Fetchlands: There are a number of more or less dubious reasons behind why people run fetches. Regardless of the motives, you always pay the price in self-damage and vulnerability to Stifle. Here are some of the reasons that often come up:
*Deck thinning: There is no such thing. It's been mathematically proven that alleged "deck thinning" through fetchlands is a false assumption. Never ever run fetches for this reason.
From the burn primer. Can someone please direct me to this proof, or just lay it out for me? This is the first I've heard of it and definitely something I want to understand.
"You know Manningham's story of the burgher's wife who bade Dick Burbage to her bed after she had seen him in Richard III and how Shakespeare, overhearing, without more ado about nothing, took the cow by the horns and, when Burbage came knocking at the gate, answered from the capon's blankets: William the conqueror came before Richard III."
That paragraph is misleading, but to sum it up what they are really trying to say is that running Fetch lands for the sole purpose of deck thinning is a bad idea because statistically more often than not the 1 point of life you spend is going to be of more consequence than any statistical advantage you get out of fetching, since the percentage is so low.
If you are using fetches for other reasons in your lists, like shuffles for brainstorm or fuel for grim lavamancer, any thinning that you get is just gravy anyways so no worries, just dont use them solely for thinning.
From the burn primer. Can someone please direct me to this proof, or just lay it out for me? This is the first I've heard of it and definitely something I want to understand.
Fetches do thin your deck. It's just that the thinning effect of fetchlands has almost no effect on your draws, as taking out one card from a 50 card deck is 2% of your remaining deck, which doesn't matter unless the game goes very very long then it will matter but if you're a deck like burn the gain is virtually nonexistant. That, and the 1 life is often more precious then the thinning fetches kind of provide.
Private Mod Note
():
Rollback Post to RevisionRollBack
"Yawgmoth," Freyalise whispered as she set the bomb, "now you will pay for your treachery."
I disagree with the conventional thinking here, if only slightly. The thing about the article everyone always cites is that it uses a muddled definition of "extra cards". For it, an extra card means replacing a land with a spell. That may be better or worse than drawing a card, depending on the deck. An extra card draw has some chance of being an island. That's acceptable, if a deck can use extra mana lands -- it has mana sinks or high cost cards, or it has ways to turn them into better stuff, like Brainstorm. On the other hand, turning a land into a spell won't get you any card advantage, but it may be a big enough improvement in decks that can't use lands after the first 3 or so.
When the Mathemagics article talks about trading life for cards, it makes you think about the trade off of things like Necropotence or Dark Confidant, which price cards at 1-2 life each. Those are card draws though, which (as I argued) may be better or worse than the effect of fetch lands to turn would-be lands into spells. Sign in Blood is my typical example here. In a 20 land deck with 8 fetch lands (as in the simulations from the article), Sign in Blood will net you an additional 0.368 non-land cards per casting. In decks that only want to draw spells after reaching a critical amount of mana, this is all that matters. With fetchlands, you accumulate this advantage by about turn 14, at a cost of 2.75 life and 0 mana. That's still not very good, but it's a price that's worth considering.
This is how I think about it: with a deck like burn or zoo, life does not matter in many matches. In matches where life matters, fetches hurt you a lot, as evidenced by the success of non-fetch builds in burn mirror matches. In matches where life doesn't matter -- the opponent stabilizes with life-gain, fatties, or enough draw and counters to stop everything you do, or the opponent is playing a combo that wins in one glorious turn -- fetches help a very little bit. In fact, they get you an extra non-land card by turn 6 (a cut off for relevance in fast decks) in one about one out of 10 games. Hypothetically, there's a point at which the tradeoff between the wins in the second type of match outweigh the losses in the first type. Via some loose estimates of these effects, I've guessed that a meta would need to be 90%+ decks of the second type before fetches were advantageous just for deck thinning purposes. This ignores, of course, the fact that fetches also open up vulnerability to Stifle. It's more relevant, though, for the occasional (but unlikely) Modern or Extended meta-game that meets that cut-off point.
Sign in Blood is my typical example here. In a 20 land deck with 8 fetch lands (as in the simulations from the article), Sign in Blood will net you an additional 0.368 non-land cards per casting.
In a 20-land deck, each card drawn has a roughly 2/3 chance of being a non-land card. Sign in Blood should therefore net you roughly 4/3 non-land cards per casting. This is much, much more efficient than trying to thin your deck with fetch-lands.
(The probabilities are only approximate because each card drawn is no longer in your library and therefore cannot be drawn again. However, since Sign in Blood costs more than a single mana, you must have multiple mana sources to cast it and therefore very slightly fewer {on average} lands left in the deck as a proportion of the whole deck than what you start with for your opening hand.)
Private Mod Note
():
Rollback Post to RevisionRollBack
I am the author of the "Resource Advantage in Magic" series over on EternalCentral.com
Sign in Blood should therefore net you roughly 4/3 non-land cards per casting.
The way I'm measuring it, that's not really possible, as Sign in Blood only nets you +1 card; so it certainly doesn't get you +1.33... non-land cards. The issue is that it costs you a non-land card, but you might draw only 0 or 1 non-land cards (or get only an additional 0 to 1 spells by the end of the game due to having drawn 2 cards). Fetchland thinning might get you additional non-land cards without also counting as one of your previous non-land draws. You have to count the fact that a card draw spell costs one of your non-land draws, in whatever probability distribution you get them, but is not itself a threat or an answer, which is what you ultimately want to accrue by casting it. In other words, it takes the position of a "business spell", so it only nets you resource-advantage (in comparison with just playing another threat or answer) if you ultimately get 2+ "business spells" by casting it.
From the burn primer. Can someone please direct me to this proof, or just lay it out for me? This is the first I've heard of it and definitely something I want to understand.
Its not that it doesn't thin the deck, but rather that the desired effect is almost negligible in comparison to not using any fetch's. So use them for mana fixing and nothing else.
Also, fetchlands give multiple and instant speed landfall, e.g. for Steppe Lynx. They also fill the graveyard with lands for cards that can benefit from that, like Nimble Mongoose, Knight of the Reliquary and Life from the Loam. Granted, most decks that run those also want some mana fixing, but these cards, and free shuffles post cantrip, are the main reason that some decks run 10+ fetches with only 6-8 target lands. The mana-fixing advantage from this set-up, versus having equal fetches and targets, is miniscule, and it leaves you more vulnerable to Stifle while helping aggro decks' game plan. The only good reason to go extreme with fetchland numbers are cards like the ones I mentioned above, plus Brainstorm and company.
I've always hated that article. People have misunderstood that you are still thinning your deck. Sure, you have to run multiple fetches for it to have a significant effect, but even that one fetch does get rid of a land that you would otherwise draw.
Statistically it might not make a big difference to do one or two fetches, but in individual games it can pay off significantly.
That's the difference your paying 1 or 2 life for the chance to not draw a land. Plus color fixing is always good etc, etc.
I've always hated that article. People have misunderstood that you are still thinning your deck. Sure, you have to run multiple fetches for it to have a significant effect, but even that one fetch does get rid of a land that you would otherwise draw.
Statistically it might not make a big difference to do one or two fetches, but in individual games it can pay off significantly.
That's the difference your paying 1 or 2 life for the chance to not draw a land. Plus color fixing is always good etc, etc.
the whole point of the article is that in a single game fetching has an almost negligible effects on your chances of not drawing a land. Even when running multiple fetches, it shows that you barely decrease the amount of lands you draw. I am unsure what other situation you thought it was talking about?
if the maths are correct,you have drawn, on average, half a spell more by turn 16 when 8/20 of your lands are fetches (having cracked ~ 3-4) vs. 0/20 being fetches. This 1 spell/ 2 games that last to turn 16 is often less relevant than say, stifle, or beating yourself up with fetches.
you WILL thin your deck. but the point of the article is that it thins your deck slightly. to the point where the 1 life you pay is probably not worth it. (not to mention the stifle bait fetches are in legacy)
so running fetches for, and only for, thinning is worthless. fetches have other uses, but thinning isnt an important one.
the whole point of the article is that in a single game fetching has an almost negligible effects on your chances of not drawing a land. Even when running multiple fetches, it shows that you barely decrease the amount of lands you draw. I am unsure what other situation you thought it was talking about?
if the maths are correct,you have drawn, on average, half a spell more by turn 16 when 8/20 of your lands are fetches (having cracked ~ 3-4) vs. 0/20 being fetches. This 1 spell/ 2 games that last to turn 16 is often less relevant than say, stifle, or beating yourself up with fetches.
you WILL thin your deck. but the point of the article is that it thins your deck slightly. to the point where the 1 life you pay is probably not worth it. (not to mention the stifle bait fetches are in legacy)
so running fetches for, and only for, thinning is worthless. fetches have other uses, but thinning isnt an important one.
You missed my point. The article is right in suggesting that statistically it does not make a big difference. But wrong in assuming that statistics carry over to each game. What I mean by this is that their will be games that fall out side of the curve. These are the games that we care about where it does make a difference.
You missed my point. The article is right in suggesting that statistically it does not make a big difference. But wrong in assuming that statistics carry over to each game. What I mean by this is that their will be games that fall out side of the curve. These are the games that we care about where it does make a difference.
The article is saying that you will encounter more games lost due to stifle or paying 3-4 life than you will encounter games won because of drawing 1 extra spell. The question is whether paying 2 life every game by turn 8 so that you draw an extra card in 1/10 of isn't going to increase the number of games you win for many decks. That 1 case where 2 life and a land card bought you a spell has to outweigh 9 cases of paying 2 life for no benefit. This isn't even considering the impact of stifle yet. Is 20 life spread over 10 games worth making 1 land card a spell. Life to cards is obviously powerful at 1 life to +1 card. Is it so powerful at 20 life to +1 spell - 1 land?
Fetchlands will tend towards 1 life for +1 spell -1 land as the game reaches turns 40-50. The problem is that the vast majority of games are going to be decided in turns 4-10 where you aren't getting anywhere close to that amount of value out of them.
To see the reality of fetchlands number your lands somehow. Use different art, write numbers on them, it doesn't matter. Then play games where when you fetch you choose a land number, say Island #2, and play as if the fetch were that land. Then act as though Island #2 automatically cycles when you draw it. See how frequently you just hit another land card from that cycling, see how infrequently you are even going to get to cycle Island #2. This simulation models exactly the results of fetchlands, in a way that you can see just how many times it would matter.
Is 20 life spread over 10 games worth making 1 land card a spell. Life to cards is obviously powerful at 1 life to +1 card. Is it so powerful at 20 life to +1 spell - 1 land?
You say this as if +1 spell -1 land is always worse, or at least no better than +1 card. This is the error for which I was critiquing the article. For some decks, i.e. decks with low curves, no additional uses for land, and no mana sinks, this isn't true. In such decks, trading a land for a spell is much more useful than drawing a card, which has a fair chance of being a useless land. With Sign in Blood, which is the example I always use, here, you gain about 0.368 resource or resource-trading spells, versus just playing a resource-influencing spell in Sign in Blood's place. For decks that only care about influencing the game state via low cost spells, that's the number that matters. It does this at a cost of 2 life and 2 mana, which is comparable to the influence of fetchlands through 14 turns (2.75 life and 0 mana). That's still, usually, too slow to make a difference, but it makes fetchlands seem a lot better in hyper-aggressive decks. Of course, the existence of Stifle makes this all a moot point in legacy.
U start with 23/60 lands/cards in a deck
draw a hand 3/7 are lands 20/53 are left= 37.74% chance ull draw a land next
with fetchland 19/52= 36.34% chance u will draw a land next
next turn assuming u dont draw a land it becomes 20/52 = 38.46%, and 19/51 = 37.25%
playing another fetch will bring it down to 18/50= 0.360% so it does mathamaticly thin down your deck.
(*its the 1% that make you a winner ;))
Private Mod Note
():
Rollback Post to RevisionRollBack
Standard: pft, i love to see my cards depreciate in value once they rotate :S
Mordern: Melira, UR Storm, RDW, Infect, W life/control, UW Tron
Legacy: RDW, Pox
Vintage: Dark Depths, R Grey Orge
23 lands is a bit steep to be honest. You'd be more likely looking at 18-22 for the average curve in the format. So let's say 20.
1/3 of your cards are lands. If you crack a fetchland before your draw on t2 if you're on the play, you're going to have 52 cards left in your library. If 18 of those are still lands, you have a 34% chance of drawing a land. If you do another, the turn after you have a 32% chance of drawing another land. Is a two percent increase enough to merit you playing fetchlands? In some competitive builds, any increase in your chances is great. But they don't make it significant enough for it to matter.
...unless you're playing crucible and exploration.
From the burn primer. Can someone please direct me to this proof, or just lay it out for me? This is the first I've heard of it and definitely something I want to understand.
Google would've yielded the same result.
Thanks to Heroes of the Plane Studios for the amazing sig.
NO RUG: Primer
Tempo Thresh: Primer
If you are using fetches for other reasons in your lists, like shuffles for brainstorm or fuel for grim lavamancer, any thinning that you get is just gravy anyways so no worries, just dont use them solely for thinning.
Http://www.fantasticneighborhood.com/
Comedy gaming podcast. Listening to it makes you cool.
Here you go.Mathemagics: Onslaught Fetchlands
EDIT: And that's why you don't open multiple tabs, you get ninja'ed.
Legacy:
UWRStoneblade
UWGBant Stoneblade
UWBEsper Stoneblade
BUGBUG
BUGTeam America
BUGDark Tempo Thresh
RUGTempo Thresh
BGWJunk
BGWThe Truffle Shuffle
WBDeadGuy Ale
GWMaverick
GWRPunishing Maverick
BGNic Fit
URSneak and Show
UBRWGThe Epic Storm
UBRWGDoomsday
UBReanimator
USolidarity
RGBelcher
UBWGRDredge
UWGBCephalid Breakfast
RBurn
Standard:
UWUB Control
RGRed Deck Wins
Currently Playing:
Retired
I disagree with the conventional thinking here, if only slightly. The thing about the article everyone always cites is that it uses a muddled definition of "extra cards". For it, an extra card means replacing a land with a spell. That may be better or worse than drawing a card, depending on the deck. An extra card draw has some chance of being an island. That's acceptable, if a deck can use extra mana lands -- it has mana sinks or high cost cards, or it has ways to turn them into better stuff, like Brainstorm. On the other hand, turning a land into a spell won't get you any card advantage, but it may be a big enough improvement in decks that can't use lands after the first 3 or so.
When the Mathemagics article talks about trading life for cards, it makes you think about the trade off of things like Necropotence or Dark Confidant, which price cards at 1-2 life each. Those are card draws though, which (as I argued) may be better or worse than the effect of fetch lands to turn would-be lands into spells. Sign in Blood is my typical example here. In a 20 land deck with 8 fetch lands (as in the simulations from the article), Sign in Blood will net you an additional 0.368 non-land cards per casting. In decks that only want to draw spells after reaching a critical amount of mana, this is all that matters. With fetchlands, you accumulate this advantage by about turn 14, at a cost of 2.75 life and 0 mana. That's still not very good, but it's a price that's worth considering.
This is how I think about it: with a deck like burn or zoo, life does not matter in many matches. In matches where life matters, fetches hurt you a lot, as evidenced by the success of non-fetch builds in burn mirror matches. In matches where life doesn't matter -- the opponent stabilizes with life-gain, fatties, or enough draw and counters to stop everything you do, or the opponent is playing a combo that wins in one glorious turn -- fetches help a very little bit. In fact, they get you an extra non-land card by turn 6 (a cut off for relevance in fast decks) in one about one out of 10 games. Hypothetically, there's a point at which the tradeoff between the wins in the second type of match outweigh the losses in the first type. Via some loose estimates of these effects, I've guessed that a meta would need to be 90%+ decks of the second type before fetches were advantageous just for deck thinning purposes. This ignores, of course, the fact that fetches also open up vulnerability to Stifle. It's more relevant, though, for the occasional (but unlikely) Modern or Extended meta-game that meets that cut-off point.
In a 20-land deck, each card drawn has a roughly 2/3 chance of being a non-land card. Sign in Blood should therefore net you roughly 4/3 non-land cards per casting. This is much, much more efficient than trying to thin your deck with fetch-lands.
(The probabilities are only approximate because each card drawn is no longer in your library and therefore cannot be drawn again. However, since Sign in Blood costs more than a single mana, you must have multiple mana sources to cast it and therefore very slightly fewer {on average} lands left in the deck as a proportion of the whole deck than what you start with for your opening hand.)
Check it out!
http://www.eternalcentral.com/resource-advantage-in-magic-part-1-one-shot-resources/
http://www.eternalcentral.com/resource-advantage-in-magic-part2-tempo/
I've also written a short primer on Manaless Dredge in Vintage:
http://www.eternalcentral.com/the-dredge-of-glory-an-introduction-to-manaless-dredge-in-vintage/
The way I'm measuring it, that's not really possible, as Sign in Blood only nets you +1 card; so it certainly doesn't get you +1.33... non-land cards. The issue is that it costs you a non-land card, but you might draw only 0 or 1 non-land cards (or get only an additional 0 to 1 spells by the end of the game due to having drawn 2 cards). Fetchland thinning might get you additional non-land cards without also counting as one of your previous non-land draws. You have to count the fact that a card draw spell costs one of your non-land draws, in whatever probability distribution you get them, but is not itself a threat or an answer, which is what you ultimately want to accrue by casting it. In other words, it takes the position of a "business spell", so it only nets you resource-advantage (in comparison with just playing another threat or answer) if you ultimately get 2+ "business spells" by casting it.
Its not that it doesn't thin the deck, but rather that the desired effect is almost negligible in comparison to not using any fetch's. So use them for mana fixing and nothing else.
Statistically it might not make a big difference to do one or two fetches, but in individual games it can pay off significantly.
That's the difference your paying 1 or 2 life for the chance to not draw a land. Plus color fixing is always good etc, etc.
the whole point of the article is that in a single game fetching has an almost negligible effects on your chances of not drawing a land. Even when running multiple fetches, it shows that you barely decrease the amount of lands you draw. I am unsure what other situation you thought it was talking about?
if the maths are correct,you have drawn, on average, half a spell more by turn 16 when 8/20 of your lands are fetches (having cracked ~ 3-4) vs. 0/20 being fetches. This 1 spell/ 2 games that last to turn 16 is often less relevant than say, stifle, or beating yourself up with fetches.
you WILL thin your deck. but the point of the article is that it thins your deck slightly. to the point where the 1 life you pay is probably not worth it. (not to mention the stifle bait fetches are in legacy)
so running fetches for, and only for, thinning is worthless. fetches have other uses, but thinning isnt an important one.
UWUW ControlUW
UGWSpiritsUGW
GHardened ScalesG
WGRUKiki PodWGRU [RIP]
You missed my point. The article is right in suggesting that statistically it does not make a big difference. But wrong in assuming that statistics carry over to each game. What I mean by this is that their will be games that fall out side of the curve. These are the games that we care about where it does make a difference.
The article is saying that you will encounter more games lost due to stifle or paying 3-4 life than you will encounter games won because of drawing 1 extra spell. The question is whether paying 2 life every game by turn 8 so that you draw an extra card in 1/10 of isn't going to increase the number of games you win for many decks. That 1 case where 2 life and a land card bought you a spell has to outweigh 9 cases of paying 2 life for no benefit. This isn't even considering the impact of stifle yet. Is 20 life spread over 10 games worth making 1 land card a spell. Life to cards is obviously powerful at 1 life to +1 card. Is it so powerful at 20 life to +1 spell - 1 land?
Fetchlands will tend towards 1 life for +1 spell -1 land as the game reaches turns 40-50. The problem is that the vast majority of games are going to be decided in turns 4-10 where you aren't getting anywhere close to that amount of value out of them.
To see the reality of fetchlands number your lands somehow. Use different art, write numbers on them, it doesn't matter. Then play games where when you fetch you choose a land number, say Island #2, and play as if the fetch were that land. Then act as though Island #2 automatically cycles when you draw it. See how frequently you just hit another land card from that cycling, see how infrequently you are even going to get to cycle Island #2. This simulation models exactly the results of fetchlands, in a way that you can see just how many times it would matter.
You say this as if +1 spell -1 land is always worse, or at least no better than +1 card. This is the error for which I was critiquing the article. For some decks, i.e. decks with low curves, no additional uses for land, and no mana sinks, this isn't true. In such decks, trading a land for a spell is much more useful than drawing a card, which has a fair chance of being a useless land. With Sign in Blood, which is the example I always use, here, you gain about 0.368 resource or resource-trading spells, versus just playing a resource-influencing spell in Sign in Blood's place. For decks that only care about influencing the game state via low cost spells, that's the number that matters. It does this at a cost of 2 life and 2 mana, which is comparable to the influence of fetchlands through 14 turns (2.75 life and 0 mana). That's still, usually, too slow to make a difference, but it makes fetchlands seem a lot better in hyper-aggressive decks. Of course, the existence of Stifle makes this all a moot point in legacy.
draw a hand 3/7 are lands 20/53 are left= 37.74% chance ull draw a land next
with fetchland 19/52= 36.34% chance u will draw a land next
next turn assuming u dont draw a land it becomes 20/52 = 38.46%, and 19/51 = 37.25%
playing another fetch will bring it down to 18/50= 0.360% so it does mathamaticly thin down your deck.
(*its the 1% that make you a winner ;))
Mordern: Melira, UR Storm, RDW, Infect, W life/control, UW Tron
Legacy: RDW, Pox
Vintage: Dark Depths, R Grey Orge
1/3 of your cards are lands. If you crack a fetchland before your draw on t2 if you're on the play, you're going to have 52 cards left in your library. If 18 of those are still lands, you have a 34% chance of drawing a land. If you do another, the turn after you have a 32% chance of drawing another land. Is a two percent increase enough to merit you playing fetchlands? In some competitive builds, any increase in your chances is great. But they don't make it significant enough for it to matter.
...unless you're playing crucible and exploration.
Legacy:
:symg::symb::symg:Pox:symg::symb::symg:
Edh:
:symg::symw::symu:Angus Mackenzie:symg::symw::symu: