[Article] Unlocking Legacy - Investigating Fetchlands

[Zach Tartell]

Link (http://www.starcitygames.com/php/news/article/15815.html)

Christopher Coppola analyzes the deck thinning properties of fetchlands and dispels some misconceptions about their effects.

In other news, Zach Tartell continues to be the only Legacy player who reads Star City.

[Bryant Cook]

I actually liked this article. It's not on Threshold or some crappy deck. Good job.

[Cait_Sith]

Informative, but the topic has been beaten to death. Many times over.

[Bryant Cook]

Informative, but the topic has been beaten to death. Many times over.

I disagree, find me articles with the same topic. People need to quit reading Unlocking Legacy articles with the mind set that before they read the article that it's going to be awful.

[Cait_Sith]

I mean in general. This has been a long on going debate over how important are fetchlands. An article like this is certainly nice, but it feels more excessive than anything else.

[Nihil Credo]

The SCG forums actually had a useful (albeit obvious) point to make! I'm shocked.

[Bovinious]

I actually really liked this one, the only suggestion I would have is showing how you calculated some of those percentages, although I realize this would be long/complex, and also probably go over 80% of the readers' heads, so I can see why they werent included also. Good job :wink:

[BreathWeapon]

The article was pertinent to both Legacy and Extended, so I didn't really see it as a Legacy article so much as a math article, but I think for people who haven't bothered to investigate the issue for themselves it was reasonable subject matter to re-hash.

Well written if nothing else.

[Phantom]

I also really enjoyed the article. Short, sweet, with important info. I also always suspected that the myth of fetchlands not affecting future draws was just that, and am glad to see it bear out in numbers. The only two things I would have liked to see is some sort of nod to the deckbuilder (perhaps some advice on choosing your number of fetches) and a section like 'the dangers of fetches'.

Still, solid, solid work.

[TheAardvark]

What "2003 articles" were you referring to (regarding deck thinning), Chris? Or did the article say and I totally no-sold?

Anyway, I suggest anyone who is interested in the thinning effect of fetchlands to read this (http://magic.tcgplayer.com/db/print.asp?ID=3096). It involves an MC simulation, for those who care.

[Tog]

Thank you. Something besides Threshold or Landstill. =]

[AngryTroll]

I thought it was a great article- I hope we see more like this in the future. It applied to Extended, Legacy, and even a little bit to Vintage, was easy for someone that hasn't played Legacy before to read (and showed off some cool cards we get to use that Type 2 doesn't -Brainstorm, Fact or Fiction).


I would be very interested to know more about how (if) you used a Monte Carlos analysis on the probabilities. I've gotten familiar with programming Monte Carlo in MATLAB; I'd be interested to see how it can be applied to the Fetchland question.

[freakish777]

I actually really liked this one, the only suggestion I would have is showing how you calculated some of those percentages, although I realize this would be long/complex, and also probably go over 80% of the readers' heads, so I can see why they werent included also. Good job :wink:

Why not have Chris produce the math here?

[T is for TOOL]

The article can be summarized very succintly.

1. Deckthining is good (numbers)
2. Fetchlands thin your deck (more numbers)
3. Therefore, running max Fetchlands is good.

The biggest controversy surrounding fetchlands as deck-thinners has always occured when comparing the benefits of deckthining to lifeloss, yet this issue was completely ignored in the article.

[Willoe]

I second that. I prefer fetches in a non-thrash heavy meta (which is my meta :P) or simply a meta without too many stifles. However, like T is for Tool pointed out, the life loss was completely ignored. 1 life may not be that much, but 2 life is -1 to the opponent's storm count. I think that this is a point that at a minimum should be considered.

[Bryant Cook]

I second that. I prefer fetches in a non-thrash heavy meta (which is my meta :P) or simply a meta without too many stifles. However, like T is for Tool pointed out, the life loss was completely ignored. 1 life may not be that much, but 2 life is -1 to the opponent's storm count. I think that this is a point that at a minimum should be considered.

I love it when control players fetch 3 times, that's prime for the time to combo off. They Force Chant, then play the rest of your hand with Tendrils in it.

[Tacosnape]

In other news, Zach Tartell continues to be the only Legacy player who reads Star City.

[Jander78]

The biggest controversy surrounding fetchlands as deck-thinners has always occured when comparing the benefits of deckthining to lifeloss, yet this issue was completely ignored in the article.
I honestly think this is the only valid argument when discussing whether or not to include fetchlands. We all know that they help thin the deck, smooth out draws, and should be an auto-include in multicolored decks. It's all questionable and opinion whether or not the life loss makes it a valid option in certain decks.

[Nydaeli]

The article's main flaw is that it only seriously considers running 0 or 8+ fetchlands. Their drawbacks - life loss, and their tendency to turn into useless Mountains in this format - tend to encourage moderation, and I would have liked to see some discussion of how to balance fetchlands' benefits and drawbacks.

[Machinus]

I'd be happy to talk about lifeloss but I haven't heard any coherent questions on it, just "lifeloss bad!" But many good decks run cards that say "pay X life draw Y cards" and we think of those as good. Trading life for card is a desirable effect, and in this case there is no mana investment and it's undisruptable.


What "2003 articles" were you referring to (regarding deck thinning), Chris? Or did the article say and I totally no-sold?

Anyway, I suggest anyone who is interested in the thinning effect of fetchlands to read this (http://magic.tcgplayer.com/db/print.asp?ID=3096). It involves an MC simulation, for those who care.

Yes.

[Nihil Credo]

I'd be happy to talk about lifeloss but I haven't heard any coherent questions on it, just "lifeloss bad!" But many good decks run cards that say "pay X life draw Y cards" and we think of those as good. Trading life for card is a desirable effect, and in this case there is no mana investment and it's undisruptable.
Oh, come on, you know perfectly well that it's only true for some values of X and Y.

In the example you bring up in the article, an aggro player with fetchlands is paying 6 life for an extra card. There are zillions of scenarios where this is a bad trade, and Orlove brought up a few common ones in the forums. A few life points can cost games just as much as a bad topdeck, and deciding which happens more often is by far the biggest issue in deciding how many fetchland to run beyond the minimum needed for colour stability.

[BreathWeapon]

The article can be summarized very succintly.

1. Deckthining is good (numbers)
2. Fetchlands thin your deck (more numbers)
3. Therefore, running max Fetchlands is good.

The biggest controversy surrounding fetchlands as deck-thinners has always occured when comparing the benefits of deckthining to lifeloss, yet this issue was completely ignored in the article.

I'm not certain it's Deck Thinning vs Damage so much as Deck Thinning + Shuffling vs Damage, I run maximum blue fetchlands just for their synergy with Brainstorm, Ponder and Sensei's Divining Top. When you consider those 3 cantrips are 99% of the draw spells cast in this format, I think mix/maxing Fetchlands is kind of a given for Legacy.

[Machinus]

I think players who discuss this issue are far too afraid to lose life.

First of all, there is no card investment or mana investment here. It is an investment of life and time only. Considering life only, it is fair to pay more life for the same amount of cards because you aren't spending one in the first place. It's also fair to pay more life for the same amount of cards because you aren't paying any mana at all.

In a short game, the effect is very small and it's not as strong a trade since you are investing the life early. But in a long game the higher spell density will result in X more cards drawn. If you have any manipulation at all there will be substantial differences in your "virtual" card advantage.

Just imagine you are playing vanguard with the "Wooded Foothils" avatar. You could pay 1 life a few times during the game, and you would be more likely to draw spells afterwards depending on how much total life you had paid.

If it's not worth it to make this trade, then why do you let Lava Spike resolve?

[Mad Zur]

Was this article just aimed at Dragon Stompy players? Are there any other decks that can run fetchlands but don't already have a much more compelling reason to? If so, are any of those decks good?

Assuming it's even relevant to consider the deckthinning effect, there were other problems with the article.

First of all, I'd like to know the reasoning behind this claim:

Monte Carlo methods may be unnecessary for the analysis of this topic, and may even be worse than a classical approach given the variety of questions asked about different game states.
Why is it more useful to focus on the few specific examples you've created than to look at actual averages?

Let's look at those examples, which all turn out to be weighted in your favor.

Consider two Threshold decks that play seventeen lands – one with zero and one with eight fetchlands. Suppose all draw events are identical, and they both have thirty-five spells left in their decks. The Threshold deck without fetchlands has twelve lands left in its library, and the Threshold deck with fetchlands has nine left. On the play, on turn five, if both decks cast a Brainstorm hoping to fetch away the extra land they drew, what will happen?
We can deduce that, in your example, both decks have drawn 13 cards, 5 of which are lands, 3 of which are fetchlands in the second case. On average, in 13 cards, a deck with 17 lands will draw [(17/60)*13 =] 3.68 lands, including [(8/60)*13 =] 1.73 fetchlands. You have the deck drawing an above average number of fetchlands to exaggerate the deckthinning effect.

Consider two Landstill decks that play twenty five lands – one with zero and one with eight fetchlands. Again, suppose everything that happens in the game is the same, except the second deck has fetched four lands out of its library. The first deck has seventeen lands remaining in its library, and the second deck has thirteen. If both decks have twenty-five spells remaining, what will happen if they both cast Fact or Fiction on turn eight?
Each deck has drawn 18 cards, 8 of which are lands, including 4 fetchlands. Again, this is exaggerated. In 18 cards, this deck can be expected to draw [(8/60)*18 =] 2.4 fetchlands.

Consider two Aggro-like decks with twenty-two lands, one with zero fetchlands and one with ten, playing against each other. At some point in the late game, the decks have reached a stalemate since they cannot kill each others’ Tarmogoyfs, and each is hoping to draw out of it. Suppose we are on turn ten, and the second deck has fetched six times. How many turns will it take for the deck with fetchlands to draw one more spell than the opponent?
There are no details in this example about how many cards have been drawn so far or what's left in the deck, but it would take, on average, [6/(10/60) =] 36 draws for such a deck to draw 6 fetchlands.

In short, setting up examples with above-average numbers of fetchlands drawn in order to establish that running fetchlands can be expected to yield significant results is somewhere between unconvincing and dishonest.

I'm also sick of vague criticism. If someone else made a mistake when analyzing this issue, I want to know who they are, what they did, why it was wrong, and why your analysis is better. If you can't show me that, or even point me to what you're talking about ("a pair of articles written in 2003" is not helpful), I'm not going to be inclined to agree with you that your analysis is better. I'm going to wonder why you don't have specific complaints and why you don't even want me to know what other analysis is out there, and most of the likely answers to that do not involve you being correct.

My other objections have already been mentioned, primarily that you didn't address any of the actual objections to deckthinning (I don't think anyone contends that it is literally not real, I think they contend that it isn't worth it).

Edit: There are many situations in Magic where it is incorrect to trade some amount of life for some amount of cards, even when there is no other cost. I've chosen not to keep Sylvan Library cards an awful lot of times. Unfortunately, to answer that kind of question, you have to actually look at the specifics of the situation, not just say "trading life for card is a desirable effect."

[AnwarA101]

I also wished there was some discussion of drawbacks. I'm mainly considering cards like Stifle and Magus of the Moon. These might sound narrow, but I've actually lost to these cards because I had fetchlands instead of real lands. While this didn't make me not play fetchlands, it still means that there is a drawback.

[T is for TOOL]

In a short game, the effect is very small and it's not as strong a trade since you are investing the life early. But in a long game the higher spell density will result in X more cards drawn.
There is no myth, as you suggest in your article, that people don't believe this to be true. The Garrett Johnson study that you reference was written for the 'context of an aggro, usually mono-colored deck' and as such his conclusions apply to decks without library manipulation and the need for color fixing. The reason that people are hesitant to run the maximum number of fetches in Legacy decks includes not only lifeloss but also manabase vulnerability and many other reasons. Maybe these fears are unfounded and maybe not, but such an analysis was noticably absent from your article.

[Nihil Credo]

I think players who discuss this issue are far too afraid to lose life.

I disagree. Moving on.


First of all, there is no card investment or mana investment here. It is an investment of life and time only.No-one claimed otherwise.


Considering life only, it is fair to pay more life for the same amount of cards because you aren't spending one in the first place. It's also fair to may more life for the same amount of cards because you aren't paying any mana at all."Pay 19 life: Draw a card". According to what you just said, this is fair.


In a short game, the effect is very small and it's not as strong a trade since you are investing the life early. But in a long game the higher spell density will result in X more cards drawn.Both the life loss and the improvement in draw quality increase with the number of fetchlands you sacrifice (which generally correlates to how long the game goes). This is a non-point.


If you have any manipulation at all there will be substantial differences in your "virtual" card advantage.Agreed. The deck thinning benefits of fetchlands are more visible the more cards you see in a game. This still doesn't support your point that paying life for cards is always a good trade.


Just imagine you are playing vanguard with the "Wooded Foothils" avatar. You could pay 1 life a few times during the game, and you would be more likely to draw spells afterwards depending on how much total life you had paid.So, you're picturing the following avatar: "Hand +0, Life +0; Pay 1 life: Remove a land card in your library from the game"?

Nobody would activate its ability outside of corner cases; say, when playing against Solidarity, or when digging for answers against a non-damage-based game-winner like Counterbalance or Humility. Assuming that answer isn't Krosan Grip, because then you don't care as much about your opponent drawing counters in the meanwhile, and depending on the clock you're on it can be a better deal to save your life points to buy yourself extra turns to have more shots at drawing that Grip.


If it's not worth it to make this trade, then why do you let Lava Spike resolve?If Lava Spike dealt 6 damage, I would often counter it.

[kirdape3]

The conditions where running the maximum number of fetchlands in a deck would be undesirable are as follows:

1. Existence of decks that put incremental pressure on your life total - Red Deck Wins and strong traditional attack decks mainly. In Legacy, do we even really bother playing decks like this? R/g Sligh exists but that's about the only deck that wants to win like that - Threshold is going to want to hit you with a Tarmogoyf four or five times and every other deck that closes through the attack phase is either going to do 40 to you or not going to care how many turns it takes.

2. Decks that can kill all of the lands that you have, thereby making remaining fetchlands worthless. If you have say 12 fetchlands and 8 real lands (like Zoo did in Extended last season), and they kill a bunch of your lands, those fetchlands are blanked. There is Wasteland in this format, and some of the decks can't afford to have only 10 or 11 mana producing lands in their deck. Even Threshold doesn't want lose mana-producing lands.

Neither of those conditions happens all that often in Legacy, but combined they might happen often enough to at least give a competent deckbuilder pause before blindly shoving eight (or 12) fetchlands into their deck.

[Phantom]

Aren't there like, a ton more considerations when running fetchlands than those two??? Sure, some show up less than others, but off the top of my head:

Blood freaking Moon
Stifle
Needle
Shit that makes your shit come into play tapped (Root maze maybe?)
Ankh of mishra and Zo Zo the (Yo Yo) Punisher or whatever
Storm count
Other stuff in you deck that makes you lose life like oh say Fow, Ancient Tomb, Confidant and Thoughtseize
Everything I'm forgetting


I still think fetches are the bees knees in multicolored decks, but with all the variables that go into the choice, isn't it tough to quantify?

[Machinus]

I used examples that would highlight the effect of fetchlands because it's much easier to measure the effect if the fetching is basically a singular event. The examples aren't "weighted," they are just demonstrating the effect of the cards. Real life situations are more complicated because there are many more things happening, like fetching occuring throughout the game and brainstorming and so forth, and there are of course too many different game states to examine.

I hope it's obvious to everyone that a deck is likely to draw less than six fetches by turn ten, etc. That's really not the point of the example.

Also, it's pretty easy to find the referenced articles for yourself. If you think I was being misleading by omitting the reference then I'll just have to take that. But the real reason was that I was just being polite, as initially I was going to talk about them more directly and decided to not do that.

[frogboy]

I hope it's obvious to everyone that a deck is likely to draw less than six fetches by turn ten, etc. That's really not the point of the example.

So what is the point of the example? In your average game if you use a few fetchlands it will have a very small effect on the number of spells you draw?

[TheInfamousBearAssassin]

I used examples that would highlight the effect of fetchlands because it's much easier to measure the effect if the fetching is basically a singular event. The examples aren't "weighted," they are just demonstrating the effect of the cards. Real life situations are more complicated because there are many more things happening, like fetching occuring throughout the game and brainstorming and so forth, and there are of course too many different game states to examine.

"Of course I'm not really saying that Obama would turn out to be a member of Hizbollah and help terrorists bring in nukes to blow up your cities to pave the way for underpriced Mexican labor. I'd hope everyone would realize that. I'm just highlighting the possibility that it would happen to make a stronger case for why Clinton's a better candidate."


No dice. If you wanted to prove a valid point, you should use valid argumentation, not biased samples.

[Mad Zur]

I used examples that would highlight the effect of fetchlands because it's much easier to measure the effect if the fetching is basically a singular event. The examples aren't "weighted," they are just demonstrating the effect of the cards. Real life situations are more complicated because there are many more things happening, like fetching occuring throughout the game and brainstorming and so forth, and there are of course too many different game states to examine.
This makes sense if you simply wanted to demonstrate that deckthinning is not a myth. If I had come into the article thinking that there was literally no difference in average draws between two decks with different numbers of lands, I'm sure it would have convinced me. But I think your point is more than that; you're arguing that including fetchlands in your deck yields significant gains because of deckthinning. When we're discussing the significance of the effect (as opposed to the mere existence), we want to consider how much of an effect we expect to get, not how much of an effect we can get in a specific unlikely situation.

I hope it's obvious to everyone that a deck is likely to draw less than six fetches by turn ten, etc. That's really not the point of the example.
If the point is "Deckthinning is not imaginary," that's fine. If the point is "Running X fetchlands will let you draw an extra spell by turn Y," the example is invalid.

Also, it's pretty easy to find the referenced articles for yourself.How do I know which articles you're talking about? If I find an article from 2003 that mentions deckthinning, should I assume that it's one of the ones you have complaints with? Isn't that kind of unfair to any articles concerning deckthinning that you don't have complaints with?

If you think I was being misleading by omitting the reference then I'll just have to take that.
You could provide the reference here. I still want to know.

But the real reason was that I was just being polite, as initially I was going to talk about them more directly and decided to not do that.
This is arguably being polite to those authors, but it is being exceptionally impolite to your readers. It implies that either we are unqualified to judge whose analysis is better, or there's no reason we should want to. If the article TheAardvark linked to is one of the ones you were talking about, I'd like to know what problems you have with it. At the moment, I'm more impressed with that analysis than yours, but if you could provide (or if you had provided in the article) specific reasons that yours is better, I could easily be convinced.

[GreenOne]

I'd be happy to talk about lifeloss but I haven't heard any coherent questions on it, just "lifeloss bad!" But many good decks run cards that say "pay X life draw Y cards" and we think of those as good. Trading life for card is a desirable effect, and in this case there is no mana investment and it's undisruptable.


The other cards that say "Pay X life, draw Y cards" that see some playing have mana requirements and draw a savage amount of cards (Infernal Contract and Cruel Bargain).
The only card comparable I recall is Street Wraith. And FT, the only deck playing it, has dropped it lately. Noone will play it only for the thinning effect, but for some interaction with the deck (mystical tutor, threshold, goyf pump or the like).
Same thing about fetchlands. Obviously they help manabases in >1 color decks, but they're almost always played for something that it's not the thinning effect, but the shuffle or the grave effect (Brainstorm, Ponder, SDT, Threshold, Goyf pump, Crucible, Loam, Grim Lavamancer...).

The thinning effect is negligible unless you are planning to see a lot of cards during the game (with draw spells, staying alive till turn 20, Ringleader and the like,...)

[Illissius]

What Phantom said. We could (and, it seems, we do) argue endlessly about whether the life loss or the deck thinning is more negligible, but the popularity of Stifle makes the whole argument pointless. Paying one life to minimally increase your odds of drawing a spell may or may not be worth it, but if it also means giving your opponents occasional Time Walks, it's sure as hell not.

[SpatulaOfTheAges]

2. Decks that can kill all of the lands that you have, thereby making remaining fetchlands worthless. If you have say 12 fetchlands and 8 real lands (like Zoo did in Extended last season), and they kill a bunch of your lands, those fetchlands are blanked. There is Wasteland in this format, and some of the decks can't afford to have only 10 or 11 mana producing lands in their deck. Even Threshold doesn't want lose mana-producing lands.

This would really only matter to decks that plan on seeing a late, late game.

[FoolofaTook]

It's a decent article. I'd have liked to see some discussion of how Fetchlands are best used. Do you want to pre-emptively thin in the absence of a Brainstorm or SDT in hand or do you want to wait to get the chance to turn chaff into a better hand a few turns down the road?

When you have a multi-land and a fetch in hand which do you want to play first? The land that can cast your current spells or the land that can put whatever mana you need in play in a few turns?

Fetches create more basic play questions, and earlier, than any other card.

[Lego]

and in this case there is no mana investment and it's undisruptable.

Magus of the Moon, Pithing Needle, and Stifle, among others, all call bullshit.

[HdH_Cthulhu]

I disagree. Moving on.

So, you're picturing the following avatar: "Hand +0, Life +0; Pay 1 life: Remove a land card in your library from the game"?



Who needs manaless Belcher?^^

I think it is very easy to deside how much fetchlends you should use in deckbuilding.

There are only a few decks where you couldnt clearly say with or without the fetches like mono red Goblins

For me whene it only comes down to deck thinning i dont use fetches...

[Curby]

There are only a few decks where you couldnt clearly say with or without the fetches like mono red Goblins

For me whene it only comes down to deck thinning i dont use fetches...

The question of fetches also came up in a monogreen elf thread on mtgsalvation. Something else to consider is that if you're going to be recurring Elvish Messenger et. al., you're going to stack a bunch of land at the bottom of the deck, and cracking a fetch at that point will mix the land-heavy bottom with the gas-heavy top. In this case, the deck is thinned of 1 land, but the topdeck is made worse overall. On the other hand, once you start recurring Messengers you're about to win anyway, so it's probably not too significant. =P


What if we don’t even take into consideration draw spells? Drawing multiple cards at once does make a thinned deck much more noticeable, but the same effect will happen in a long game with many draw steps.

Consider two Aggro-like decks with twenty-two lands, one with zero fetchlands and one with ten, playing against each other. At some point in the late game, the decks have reached a stalemate since they cannot kill each others’ Tarmogoyfs, and each is hoping to draw out of it. Suppose we are on turn ten, and the second deck has fetched six times. How many turns will it take for the deck with fetchlands to draw one more spell than the opponent?

The answer is ten turns. After ten turns, the first deck will have drawn 6.36 spells, and the second deck will have drawn 7.37 spells. I have played many magic games at least this long in tournaments, and it is not unreasonable to conclude that with correct play, the deck that draws more threats or removal in this situation has a better chance of winning.

This is extremely suspect. He's arguing that 6 fetches will be cracked by a single deck in 10 turns without library manipulation (e.g. Brainstorm). In other words, he's expecting to find 6 out of 10 fetches in 16-17 cards? The chances of drawing over half of the number of a given card after drawing less than a third of the deck is astonishingly small, and this puts the rest of his argument (the third quoted paragraph) into question because you're basing the argument on an unrealistic set of initial conditions, namely 6 fetches in 10 turns.

By the way, by cracking 6 fetches more than your opponent, you've lost nearly a third of your life just due to your own land. This is as likely to make a difference in the game as the minor amount of deck thinning that it does.


I hope it's obvious to everyone that a deck is likely to draw less than six fetches by turn ten, etc. That's really not the point of the example.

Examples of unlikely possibilities do not make a convincing argument. You must look at likelihoods and averages. If you truly believe that using fetches helps thin the deck in a meaningful way, why can't you provide realistic examples that support your argument? If your example is not very realistic, at least point to its assumptions, and why they might not always be applicable. I'm hoping that the point of the article is to provide a balanced look at fetchland thinning, and when they are useful for that purpose, not a blanket statement that fetches are useful thinners with a sweeping dismissal of all possibilities to the contrary.

I'm curious why you arrived at a different conclusion than this article, which likewise does not look at individual drawing events, but rather the chances of drawing more gas over the course of a game: http://magic.tcgplayer.com/db/print.asp?ID=3096

You dismiss Monte Carlo methods as "may be unnecessary" and "may even be worse" but that's not exactly a convincing argument. Is it "may be" or "is definitely"? And why? I respect your experience and insight into the game, but rigor is required when you're talking about hard math, and that seems a bit lacking in this article.


I have played many magic games at least this long in tournaments, and it is not unreasonable to conclude that with correct play, the deck that draws more threats or removal in this situation has a better chance of winning.

This is another quote that lacks true information. You play in a lot of tourneys, so obviously you've played a lot of long matches. With all of the combo decks in Legacy, and the fundamental turns on many decks being between 1-3 (combo) and 3-5 (aggro) it's "not unreasonable to conclude" that the average game of Legacy takes a lot less than the 20 turns claimed (10 to set up your questionable initial conditions, and 10 to realize the extra spell drawn after that point).

Of course long games CAN happen, but that's like looking at individual draw events, an analysis you dismiss. Instead, how often do you crack 6 fetches in 10 turns, and how often does the game progress a further 10 turns after that for you to draw an extra card? No one ever argued that fetches don't thin the deck at all. The argument is that it's not a meaningful change to the deck, over the course of a typical game. That combined with the metagame hate like Stifle and inherent life loss make fetches a bad choice when used solely to thin the deck.

Overall, this article is interesting to me because it bucks the established belief that fetches are not statistically meaningful when included only to provide deck-thinning. It's a shame that it doesn't provide a more convincing argument to supports its thesis. It would be great if you wrote a part 2 that further defends these new insights against the conventional wisdom.

[Anusien]

On a Top8Magic podcast, Mike Flores was playing a RG deck (maybe Zoo) while being watched by the deck's creator, Mark Herberholz. Flores didn't want to fetch, but Herberholz made him. He said something to the effect of, "So what if the added 'spell probability' is incredibly minor? It's non-zero, and we need every advantage we can get." Considering the way Flores talks, it was more like, "Drawing lands is death for the deck, so do everything you can to not draw lands."
I think more decks should run Sensei's Divining Top than they currently do, and I think more decks should run fetchlands than they currently do. Here's a way to see; what average life total are you winning your games on? If it's more than 1 and you currently are not running fetchlands, you can afford to run some.

I think you were on the right track, but the ending is sort of abrupt. I mean, I love me some fetchlands, and I think you were pretty much dead-on with the article. What you wanted to say and what you said was good. But I think you should acknowledge the downsides of fetchlands. Also, if you're talking about the famous article on the mathematics of fetchlands, you need to rebut that head-on or you look like you don't know what you're talking about. By the way, saying that Monte Carlo simulations are wrong or unnecessary makes you sound very silly. Monte Carlo, in this case, is intended to simulate playing Magic.

That said, the average game seems to be somewhere between 6-8 turns, with 5 and 9-10 as either really high or really low. So you're talking about a free card once per match, rather than per game.

Then again, free card!

The articles I think he's referring to, thanks to Ben Goodman:
http://magic.tcgplayer.com/db/article.asp?id=3096
http://magic.tcgplayer.com/db/article.asp?id=3121

[Apex]

You talked about the 2003 articles out of the context of its discussion. Things taken out of context will often mislead readers.

Those articles by Garrett Johnson talked about Fetchlands strictly in the question of "if you run a mono-coloured deck (like a mono-black aggro deck), then is it better to run 20 swamps or 12 swamps and 8 fetches"? In the end, he concluded that if you ran 8 fetches, it allowed you to "...only see a 7.2% chance at turn 9 of getting an extra spell of some sort, and at an average cost of 2 life." So you would have paid 28 life on average to see exactly 1 extra card on the 9th turn over the course of many games.

He did not consider the fetchlands in the context of mana fixing, shuffling effects or graveyard filling. He specifically mentioned that his calculations dealt with the deck thinning effect of fetchlands in the context of a mono coloured deck.

I personally think it's not worth the 2 life to get a 7% increase probability to see an extra card that can only be realized by the 9th turn. I do believe however, fetchlands are very valueable given their other roles in decks. If you think however, that any increase in probability of seeing an extra card is worth the cost of 2 life (even if it's only realized by the 9th turn), then that's fine and all, but that would remain an opinion, and not a conclusion derived from scientifical analysis.

Also, this is just a personal thing, since I'm a university student statistics major, I am more inclined to believe numerical data and definite calculations more than I would believe opinions, so if you had actual test numbers, I would tend to believe you more. Also, proper referencing is required in all the articles I write and research, so I found it odd that you specifically mentioned several articles, yet failed to provide any link to them. Good thing I read them before, or else I would have not picked up on what you were talking about (it was rather hard to google for Garrett Johnson's articles from the given clues).

[Nihil Credo]

Errata corrige: the articles are by Garrett Johnson, not Ben Goodman. Ben (ridiculoushat) merely posted them in Machinus' feedback thread.

[TheInfamousBearAssassin]

Life and cards are both valuable resources that you need in order to win. Trading one for the other is only good if you're making a net profit towards your actual win percentages. This is the base argument at the heart of the Fetchland debate (and a number of other such debates), and no amount of hand-waving, catcphrasing or lip-curling condescension can make this fact go away. Any discussion of the advantages or disadvantages of a card that trades life for cards, or vice versa, which does not take this into account, is fundamentally useless and fraudulent.

[Anusien]

Errata corrige: the articles are by Garrett Johnson, not Ben Goodman. Ben (ridiculoushat) merely posted them in Machinus' feedback thread.
Well yeah, that's what I meant.

[Braves]

Personally I feel fetches are not that bad against a moon. Because play a fetch turn one, Dragon stompy or what ever gets turn one moon, you then can fetch a basic in response. Late game they suck, but getting one basic in a 3-4 color deck is a lot of times all you need to deal with the moon.

[DoNkEyBaLLs2000]

By the way, saying that Monte Carlo simulations are wrong or unnecessary makes you sound very silly. Monte Carlo, in this case, is intended to simulate playing Magic.

Using MC is pretty silly. There is no advantage to using it to calculate these probabilities.

And no method can "simulate playing Magic."

Really Chris? Did you think we wouldn't be able to tell this was you?

[bruno_tiete]

The funny thing is that when I started reading the article, I really though he was going the other way around. I mean, I often slap people for playing fetches in Burn, based on the fact that you are opening yourself to lose to a single stifle for trying to draw two extra cards per tournament, while paying a bunch of life. It seemed logical I would receive support from the article.

I can understand Machinus point in exagerated examples, specially because the life you pay in a game is directly correlated to the extra odds of business spells in that same game, so saying "your 6 fetchs example is non-sense" could be replied with a "if the example used a couple fetches, but the situation happened in 3 different games or so, results would be the same".

Whenever I get into this kind of discussion I try to find what are the accepted trade values. I mean, how much mana is worth an extra card? How much life is worth an extra card? How much life is worth mana? I am pretty sure this answers depend on the turn you are doing it, the possibility for recurring the effect and your game plan, among other factors.

Greed, the crappy black enchantment, will let you trade B and 2 life for a card and we are not seeing it anywhere. Off course there is the 4 mana cost to put it into play, but we would play Yawgmoth's Bargain if they let us, wouldnt we?

Back in the day (the one after Ancestrall Recall was born), a card costed 4 colorless. 1 life costed 1 mana, as seen in all of the triggered from color spell artifacts, Alabaster Potion and Stream of life. This combined into the 2 life, 2 mana activation tome, which name I cant recall for the life of me.

Today, we have over 9000 cards and the eternal power creep that results from diversifying options. We wont pay 4 life nor 4 mana for a card. We wont use Sylvan Library or Greed or anything with that conversion taxes. That's why I consider the trade offered by fetchlands a bad deal if you cant make good use of its other effects. In monocolored decks with no manipulation they are a bad deal.

As Garret wrote, if your deck can trade life for cards in that rate, it cant wait long enough to reap the rewards of it.
Its a side effect you will have to life with, and sometimes it will save you, but I know that more than once I died needing an extra mana with a fetch I couldnt activate in play.

I am probably writing shitty stuff AND beating a dead horse, though.

[Malchar]

After a few years of tournament experience in Legacy with mixed performance, I've made a helpful observation. In normal analysis, I would usually side with pure probability and ride the highest probability to victory. However, I think that Magic is so random that trying to manipulate probabilities for it is useless. Rather, I would prefer to maximize the number of options that my deck has. In terms of fetchlands, I would prefer to minimize the number of vulnerabilities of my deck.

Paying X life for Y cards is good depending on the X and Y values, but it also depends on your opponent's deck/strategy. To sum it all up, I think that tossing in fetchlands gives you a significant weakness against certain decks like Blood Moon, Burn, and Combo. As a goblins player, the choice of fetchlands is actually significant to me. I figure that I already lose to Burn and the occasional Combo, so I'd like to improve my matchups as much as possible. I figure that in any given tournament, I'll have to win through 5 or 6 rounds and I'll need to minimize my weaknesses. Fundamentally, the game isn't really about winning, but it's about not losing.

[Puzzle]

After a few years of tournament experience in Legacy with mixed performance, I've made a helpful observation. In normal analysis, I would usually side with pure probability and ride the highest probability to victory. However, I think that Magic is so random that trying to manipulate probabilities for it is useless. Rather, I would prefer to maximize the number of options that my deck has. In terms of fetchlands, I would prefer to minimize the number of vulnerabilities of my deck.

Paying X life for Y cards is good depending on the X and Y values, but it also depends on your opponent's deck/strategy. To sum it all up, I think that tossing in fetchlands gives you a significant weakness against certain decks like Blood Moon, Burn, and Combo. As a goblins player, the choice of fetchlands is actually significant to me. I figure that I already lose to Burn and the occasional Combo, so I'd like to improve my matchups as much as possible. I figure that in any given tournament, I'll have to win through 5 or 6 rounds and I'll need to minimize my weaknesses. Fundamentally, the game isn't really about winning, but it's about not losing.If that was true, life-gain would be king.
There is some truth in what you say but it all comes down to balance and metagame. For example, I'd most probably always play Lightning Bolt over Nourish but if there was a 1-mana spell giving you 10-15 life at instant speed, I'd consider it, or I could prefer Nourish in an all-burn meta.

[TheInfamousBearAssassin]

Personally I feel fetches are not that bad against a moon. Because play a fetch turn one, Dragon stompy or what ever gets turn one moon, you then can fetch a basic in response. Late game they suck, but getting one basic in a 3-4 color deck is a lot of times all you need to deal with the moon.

Or just use it as a mountain.


I think Stifle was the main argument against fetches.