[Level Up Legacy] Miracles

[jrsthethird]

Zoo was a natural foil to Counterbalance decks, often applying enough pressure by turn 3 that even the CB/Top lock wouldn't really help them, and they'd have to wait until turn 4 to tap out and play a sweeper. By that point, they're low enough that a Bolt/POP/single unanswered creature could finish them off. Terminus gave them a one-mana Wrath that can clear the board at instant speed, giving the deck an answer to aggro and ultimately killing true aggro decks in the format.

Given this, Miracles ascended to the top deck in the format (or top 3, at least) for an extended time. New players want to play top decks, so now we have a huge percentage of the field playing Miracles and the people who can't play at a fast pace make the entire archetype look bad. Before when it was just Sultai CounterTop with Goyf or whatever, or Esper CounterTop with Thopters, the decks weren't DTB, so there was a much lower percentage of the player base playing the decks, with a presumably higher percentage of those players being experienced/specialist players. Because what newb would pick up a deck with an awkward combo when you could play Maverick or Zoo or Merfolk instead and have a better chance to win?

[Zombie]

Zoo was a natural foil to Counterbalance decks, often applying enough pressure by turn 3 that even the CB/Top lock wouldn't really help them, and they'd have to wait until turn 4 to tap out and play a sweeper. By that point, they're low enough that a Bolt/POP/single unanswered creature could finish them off. Terminus gave them a one-mana Wrath that can clear the board at instant speed, giving the deck an answer to aggro and ultimately killing true aggro decks in the format.

Supreme Blue did play Firespout.

[GundamGuy]

. DTB methodology is solid, but it doesn't do what most people think it does. A deck's presence (and rank) in the DTB section does not intricate the decks strength or positioning in the meta - it's skewed to favour decks which are popular and see more play. DTB does exactly what the name says - it tells us which decks we are likely to face and a should therefore be perpared "to beat".

You should have kept reading.

So how are decks selected for the DTBF?

Rather than relying on arbitrary selection or decision-making based on conjecture which can be tainted by personal bias, decks are selected for the DTBF based on their performance at recent, large, competitive Legacy tournaments. Decks which make up a very large portion of the metagame are considered DTB's. Decks which are less prevalent but appear multiple times are considered DTW's. Archetypes which appear multiple times are considered ATW's.

The data used for selection is based upon Top 8's from the most recent Legacy tournaments with 33 or more players...

Bold was added by me.

Additional Info:

In light of both Grand Prix: Chicago and the third Bazaar of Moxen tournament, it has been decided to switch to a slightly more sophisticated system for determining deck scores, so that placements in larger tournaments receive an appropriately heavier weighting:

Points per placement = {Number of Swiss rounds} + {Number of Top 8 rounds won}

So really, the Decks to Beat are the most popular because they win the most... a deck that is widely played but does not win much won't be a Deck to Beat...on the other hand a deck that isn't widely played but places on top once won't make the decks to beat either... so on one hand you are right that a deck must also be both popular and good to be a deck to beat...

But the key here... is the And Good part... and based on how many Placement Points Miracles has been getting constantly for 20 or so months... it seems to have the And Good part down. What I take from this is Miracles is widely played and consistently places well regardless of other shifts in the meta.

[Crimhead]

Did keep reading, nothing you've quoted contradicts what I quoted. It only make sense that Zilla isn't contradicting himself!

What I quoted (now in my SIG) is stating that frequency in top brackets actually =/= being tier one.

The basic idea is that everyone playing Miracles (or whatever) has some chance of placing. It follows that the chance of seeing Miracles in the top bracket is a funtion of both the power of the deck and the number of players who are pushing it. This is the most basic of mathematics (binomial expansion).

...a deck that is widely played but does not win much won't be a Deck to Beat...on the other hand a deck that isn't widely played but places on top once won't make the decks to beat either... so on one hand you are right that a deck must also be both popular and good to be a deck to beat...

But if you take this further, it follows that the more popular a deck is the less good it has to be to qualify. A deck that is merely average can still make DTB - it just needs to be very popular. Hence the quote in my sig.

So how Dow we know if a DTB is tier one or merely decent but popular? Factor out the popularity quotient to isolate the desired variable (deck strength)! In other words, divide the number of top finishes by the number of players who ran the deck in those events. Miracles has shown repeatedly that its high frequency in tops is directly proportional to its density in the field on day two (no data released for day one).

[GundamGuy]

A deck that is merely average can still make DTB - it just needs to be very popular. Hence the quote in my sig.

That's not true at all, you've also got to actually win...

And more importantly there is a causal relationship between success and how widely played a deck is...

[Crimhead]

That's not true at all, you've also got to actually win...

An average deck by definition does win - 50% of its matches. Each such deck that enters has an, eg, 1/256 chance of finishing 8-0. The more entrants there are, the more of these decks will be expected in the field. If 45-60 people out of 300 (15-20%) are playing an average deck, you can bet money you'll see that deck in the finals - even if it's a little worse than average. You claim to have a background in stats, so this should be very basic for you.

But you haven't offered an alternative explanation that accounts for the fact quoted in my sig. If you insist on rejecting (and ignoring) the philosophy behind the DTBF, maybe you should refrain from referencing the data it produces?

[GundamGuy]

An average deck by definition does win - 50% of its matches. Each such deck that enters has an, eg, 1/256 chance of finishing 8-0. The more entrants there are, the more of these decks will be expected in the field. If 45-60 people out of 300 (15-20%) are playing an average deck, you can bet money you'll see that deck in the finals - even if it's a little worse than average. You claim to have a background in stats, so this should be very basic for you.

But you haven't offered an alternative explanation that accounts for the fact quoted in my sig. If you insist on rejecting (and ignoring) the philosophy behind the DTBF, maybe you should refrain from referencing the data it produces?

The alternative explanation is that casualtity runs the other way, and the reason Miracles is the most played deck is because it's also the best.

You keep down playing this but perhaps a 1/256 chance of going 8-0 is actually far above the expected value with other decks....

Your logic is that Miracles is actually only Good because 15-20% of the field plays it... and if only 2-3% of the field played it it wouldn't be a top deck anymore... where as my explination is that 15-20% of the field plays it because it's been shown repeated success for months and months and months, and was succesful even when only 2-3% of the meta was Miracles.

I also think 1/256 is an underestimate, which assumes that every match is a coinflip. We know that's not actually the case, Miracles best matchups are well over 50/50, but it's worst matchups are pretty close to 50/50 (with the execption of a few fringe decks that don't see much play for various reasons)

[Crimhead]

The alternative explanation is that casualtity runs the other way, and the reason Miracles is the most played deck is because it's also the best.

That doesn't explain at all how a deck can be DTB but not tier one. You keep avoiding this!

You keep down playing this but perhaps a 1/256 chance of going 8-0 is actually far above the expected value with other decks....

1/256 is dead average. Are you really trained in statistics?
Think about it. If 256 players go eight rounds, exactly one deck will finish 8-0. An average deck by definition has 1/256 chance of being the winner. Any above average deck will have a better shot, and a bellow average deck will have a lower chance.

Also think 1/256 is an underestimate, which assumes that every match is a coinflip. We know that's not actually the case, Miracles best matchups are well over 50/50, but it's worst matchups are pretty close to 50/50 (with the execption of a few fringe decks that don't see much play for various reasons)

I wasn't talking about Miracles. I was talking about a hypothetical "average" deck, and explaining how such a deck would produce big numbers if it was played in big numbers.

But Miracles really only has one very good tier one pairing, being Elves.

Your logic is that Miracles is actually only Good because 15-20% of the field plays it... and if only 2-3% of the field played it it wouldn't be a top deck anymore... where as my explination is that 15-20% of the field plays it because it's been shown repeated success for months and months and months, and was succesful even when only 2-3% of the meta was Miracles.

I'm saying that if the deck is really as good as you think, 15-20% of the field playing it would result in more than 15-20% of tops! That's the difference between us - you think big numbers in the tops means a strong deck regardless of the number of those decks in the field. It makes me question your background in stats, that's for sure!

Basically what your saying is that it is not mathematically relevant that a high percentage of the field is playing Miracles because the fact that so many play it is further evidence that the deck must be good! Am I reading you right? Because there are two flaws in that logic!

Again, when we are crunching these numbers, why does it even matter why people play what they do? Even if I agreed that people play it because the have good reason to believe it's the best - even then I would expect it to put up twice as many tips as an equally good deck with half as many players.

[GundamGuy]

That doesn't explain at all how a deck can be DTB but not tier one. You keep avoiding this!

That all depends on how you define Tier 1, the whole concept of Tiers is pretty arbitrary. The short answer to this question however is this is would be pretty rare. How it would happen is if a deck like Imperial Painter snuck into a handful of Top 8ths by preying on the meta one month, it could wind up on the decks to beat for a short period of time even though it's not actually a Tier 1 deck, because it can't do so consistantly. But again this all depends on how you want to arbitrarly decide what Tier 1 vs Tier 2 vs Tier 3 is... I don't think heading down this rabbit hole is of much use....

1/256 is dead average. Are you really trained in statistics?
Think about it. If 256 players go eight rounds, exactly one deck will finish 8-0. An average deck by definition has 1/256 chance of being the winner. Any above average deck will have a better shot, and a bellow average deck will have a lower chance.

I wasn't talking about Miracles. I was talking about a hypothetical "average" deck, and explaining how such a deck would produce big numbers if it was played in big numbers.

The problem with your 1/256 calculation is that there are actually 3 possible outcomes to any Magic Match, not two, and each outcome offers a different amount of Swiss Points, so it's not actually correct to count a draw as a loss, even though it's not a win. It's better then a loss. In an event with 256 players it is indeed possible for no one go to 8-0.

But ignoring that for a moment, lets play around with this "hypotheical 'average' deck" some.

Suppose A hypotheical deck has a 1/256 calculation of being the winner. If 32 people play such a deck then that deck has a 1/8 chance of being the winner, right? No that's actually not correct. That would be correct if we were talking about pulling balls out of a bag, the chance of pulling one of 32 black balls from a bag of 256 balls with a random pull is 1/8... the problem is Swiss isn't a random pull from a bag. Each outcome isn't actually equally likely. And the density of these outcomes with different likely hoods actually changes the base probability. In short this is a far more complex problem then you are attempting to make it.

But Miracles really only has one very good tier one pairing, being Elves.

It's not like you only ever play Tier one decks on your way to the top 8....

I'm saying that if the deck is really as good as you think, 15-20% of the field playing it would result in more than 15-20% of tops! That's the difference between us - you think big numbers in the tops means a strong deck regardless of the number of those decks in the field. It makes me question your background in stats, that's for sure!


Basically what your saying is that it is not mathematically relevant that a high percentage of the field is playing Miracles because the fact that so many play it is further evidence that the deck must be good! Am I reading you right? Because there are two flaws in that logic!

Incorrect for a lot of reasons. Reason number 1, each pilot of the same deck doesn't have the same probability of success, there is varriance inherent in the game of magic, and skill that makes this translation not so clear. Second the probabilities shift depending on the distribution of decks within the field, not just the number of people playing that deck.

And no, I am not saying it is not mathematically relevant, what I am saying is your overstaing the relevance. In a way what you are saying is that it doesn't really matter how good a deck is, as long as enough people play it... what I am saying is that looking at real statistics and probabilities for a moment, more players playing the same deck doesn't inhrently mean more success. I'm sure you didn't mean to imply as much, but for that to be true, Magic would have to be entirely a game of luck, involving no skill.

Do you actually think that Burn would have the highest win percentage if 40% of the field played Burn... (No, because decks that eat burn for lunch, would dominate them) being the most played deck is both a blessing and a curse when it comes to DTB... it means that people are building there decks to beat you... and making meta decisions with you in mind... (which doesnt' really matter to Miracles since it's the honey badger of decks...) so the question is which will have a bigger impact? The fact that your a larger precent of the field means you should get more top 8's... but it also means your a bigger target for getting metagamed.