flip*, you're guaranteed to get the graveyard state you want, for the exact same reason that you are guaranteed to get the graveyard order you want in the first place - which is that, if you don't hit, you just keep going.



No, it is not. It is a number that cannot be solved for, because if we could solve for it there would be uncertainty in whether or not we can hit the graveyard order we want. You are certain to hit it because if you loop it any number of times and *don't* hit, you just loop *more*. The very instant you try to resolve that number of loops to a definite number, in any way whatsoever, you are no longer talking about a certainty. This is why such an indeterminate loop requires a rules change in the first place, if it were a certainty with a determined n we wouldn't need a rules change to allow the loop.

So even though the number would necessarily have to have been one value before the Krosan Grip, but a different value afterwards, you're saying that shouldn't matter because it can't be solved for anyway?

Keep in mind that the deck still needs to resolve Dread Return --> Sharuum at some point, which is a Sorcery effect. And after doing so, it again needs to be able to combo with Blasting Station. So it seems like you still have the requirement of burden of proof that you've actually attained the desired state where Dread Return hit the yard, but Emrakul didn't, and you have priority during your main phase. I don't see how you can say "X times" or "indeterminate times" when the solution would be fundamentally different with a different number of cards in the graveyard or library at any given point in the process.

That's why I specifically mentioned Krosan Grip - it has split second. It can't be responded to with activated abilities outside of a mana ability. With Basalt Monolith in the graveyard you're left to the whim of the stack now; there's no responding to triggers to manipulate the game state in your favor. All you can do at that point is play it out with a new card in the graveyard, so whatever value you declared when attempting to go off - whether it's a real number or not - it's still locked in. I don't understand why that would be irrelevant.

EDIT: Okay so wait, you're arguing for a rules change that allows you to simply declare an undefined non-infinite loop, so it would not matter what the opponent did because you would simply check to see that the combo could still be executed and you would presume to adjust for that state. I think that's a really slippery slope, if I've understood correctly. It would allow for players to handwave in situations where they can't actually fight their way through the hate but they can claim it is non-relevant. I'm trying to think of a better way to describe what I mean by this, but wouldn't you know it I'm at work :/ blast


I was expecting reductio ad absurdum after that line. I think the idea of X iterations as 'until condition is met' doesn't make much sense when carefully inspected.

In the context of this thread, I had wondered whether a system where the player with a 'random loop' declares some finite number N of iterations that he won't exceed, but there's still the slow play / stall issue since the random move has to be played out each time.

Have I committed a logical fallacy? I don't know that Four Horseman versus Krosan Grip is that big a straw-man. I'm just trying to grok what the intention behind just declaring "until I get there" iterations is supposed to mean, or if it is NOT about 'locking in' the iterations, the claim that a guaranteed game state can be achieved should simply allow players to fast-forward to that game state, with unintended interactions being introduced that could no longer be adjusted for. It's one kind of shortcut to announce billions of Grixis-flavored mana with Deadeye Navigator and Palinchron; we can expressly determine what the board will look like when I cast my hate card. If I can shut off a crucial piece of the 4H combo with "enough" activations on the stack, I don't think it is unreasonable to expect to land on a different game state instead of the ideal one, because the conditions of the loop have been altered.

I mean, at this point it just feels like someone wants to reserve the right to play the "I know you know I know about your hate" game. It *feels like* saying "Well I *anticipated* your hate card when I said I was going off, so NYAH. I still assemble Voltron, because I should get there somehow anyway." Unless I'm just missing something integral to the argument.

[anomie-p]

So even though the number would necessarily have to have been one value before the Krosan Grip, but a different value afterwards, you're saying that shouldn't matter because it can't be solved for anyway?

Well, I'm commenting on the idea that it somehow matters whether or not you have one more or one less card in your yard, compared to your opponent not taking an action, yes. And your statement that there is somehow a solution (because there isn't, if you define n, your probability is no longer 1).


As for Grip, it kills the combo (so they'd need another monolith in hand to keep going), but if you *can* keep going, the fact that you did (or didn't) put one more card into the yard doesn't mean anything*. (Discussing the effect of grip exactly requires a few things, such as, by 'respond' do you mean, the grip player is declining the shortcut, or accepts it and grips when they get priority, etc, but it really all boils down to 'You will need another monolith in hand, or a way to draw one' - but that's a separate thing to me than the idea you seem to keep putting forward that N is something to be solved for, or that you can resolve it to a determinate number, etc).

Edit: *and by that, I don't mean there is no situation in which it would potentially matter, just that it doesn't matter with respect to N being different or not different. The value of N really, really does not matter except that we *cannot* require that it be defined if we want to allow a nondeterministic loop shortcut.

[TsumiBand]

Well, I'm commenting on the idea that it somehow matters whether or not you have one more or one less card in your yard, compared to your opponent not taking an action, yes. And your statement that there is somehow a solution (because there isn't, if you define n, your probability is no longer 1).


As for Grip, it kills the combo (so they'd need another monolith in hand to keep going), but if you *can* keep going, the fact that you did (or didn't) put one more card into the yard doesn't mean anything*. (Discussing the effect of grip exactly requires a few things, such as, by 'respond' do you mean, the grip player is declining the shortcut, or accepts it and grips when they get priority, etc, but it really all boils down to 'You will need another monolith in hand, or a way to draw one' - but that's a separate thing to me than the idea you seem to keep putting forward that N is something to be solved for, or that you can resolve it to a determinate number, etc).

Edit: *and by that, I don't mean there is no situation in which it would potentially matter, just that it doesn't matter with respect to N being different or not different. The value of N really, really does not matter except that we *cannot* require that it be defined if we want to allow a nondeterministic loop shortcut.

I guess then the issue for me is, I understood previously that number_of_iterations could be undefined so long as it was treated as though it were unchanging, whatever it is. This would allow for declarations of exceptionally high values for loops to be executed while maintaining "no infinities in Magic", because we don't have to say what N is, necessarily; just that it is "enough to get there". However, I felt this didn't cover situations where changing the outcome before the last iteration would then cause the previous value of N - again, whether or not it was ever concretely defined, as long as it was treated as though it were 'locked in' as Magic prefers to do - to no longer apply.

I was also operating on the assumption that the operator was at some point putting multiple triggers on the stack and then passing priority, but for the purpose of shortcutting, we are assuming there is never more than one activation of Monolith on the stack. Although, now I think I see that it does not matter either way; a successful Disenchant/Krosan Grip would leave nothing to untap, so even if there were so many triggers on the stack they wouldn't actually do anything. Nothing would untap, so Mesmeric Orb would not dump additional cards in the graveyard after Monolith leaves play.

I can see I need to reconsider my position somewhat. This is interesting...

[anomie-p]

I guess then the issue for me is, I understood previously that number_of_iterations could be undefined so long as it was treated as though it were unchanging, whatever it is. This would allow for declarations of exceptionally high values for loops to be executed while maintaining "no infinities in Magic", because we don't have to say what N is, necessarily; just that it is "enough to get there". However, I felt this didn't cover situations where changing the outcome before the last iteration would then cause the previous value of N - again, whether or not it was ever concretely defined, as long as it was treated as though it were 'locked in' as Magic prefers to do - to no longer apply.

Well, from my side of it, if you're not treating the N as something you truly cannot know, you end up with a probability that is < 1. You can't have a probability of 1 without allowing an undefined loop count, and if you're not setting something up where the probability is 1 ... then the probability is not 1, the probability is less than 1, and you can't say that you are guaranteed to get the graveyard order you want.

There is no way to have both a) a definite N and b) certainty about the outcome, and if there's uncertainty about the outcome (even if the uncertainty is very, very small), I don't see any reason for your opponent to let you just do it.

Also, I'm not arguing for a rules change here. I'm just trying to point out certain things that your posts seem to indicate you're not catching. The point here is that, we are not saying 'under this nondeterministic loop rule that has been proposed, we are just ignoring some small uncertainty that it won't go our way". The only reason such a nondeterministic shortcut would work at all is because it is mathematically certain that you will get the state you want in *some* iteration, but that proof requires that the n be left undetermined and unstated, and it is not something you are solving for or even attempting to solve for.

Under the proposed loop-until-condition rule, I can't even think of a situation where you'd be able to even treat the n as 'locked in but undefined' and then somehow 'change' it, if your opponent doesn't accept your shortcut, it's not locked in, if they do accept the shortcut, then you looped whatever n was without actually knowing it or defining it, and a post-loop action by your opponent can't change that. (if they say 'nah, I counter propose that you loop until condition Y', and you accept, and then they put a Grip on the stack - it may be the case that you would have looped a different number of times if they'd just accepted your proposal, but they didn't ...)

Again, this does not mean we should have such a rule - as this thread has shown, there are good reasons for *not* allowing it, and I personally am on the side of not allowing them - but if it were to be, we wouldn't be talking about hand waving away some small uncertainty, we'd be saying "The 4H player gets to stack his graveyard this way because the probability that the graveyard will turn up in that order is actually 1 (or 100%), although we can't know how many iterations it will take".

[anomie-p]

Maybe think about it like this:

Say we're playing a game, and the rules of this game allow me to set up an infinite coin flipping loop. I say, "I have a coin, I am going to flip it, if it comes up heads, I will flip it again, if it comes up tails, I will stop. Is it ok with you if I shortcut however many heads flips I'd get and just put the coin on the ground tails up?".

Can you see that as long as I am allowed to re-flip until I get a tails, I am guaranteed to get the tails? (it's the same math, basically, you take the limit of (1/2)^n as n goes to infinity as your odds of failure, and that limit is also 0, so your probability is also 1 - 0, which is 1).

This is a similar situation mathematically, except there's only two states, instead of the large (but, like the number 2, finite) number of potential states in the 4H graveyard. If the rules of whatever game we are playing let me do that infinite coin flip, and a tails on the ground is a win, and the coin is fair, are you *really* going to say 'I don't believe you're ever, ever going to hit that tails'? If I flip heads, I'm just going to flip again ...

It's the same thing with 4H. If you don't hit it, you just shuffle your yard up and try again ... the number of possible states looks like a lot in comparison to the two states of a coin flip, but when you can just keep doing it, you will get what you want (unless you're failing to sufficiently randomize - but that's the same as the coin not being fair).

[amalek0]

at this point I will jump in and inform you that I have in fact played against T4H before, and My solution has always been the same: play it out, no you don't get to shortcut it, I want to see the kill. Twice I've had the pilot of the deck actually screw up the stacking of his triggers anyway and won because of it. The four horsemen is not really a viable deck, for more reasons than just the issue of the looping--people who *actually* can play combo usually would rather play something more stable than the 4 horsemen, and scrubs will be scrubs whether it's screwing up their show and tell or their triggers with emrakul.

[MaximumC]

at this point I will jump in and inform you that I have in fact played against T4H before, and My solution has always been the same: play it out, no you don't get to shortcut it, I want to see the kill. Twice I've had the pilot of the deck actually screw up the stacking of his triggers anyway and won because of it. The four horsemen is not really a viable deck, for more reasons than just the issue of the looping--people who *actually* can play combo usually would rather play something more stable than the 4 horsemen, and scrubs will be scrubs whether it's screwing up their show and tell or their triggers with emrakul.

You were actually being very generous with the pilot. Letting them try is not the optimal play.

The correct thing to do when Four Horseman goes off is to call a judge to watch immediately. That's because, as we discussed in this thread, the pilot does not get to "play it out" unless he or she can change the board state every time he fails to achieve what he or she wants during the loop.

In other words, if he or she tries to mill herself and hits Emrakul at the wrong time, then he or she has to do something else to change the board state before he or she can try again. So you're better off calling a judge to enforce that rule. Anyone piloting this deck should know darn well that is going to happen.

[amalek0]

I'm not going to do that When it's a local Thursday-night sanctioned 8 man legacy event and I'm judging it =).

Besides, we had plenty of time for him to try, the local guy playing Bgu pox had everything resembling a win condition extirpated out from under him and was trying to ride academy ruins/mox diamond to a win. But I digress. I'm familiar with the requirements to change the boardstate each iteration, but honestly there's no reason to enforce it in a small event where half the players still hate wizards for printing planeswalkers in the first place.

[TheInfamousBearAssassin]

I notice MaximumC still hasn't addressed how to resolve Ravenous Trap with standard infinite/determinate combos.

[MaximumC]

I notice MaximumC still hasn't addressed how to resolve Ravenous Trap with standard infinite/determinate combos.

Your posts are pretty crabby and abrasive, so I tend to just skim them. I didn't realized you asked me a question.

Are you asking me what happens if Player A proposes a regular finite shortcut of X iterations involving putting her library into her graveyard, and Player B has a Ravenous Trap and wishes to see what is in the yard before pulling the trigger on the Trap?

Well, let me take a peek at the Comp. Rules and IPG. I'm looking for a definition of "game state." Apparently, there is not such a definition. I am going to assume -- without actually knowing -- that the cards in your graveyard changing does represent a change in game state. This assumption underlies my entire analysis.

Now, to your point. Player A proposes to shortcut a loop X times putting X cards into his graveyard. Can he do this? Well, is he "describing a sequence of game choices, for all players, that may be legally taken based on the current game state and the PREDICTABLE results of a sequence of choices[?]" If the graveyard order and contents are part of the game state, then I say - NO SIR HE CANNOT.

Why not? Because he's putting cards into his yard in a random order. He cannot shortcut the milling process. He has simply do it. However, all is not lost!

It is perfectly legal to "play out" milling your entire library, because each time you mill a card, you change the game state. (Four Horseman runs into trouble because randomizing your random library is NOT a change in game state, I believe). So that works fine.

And, remember, this is how you do it anyway. People don't turn their library over. They pick it up and flip card by card, (mostly because they're looking for narcomebias.)

The upshot of this is that Ravenous Trap is no problem at all. You can't shortcut, you just do it, so when the opponent wants to trap, they just interrupt you and do it.

[TheInfamousBearAssassin]

When people lack both good faith and any kind of reasonable sense of pretty simple math I guess I tend to get a bit short with them, my apologies.

But you're clearly trying to have it both ways. Randomizing your library/graveyard order can't simultaneously be not changing the board state for purposes of slow play, but changing the board state for purposes of not being able to short cut.

This is an asinine and clearly post-hoc rationale for the status quo based on basically nothing.

[Dice_Box]

When people lack both good faith and any kind of reasonable sense of pretty simple math I guess I tend to get a bit short with them, my apologies.
Mate we have argued Math with you, but you read and translate what you like from peoples posts without reading them correctly. At lest 3 posts of mine you have read what you want to read and not what I have said. People here understand Math, they just openly think your trying to manipulate posts to your own gain and just don't argue with you any more.

Snip. Knock it off with flames and responding to them. Report and ignore them. -Jander

[Koby]

Yet, the prior objection of Ravenous Trap + Top still applies, because the mechanic of the combo is still the library hitting the graveyard. So, how would you resolve that? Allow the Top player to essentially thwart a demonstrably predictable, demonstrably infinite combo on the basis that there may be an answer on top of their library?

We've already pointed out, once the Grinding Station / Goblin Bombardment card is in play, the combo is determinant and relatively easy to resolve. It plays out as you've posted in ginormous text.

The 4H player proposes to shortcut (pre-setup, or post) Basalt Monolith/Mesmeric Orb interaction by stating: "I'd like to shortcut 60 iterations of Basalt Tapping/untapping." Opponent can accept this shortcut, and those actions occur immediately. Alternatively, he can counter by stating: "We can resolve those milling triggers by placing all incidental triggers on the stack and continuing to mill after each one." This will create a stack with 60 mill triggers interset with Narcomoeba and an Emrakul trigger. The 4H can continue to counter with "Resolve upto 60 mill triggers until a graveyard trigger appears." This creates a conditional stop based upon Narcomoeba or Emrakul. Finally, if accepted by both players, the mechanism of the shortcut would flip cards into the graveyard until either Narcomoeba or Emrakul triggers.

At this point, the 4H can choose to continue with the milling shortcut, or let the ability resolve. In the case of Narcomoeba, it enters the battlefield and the game state and graveyard is otherwise unaffected. Continuing with milling will produce a similar scenario as stated in the above paragraph.

The opponent with Ravenous Trap + SDT does not need to reveal at which point he is playing RT, nor that he even has such disruption. The mechanism of the shortcut rule forces the shortcut/loop to interrupt when any non-expected result occurs. In this case, the non-expected result is a graveyard trigger.

Thus, the strategic best time to cast Ravenous Trap is when Emrakul triggers. This is based on the 4H deck construction (1 Emrakul) and is the entire basis of the recursion. The difficulty lies in the 4H player's knowledge of these shortcuts, as any permutation of shortcut handling could result in different pieces being lost to Ravenous Trap.

Even if the MTR allowed "infinite" iterations of the shortcut, there would still need to be ways to resolve graveyard triggers. This is based upon how the 4H player sets up the shortcut proposal.

[MaximumC]

lack both good faith
clearly
asinine
clearly post-hoc
status quo.

I humbly suggest, for the good of the thread, that you remove these and similar words from your vocabulary. What actual communication happens when you use charged and/or "I took Logic 101" terms like this?

As to you reaction, I believe that changing the order of cards in a randomized library to another random order is not a change of game state. I believe changing the order of cards in your graveyard IS a change of game state. I know I've seen some description of game state somewhere, but I cannot find it in the rules. This would make sense, though, because there are cards that depend on graveyard order, and graveyard order is known information.

EDIT: For clarity's sake, I think that current rules imply the following.

Change in card order in a randomized LIBRARY = not a change in game state.
Change in card order in a GRAVEYARD = change in game state.

EDIT2: I really don't know if there are weird interactions with Search for Survivors or Fossil Find that crop up in this analysis...

[rufus]

...

Change in card order in a GRAVEYARD = change in game state.

If this were true, a 4H player could do the Orb/Monolith trigger in large batches, to make likely that some orb triggers would end up stacked 'under' the Emrakul trigger and the odds of having to take an action in a repeated game state would be astronomically small. (I suspect that, in that regime, optimal 4H play would involve a lot more reshuffles per game.)

[MaximumC]

Let's just ignore him [IBA]. Otherwise, this thread is going to just devolve into people flaming each other, and it's too interesting to fall apart like that.

I was thinking more about the implications of graveyard order being part of the game state. (Again, I can't figure out where "game state" is defined, so I might be wrong about that fact.) Let's say you have the following board:

Aphetto Alchemist equipped with Illusionist's Bracers
Llanowar Elves
Elite Arcanist "imprinted" with Search for Survivors
Island, untapped

And your graveyard contains:
Oona, Queen of the Fae
12 other non-creature cards.

Your Alchemist gives you infinite untap effects. The Elves then add infinite mana, and the Arcanist allows you infinite casts of Search for Survivors. Each time you use Search, you have a 1/13 chance of hitting Oona. If you do, then you win because you can activate her ability a million times to mill out your opponent. But, 12/13 of the time, you will end up missing Oona and return to the same game state.... but with a randomized graveyard.

Can you shortcut this? Can you play this out?

I tend to think the answers are no and yes, respectively, but I'm not sure.

EDIT @ rufus: I do not think Four Horsemen runs into problems for slow play until it reshuffles its library. My understanding is that as long as you are filling your graveyard and all that, you're good to go. The problem occurs because when you start with an empty yard and a randomized library, and then you eventually return to that exact same state with no change on the board. That's when you've done a "loop" and you can't continue what got you there. I think.

[rufus]

...

EDIT @ rufus: I do not think Four Horsemen runs into problems for slow play until it reshuffles its library. My understanding is that as long as you are filling your graveyard and all that, you're good to go. The problem occurs because when you start with an empty yard and a randomized library, and then you eventually return to that exact same state with no change on the board. That's when you've done a "loop" and you can't continue what got you there. I think.


Let's say that a 4H player has just reshuffled after digging up the third narcomoeba, and has 40 different distinct cards left in his library, 4 of which are Emrakul, Dread Return, Sharuum and Blasting Station and no cards in the graveyard.
Now, he taps and untaps the bassalt monolith 40 times, stacking the orb triggers.
Then he lets the triggers resolve, flipping cards off the top of his deck into his GY.
At some point, Emrakul hits the graveyard (most likely with a bunch of orb triggers still left on the stack).
The Emrakul trigger goes on the stack, and is resolved before the remaining Orb triggers, reshuffling the graveyard into the library.
Then the remaining orb triggers. Assuming that he doesn't hit Emrakul again, when the stack is empty he'll have some cards in his
graveyard. Most likely, this is a new and different graveyard composition than before.

N.B.: Your example case doesn't quite work the way you want.
Search for Survivors will exile non-creature cards, so you'd need a pile of 0-toughness creatures in the graveyard. (And even then, you could just stack 13 copies of the spell.)

[MaximumC]

Let's say that a 4H player has just reshuffled after digging up the third narcomoeba, and has 40 different distinct cards left in his library, 4 of which are Emrakul, Dread Return, Sharuum and Blasting Station and no cards in the graveyard.
Now, he taps and untaps the bassalt monolith 40 times, stacking the orb triggers.
Then he lets the triggers resolve, flipping cards off the top of his deck into his GY.
At some point, Emrakul hits the graveyard (most likely with a bunch of orb triggers still left on the stack).
The Emrakul trigger goes on the stack, and is resolved before the remaining Orb triggers, reshuffling the graveyard into the library.
Then the remaining orb triggers. Assuming that he doesn't hit Emrakul again, when the stack is empty he'll have some cards in his
graveyard. Most likely, this is a new and different graveyard composition than before.


Hm, that's true. Would a judge call this out as a loop that violates slow play...?



N.B.: Your example case doesn't quite work the way you want.
Search for Survivors will exile non-creature cards, so you'd need a pile of 0-toughness creatures in the graveyard. (And even then, you could just stack 13 copies of the spell.)

Good point. I didnt notice that it exiled cards out of the yard. That makes it deterministic.

EDIT: I think your proposed correction makes it non-deterministic again, though. Say your yard contains Oona and for Force of Savagery. It doesn't matter if you "stack" five activations of Search, because between each resolution of Search the game will check for state based actions and dump the Forces back in the yard. Same question arises.

Okay, so maybe the order and identity of cards in the yard is not part of the game state? But, that can't be true. If it was, then you would be committing slow play the moment you milled your first card and untapped Monolith. You're back in the same game state, the only difference being a card in your yard.

So is it or isnt it...?

[rufus]

Hm, that's true. Would a judge call this out as a loop that violates slow play...?


cdr almost certainly would. (Edit: Not to mention that keeping track of all the GY compositions is a nightmare.)
Edit 2:
http://www.mtgthesource.com/forums/showthread.php?24651-SCG-LA-Four-Horseman-in-17th-How&p=673348&viewfull=1#post673348
TPTB have apparently issued a "no-4H" edict.

[MaximumC]

cdr almost certainly would. (Edit: Not to mention that keeping track of all the GY compositions is a nightmare.)

I think you're right, but doesn't that imply that graveyards do not count as part of the game state, and therefore do not save you under slow play rules? And doesn't that imply you get into trouble much eariler, as you start to fill your yard the first time? And imply that decks like Cephallid Breakfast are also committing slow play when they fill the yard?

Or is yard-manipulation sometimes considered part of the game state and sometimes not? Based on what?

Inquiring minds want to know!

EDIT: In your linked judge ruling, it sounds like the judge also agreed that the yard WAS a change in game state:

"During Round 3 of the tournament, I was made aware of a Four Horsemen player on the feature match table. I went over to watch the match, knowing that I was likely to see a problematic line of play according to the IPG. When the player started to flip cards from the Basalt Monolith/Mesmeric Orb combination, he quickly ran into Emrakul, and was forced to shuffle his library. After doing this again, he was left in an identical game state: An empty graveyard and no other change to the game state. By performing the same loop of actions without changing the game, he was violating the shortcut policy outlined in the Magic Tournament Rules and the Slow Play policy in the Infraction Procedure Guide."

Sounds like the problem occurs when the shuffle happens, like I said.

In the example you made a minute ago, perhaps the answer is that you do not look at what is on the stack when you are deciding the game state? In your example, the only difference was you had some Orb triggers sitting around waiting to continue the loop.l

[rufus]

... Based on what? ...

(Honestly not snark.) AFAICT whether the judges like your deck.

[Koby]

Cards in graveyard, or the adding/removing of cards in graveyard are considered part of the game state. The ruling from that tournament specifically called out the result of the completed loop -- empty graveyard and no change on the battlefield. The distinction here is that the loop hit Emrakul, and reset the state. Doing so twice without successful resolution of the desired ordering of the graveyard is what the HJ ruled as Slow Play, and violation of the shortcut policy.

[MaximumC]

(Honestly not snark.) AFAICT whether the judges like your deck.

Yeah, I was worried about judge's discretion entering into the whole shebang. Since the slow play rule doesn't define "continue," it's potentially always there. In fact, my worry that Four Horseman was guilty of nothing more than annoying judges is what made this issue into a pet peeve for me in the first place.

I thought we had sorted out the rules in this thread in such a way that the rules were logical, consistent, and kept Horseman from working for a good reason. I had kind of made peace with that issue.

Now this "graveyard order" issue is bugging the hell out of me and putting us back, maybe, in the realm of judge's discretion.


Cards in graveyard, or the adding/removing of cards in graveyard are considered part of the game state. The ruling from that tournament specifically called out the result of the completed loop -- empty graveyard and no change on the battlefield. The distinction here is that the loop hit Emrakul, and reset the state. Doing so twice without successful resolution of the desired ordering of the graveyard is what the HJ ruled as Slow Play, and violation of the shortcut policy.

Yeah, that's what I thought. Where can I find a statement of what makes up the "game state" and what does not, though? Where does the rule about graveyard manipulation being part of the game state come up?

[TheInfamousBearAssassin]

Cards in graveyard, or the adding/removing of cards in graveyard are considered part of the game state. The ruling from that tournament specifically called out the result of the completed loop -- empty graveyard and no change on the battlefield. The distinction here is that the loop hit Emrakul, and reset the state. Doing so twice without successful resolution of the desired ordering of the graveyard is what the HJ ruled as Slow Play, and violation of the shortcut policy.

The nominal logic doesn't care about changes in the battlefield per se, only changes in the game state.

I am amused that MaximumC waited until someone else made the same point to acknowledge it.

[TsumiBand]

I think you're right, but doesn't that imply that graveyards do not count as part of the game state, and therefore do not save you under slow play rules? And doesn't that imply you get into trouble much eariler, as you start to fill your yard the first time? And imply that decks like Cephallid Breakfast are also committing slow play when they fill the yard?

Or is yard-manipulation sometimes considered part of the game state and sometimes not? Based on what?

Inquiring minds want to know!

EDIT: In your linked judge ruling, it sounds like the judge also agreed that the yard WAS a change in game state:

"During Round 3 of the tournament, I was made aware of a Four Horsemen player on the feature match table. I went over to watch the match, knowing that I was likely to see a problematic line of play according to the IPG. When the player started to flip cards from the Basalt Monolith/Mesmeric Orb combination, he quickly ran into Emrakul, and was forced to shuffle his library. After doing this again, he was left in an identical game state: An empty graveyard and no other change to the game state. By performing the same loop of actions without changing the game, he was violating the shortcut policy outlined in the Magic Tournament Rules and the Slow Play policy in the Infraction Procedure Guide."

Sounds like the problem occurs when the shuffle happens, like I said.

In the example you made a minute ago, perhaps the answer is that you do not look at what is on the stack when you are deciding the game state? In your example, the only difference was you had some Orb triggers sitting around waiting to continue the loop.

Honestly I'm not sure that this is a good example of an unchanged game state. I feel like the judge made a bad call here.

If Emrakul had a replacement effect, I could understand the reasoning here. But Emmy has a triggered ability; it goes to the yard and both players have to pass in succession. It can't be the case that 'the game state' is only checked at sorcery speed, can it? By this reasoning, wouldn't you be unable to, like, combo out with Aluren and Man-o-war? The creature enters the battlefield, targets itself, returns to hand -- and the game is precisely where it was previously.

sent from phone, don't be a dick

[MaximumC]

It can't be the case that 'the game state' is only checked at sorcery speed, can it?


I don't know...? I still can't find the rules related to how a "game state" is determined, and it sounds like no one else can either!

But if it were true that the stack is not part of the "game state," then the judge's ruling would make sense, no? You can keep going after Emrakul's trigger is on the stack, sure. But once you let it resolve and your yard is empty, it's possible to be back where you started (sans Orb triggers).

I mean, ideally you hit a narco or something and you've changed the game state by putting a new creature out. But let's say you got all three out, and are now looking to mill yourself into the three combo pieces without milling Emrakul. That's the state where you get into trouble... I think. Because there you can't do anything until you hit the right order in the yard -- Dread Return is sorcery speed.

EDIT: None of this clarifies what happens in the non-deterministic graveyard loop involving Search for Survivors on the last page, though. Seems like that would work and you'be be immune from slow play, yes?

[Dice_Box]

I think the simplest way to look at it is to take a snapshot of the game when the stack is empty, compare it to the next time the stack is empty and check for changes. If you do this over a period of time taking snapshots and continually comparing them and see a pattern of no change repeating itself, then that's a stale game state.

I view the "Gamestate" as a save point that you can compare or wind back to should the judge need you to do so.

Do note, I have seen games winded back with a stack intact, so I can admit this is not an optimal explanation, but it's how I view it.

[Koby]

The nominal logic doesn't care about changes in the battlefield per se, only changes in the game state.


Yes, that was my lapse. Under this scenario, the delta in game state is assumed to only effect the battlefield & graveyard by the nature of the 4H combo. Resetting the graveyard excludes that consideration, thus leaving only the battlefield.


I don't know...? I still can't find the rules related to how a "game state" is determined, and it sounds like no one else can either!

Let's call the initial state prior to the loop as <hand>, no graveyard, and the 4H setup combo in play. Iterating through the milling at least 60 times returns you to this very state. You've added triggers to the stack and resolved them all, but nothing was done to change the initial game state. You're basically just eaten ~1 minute of game time with no advancement. Don't get too hung up on the "game state," it's just a snapshot of the cards and their locations. The real crux of the matter from officiating the game is verifying players advance the game with a reasonable pace of play.

Game state is referred in the Infraction Penalty Guide for Judges to follow:



Definition
A player takes longer than is reasonably required to complete game actions. If a judge believes a player is intentionally playing slowly to take advantage of a time limit, the infraction is Unsporting Conduct — Stalling.

It is also slow play if a player continues to execute a loop without being able to provide an exact number of iterations and the expected resulting game state.

Examples
A. A player repeatedly reviews his opponent’s graveyard without any significant change in game state.
B. A player spends time writing down the contents of an opponent’s deck while resolving Thought Hemorrhage.
C. A player takes an excessive amount of time to shuffle his deck between games.
D. A player gets up from his seat to look at standings or goes to the bathroom without permission of an official.

Philosophy
All players have the responsibility to play quickly enough so that their opponents are not at a significant disadvantage because of the time limit. A player may be playing slowly without realizing it. A comment of “I need you to play faster” is often appropriate and all that is needed. Further slow play should be penalized.

Additional Remedy
An additional turn is awarded for each player, to be applied if the match exceeds the time limit. If multiple players on each side are playing the same game (such as in Two-Headed Giant) only one additional turn is awarded per team. This turn extension occurs before any end-of-match procedure can begin and after any time extensions that may have been issued.

No additional turns are awarded if the match is already in additional turns, though the Warning still applies.

If Slow Play has significantly affected the result of the match, the Head Judge may upgrade the penalty.

[MaximumC]

I think the simplest way to look at it is to take a snapshot of the game when the stack is empty, compare it to the next time the stack is empty and check for changes. If you do this over a period of time taking snapshots and continually comparing them and see a pattern of no change repeating itself, then that's a stale game state.

I view the "Gamestate" as a save point that you can compare or wind back to should the judge need you to do so.

Do note, I have seen games winded back with a stack intact, so I can admit this is not an optimal explanation, but it's how I view it.

That's a very sensible suggestion on how to do it. But, do the rules really work that way? If they did, then wouldn't rufus be right, and Four Horseman would be able to avoid any slow play penalty? Remember, the Four Horseman player has control of Monolith and Orb, and can put as many Orb activations on the stack as he or she likes whenever he or she likes. If you only checked for a repeated game state when the stack was empty, it would be trivial for the Horsey player to make sure the stack was never empty while they tried to go off. They could go on all day that way.

So, I don't think your suggestion describes the rules as they are.