What an appropriate day to publish this, Wizards just banned the most cards in one B&R announcement in the history of the game and it was because of practicality. Something I really love about 3-Card Blind is how community driven it is, and when something is this disruptive to the community I think it’s entirely reasonable to take this action. I imagine that Hangarback would be banned anyway once the next ban vote commences, but since it’s creating huge problems that need to resolved immediately I am okay with emergency banning it right now.
I played many EDH / Commander games with a player playing Sharazad frequently and even copy it. It was great and one of my best memories from early Commander before the ban. So sad that its banned. 😢
A lot of the best puzzles are also very simple and elegant. This problem is so simple yet so complex! Only 3 unique cards can create an unsolved problem. Reminds me of Fermat's last theorem. Hopefully it takes less than a century to solve this.
I had a heart attack when I saw this video because my friend subbmitted a deck with elixer. Also, thanks for being really open about this, that's quite wonderful. As for the banning, maybe you choose to ban a card, but let the community choose which one.
Not sure why this video was recommended to me in particular, but I remember playing 3CB back on MTGSal and GFAQs. I wasn't any good, but it was fun. Awesome to see it's still going on, might look into it again.
I haven't seen anything about 3CB in probably 20 years. I'm rather amused to see that there's still a community of people that are interested in it. I'm not surprised given that there are now way more options available that some of them would lead to games where proving the optimal strategy would be beyond the capabilities of people.
We use Elo, ties favor the player with lower Elo. Omitting the match lets us present more statistically accurate information and does not favor either player.
@@MTG3CardBlind Then you could take inspiration on chess "50 turns without a capture or pawn advance is a tie" rule. Ties for practicality are compatible with ELO as long as they are clearly defined. So something like "50 turns without a player getting closer to 0 HP" (where only the minimum count, if you gain life and lose life again it doesn't count, and negative hitpoint if you are immune to losing by hitpoints also don't count). The goal would be to make exhaustive approaches through computers possible. Admittedly, I'm not sure that the rule I propose is tight enough to solve this problem, computer simulation might still take an eternity in complex situations.
@@MoiMagnus1erthat definition would preclude a game state that does go for a long time but does win, inevitably and in a way that can be shown. whilst i dont believe such a state is currently possible, it is one of the problems with implementing such a rule somewhere where the possibility space expands exponentially. it would probably be fine to add it, and readdress it should such an entry become possible, but some consideration would be necessary for a reasonable game which would be forced into a tie
@@MoiMagnus1er We already mark games as ties, but ties have a specific definition in this format. "We can't solve the match" is not a tie, nor is marking it at a tie a superior way to handle these situations than omitting the match.
@@MTG3CardBlind I agree that as an "after the fact" decision, you should not declare it a tie, since the current definition of tie doesn't includes it. And omitting the match, while not "perfect" (it still has some statistical influence, from the omniscient point of view of someone who knows who should have won), is the best compromise. What I'm saying is that Chess has updated their tie rules along the years to handle impractical situations (last update for Chess was in 1992, so 20 years after Elo was introduced), and if that debates about bans don't lead to a clear answer, the debate could also include redefining what a draw is for future matchs.
I'm curious if letting players play the match manually would be another way to solve this. Like a game of chess, it may turn out to be a boardstate that's too hard to actually determine who wins, and the playing of the match itself could provide strategic options for either player to gain an edge. Of course, this would change the dynamic from one of deckbuilding alone to also one of skill in the matchup and gameplay itself.
I think it would be very interesting to see players play this match out, but it would result in the outcome being different for different players, which is antithetical to what we do in the Main Event. It could be a fun mini game though.
I be sad to see either of them ban but i do understand at some level praticality has to win out. Also I knew some complexity beyond what I would have guess could happen from 3 card blind but this is beyond what i thought could happen. I hope if you get a confident answer, that you can share it with us. Also it may be fun to have "what is the most complex match you can make" as a general soft challenge for some players. Thanks for the video and I hope an answer that makes the tournament runnable but also still keep some of the complexity is reached.
Agreed, amazing art. His artstation profile features works he has done for other brands too that are equally amazing. If you have not checked out some of his non-MTG works, I would highly recommend it 👍
First of all, it seems to me that this case where it's practically impossible to determine what's the win state for either player is a great sign that this is going to be a tie. Second of all: how about giving the players the burden of proof for these kinds of cases?
To respond to the first, it is an unfortunately not an indication that the game is going to be a tie. The good news is progress is being made on resolving this match and there is compelling evidence that Elixir wins. To respond to the second, giving players the burden of proof in this case unfortunately does not solve the issue as there is reason to believe that only a tiny fraction of our players could handle resolving a match like this. It is not the fault of those players for submitting these decks either, the decks do not seem complicated in isolation of each other. I also want to make sure that having signficant game knowledge is not a barrier to participate in this community.
Surely the double Hangarback would win slightly more often, no? The single Hangarback hand requires at minimum three full turns of actions to do their full loop with only City of Traitors ("Cast Hangarback for X=1", "Pump Hangarback and/or Cast Elixir", "Crack Elixir"). Any scenario where the Double-Hangarback hand can successfully deal 2 damage per turn, they're guaranteed to outpace the life-gain from Elixir. Probably only fractions of a % and play/draw dependent though.
The current most developed proof has the Elixir player winning, which is confirmed when they are on the draw and strongly suspected when they are on the play. I am not even going to try to attempt to explain it, but if are curious to see the lines used, you can join our discord. There is a massive thread there that has collected all the work done on this match.
Seems like the one with Elixir would win since he can afford to keep his smaller than the opponent and attack. Gaining 5 life a turn as well. The one with the double Hangers will get to get his thopters out but if he doesn't do it at a good time while building the other, then he won't get enough thopters to win. Also it seems like this favors the first player more if they have the elixir, but favors the other if he goes second, which is strange.
Elixir can only proc every 2 turns as it takes 3 mana to do, I lean towards your guess too, but I don't think the logic is simple the tempo vs sustain is so murkey
I really can't imagine a world in which the deck running Elixir ever wins. The deck with two Hangarback Walkers will have a single Walker getting larger at double the speed of the opponent if they're playing Elixir and using it every other turn. They'll keep their second Walker matching the stats of the player with Elixir so it can be a trade and the Thopters can block each other. Eventually the one continuously growing Hangarback will breaker through all the chump blockers and overcome the 2.5 lifegain each turn from the Elixir player.
I definitely think the Elixir deck is at a disadvantage, but whatever strategy you might come up with for the other player, there's presumably a counter to it available, or else this wouldn't be such a problem. There are so many possible ways that the game could branch that trying to figure out whether a specific strategy has no counterplay is very difficult. The optimal strategy likely requires a very complex decision tree that's just beyond the reach of anyone with a finite amount of computing power in the same way that solving chess is.
@@stevenglowacki8576 but the deck with elixir will be able to chump the big walker and then use the elixir to get it back. So... When do you attack, as the double walker player? The elixir is not so much there to be used to heal every turn. It's an option to recover the walker and keep your health up. I actually can't imagine the player with finite resources losing
In case you are curious, the current most comprehensive solution has the Elixir player winning on the play and the draw. How you might ask? I have no idea, I am avoiding this puzzle like the plague at this point. We are waiting to see if someone can put together a comprehensive document covering the solution now just to be sure, but the people who have been working on this sound pretty convinced that Elixir wins.
@@MTG3CardBlindThat’s wild, I definitely would have guessed it going the other way too, now I kinda wanna buy some hangarback walkers, an elixir, and some temple of the false gods to redact and play it out
I think the situation is that after a certain point you don't want to grow your hangarbacks, because you don't want to be forced to kill the enemy hangarback. For instance, by the time all the hangarbacks are 10/10, attacking is losing. But if there's such a point, then the elixir gains life
Currently the 3 Card Blind database has 33,682 matches saved, and only this one raised this level of concern. Assuming it is a somewhat representative sample, that means only like 0.003% of matches are like this. That is likely an oversimplification, but it is an indicator that these events are probably very rare (unless you are actively trying to cause them to happen).
Short answer is that I am not sure. We have some pretty talent folks in our discord, so it would not suprise me, but these are subjects I am personally not familiar with.
Sort of? There is concensus among folks who spent the most time working on it, and there was someone who created code to try to solve it that comes to the same conclusion, but no one has put together a comprehensive proof. Allegedly, the Elixir deck wins the game when it is on the play, and the game ends in a stalemate when it is on the draw, but we don't have conclusive proof that this is correct.
@@enugie No, I think the folks who devoted the most energy to this feel pretty confident about the conclusion, but no one wants to actually go through and create the comprehensive proof
City might get banned anyways, but the reason not to ban it due to this situation is because City supports many strategies that do not generate these sort of issues, such as Isochron Scepter decks.
What do you mean, what do you do!? They play the game. Time out rules may come into play. If it goes to extra turns, the rules are already stated (I believe it’s higher life total at end of extra turns wins, if there still is no winner). Players will be given the opportunity to play optimally and potentially time out or misplay and actually win or lose.
It sounds like you might not be familiar with this specific format. Frankly, it might not be your cup of tea, and that is okay, but the comment you made is descriptive of "normal Magic" not 3 Card Blind. Intro to the format: ua-cam.com/video/29P243ggtU8/v-deo.htmlsi=_YB32bVBAwDGKfj5
What an appropriate day to publish this, Wizards just banned the most cards in one B&R announcement in the history of the game and it was because of practicality. Something I really love about 3-Card Blind is how community driven it is, and when something is this disruptive to the community I think it’s entirely reasonable to take this action. I imagine that Hangarback would be banned anyway once the next ban vote commences, but since it’s creating huge problems that need to resolved immediately I am okay with emergency banning it right now.
"If Death's Shadow is so hard to compute, why isn't there Death's Shadow 2" - 3CB detractors, probably
there is, it’s called Shadow of Mortality and sees no play
Scourge of the Skyclaves
oh god elixir can shuffle the walker back into your library.
I missed this part at first. It definitely makes the game seem more tilted toward the Elixir player than I first imagined.
I played many EDH / Commander games with a player playing Sharazad frequently and even copy it. It was great and one of my best memories from early Commander before the ban. So sad that its banned. 😢
A lot of the best puzzles are also very simple and elegant. This problem is so simple yet so complex! Only 3 unique cards can create an unsolved problem.
Reminds me of Fermat's last theorem. Hopefully it takes less than a century to solve this.
I had a heart attack when I saw this video because my friend subbmitted a deck with elixer. Also, thanks for being really open about this, that's quite wonderful. As for the banning, maybe you choose to ban a card, but let the community choose which one.
Not sure why this video was recommended to me in particular, but I remember playing 3CB back on MTGSal and GFAQs. I wasn't any good, but it was fun. Awesome to see it's still going on, might look into it again.
Can't wait for all the comments trying to solve this
Not sure if you are in our discord or not, but we have people literally typing up dissertations at this point
@@MTG3CardBlindI was going to say you could get a solid math paper if you crack this. Glad to hear people are trying.
@@MTG3CardBlind wouldn't be surprised if Matt Parker gets involved :D
I haven't seen anything about 3CB in probably 20 years. I'm rather amused to see that there's still a community of people that are interested in it. I'm not surprised given that there are now way more options available that some of them would lead to games where proving the optimal strategy would be beyond the capabilities of people.
I'm kind of in the same situation, though I've been thinking about the format again recently. I had no idea it was still around though.
Taking breaking the game to the next level
Personally, I would rule any game complex enough to result in event delays as a "tie and move on" kind of thing.
We use Elo, ties favor the player with lower Elo. Omitting the match lets us present more statistically accurate information and does not favor either player.
@@MTG3CardBlind Then you could take inspiration on chess "50 turns without a capture or pawn advance is a tie" rule. Ties for practicality are compatible with ELO as long as they are clearly defined.
So something like "50 turns without a player getting closer to 0 HP" (where only the minimum count, if you gain life and lose life again it doesn't count, and negative hitpoint if you are immune to losing by hitpoints also don't count).
The goal would be to make exhaustive approaches through computers possible. Admittedly, I'm not sure that the rule I propose is tight enough to solve this problem, computer simulation might still take an eternity in complex situations.
@@MoiMagnus1erthat definition would preclude a game state that does go for a long time but does win, inevitably and in a way that can be shown.
whilst i dont believe such a state is currently possible, it is one of the problems with implementing such a rule somewhere where the possibility space expands exponentially.
it would probably be fine to add it, and readdress it should such an entry become possible, but some consideration would be necessary for a reasonable game which would be forced into a tie
@@MoiMagnus1er We already mark games as ties, but ties have a specific definition in this format. "We can't solve the match" is not a tie, nor is marking it at a tie a superior way to handle these situations than omitting the match.
@@MTG3CardBlind I agree that as an "after the fact" decision, you should not declare it a tie, since the current definition of tie doesn't includes it. And omitting the match, while not "perfect" (it still has some statistical influence, from the omniscient point of view of someone who knows who should have won), is the best compromise.
What I'm saying is that Chess has updated their tie rules along the years to handle impractical situations (last update for Chess was in 1992, so 20 years after Elo was introduced), and if that debates about bans don't lead to a clear answer, the debate could also include redefining what a draw is for future matchs.
I'm curious if letting players play the match manually would be another way to solve this. Like a game of chess, it may turn out to be a boardstate that's too hard to actually determine who wins, and the playing of the match itself could provide strategic options for either player to gain an edge. Of course, this would change the dynamic from one of deckbuilding alone to also one of skill in the matchup and gameplay itself.
I think it would be very interesting to see players play this match out, but it would result in the outcome being different for different players, which is antithetical to what we do in the Main Event. It could be a fun mini game though.
I be sad to see either of them ban but i do understand at some level praticality has to win out. Also I knew some complexity beyond what I would have guess could happen from 3 card blind but this is beyond what i thought could happen. I hope if you get a confident answer, that you can share it with us. Also it may be fun to have "what is the most complex match you can make" as a general soft challenge for some players.
Thanks for the video and I hope an answer that makes the tournament runnable but also still keep some of the complexity is reached.
Dominik Mayer is the creator of what is arguably some of the best MTG art ever since Zendikar Rising. Cool that you show off his work here.
Agreed, amazing art. His artstation profile features works he has done for other brands too that are equally amazing. If you have not checked out some of his non-MTG works, I would highly recommend it 👍
First of all, it seems to me that this case where it's practically impossible to determine what's the win state for either player is a great sign that this is going to be a tie.
Second of all: how about giving the players the burden of proof for these kinds of cases?
To respond to the first, it is an unfortunately not an indication that the game is going to be a tie. The good news is progress is being made on resolving this match and there is compelling evidence that Elixir wins.
To respond to the second, giving players the burden of proof in this case unfortunately does not solve the issue as there is reason to believe that only a tiny fraction of our players could handle resolving a match like this. It is not the fault of those players for submitting these decks either, the decks do not seem complicated in isolation of each other. I also want to make sure that having signficant game knowledge is not a barrier to participate in this community.
@@MTG3CardBlindI like your reasoning on the second point
And if elixir wins, good for them! But I would have placed my bets on draw
I think I’m gonna make this into a small format that I play with my friends
Surely the double Hangarback would win slightly more often, no? The single Hangarback hand requires at minimum three full turns of actions to do their full loop with only City of Traitors ("Cast Hangarback for X=1", "Pump Hangarback and/or Cast Elixir", "Crack Elixir"). Any scenario where the Double-Hangarback hand can successfully deal 2 damage per turn, they're guaranteed to outpace the life-gain from Elixir. Probably only fractions of a % and play/draw dependent though.
The current most developed proof has the Elixir player winning, which is confirmed when they are on the draw and strongly suspected when they are on the play. I am not even going to try to attempt to explain it, but if are curious to see the lines used, you can join our discord. There is a massive thread there that has collected all the work done on this match.
Seems like the one with Elixir would win since he can afford to keep his smaller than the opponent and attack. Gaining 5 life a turn as well. The one with the double Hangers will get to get his thopters out but if he doesn't do it at a good time while building the other, then he won't get enough thopters to win. Also it seems like this favors the first player more if they have the elixir, but favors the other if he goes second, which is strange.
Elixir can only proc every 2 turns as it takes 3 mana to do, I lean towards your guess too, but I don't think the logic is simple the tempo vs sustain is so murkey
I really can't imagine a world in which the deck running Elixir ever wins. The deck with two Hangarback Walkers will have a single Walker getting larger at double the speed of the opponent if they're playing Elixir and using it every other turn. They'll keep their second Walker matching the stats of the player with Elixir so it can be a trade and the Thopters can block each other. Eventually the one continuously growing Hangarback will breaker through all the chump blockers and overcome the 2.5 lifegain each turn from the Elixir player.
I definitely think the Elixir deck is at a disadvantage, but whatever strategy you might come up with for the other player, there's presumably a counter to it available, or else this wouldn't be such a problem. There are so many possible ways that the game could branch that trying to figure out whether a specific strategy has no counterplay is very difficult. The optimal strategy likely requires a very complex decision tree that's just beyond the reach of anyone with a finite amount of computing power in the same way that solving chess is.
@@stevenglowacki8576 but the deck with elixir will be able to chump the big walker and then use the elixir to get it back. So... When do you attack, as the double walker player?
The elixir is not so much there to be used to heal every turn. It's an option to recover the walker and keep your health up.
I actually can't imagine the player with finite resources losing
In case you are curious, the current most comprehensive solution has the Elixir player winning on the play and the draw. How you might ask? I have no idea, I am avoiding this puzzle like the plague at this point. We are waiting to see if someone can put together a comprehensive document covering the solution now just to be sure, but the people who have been working on this sound pretty convinced that Elixir wins.
@@MTG3CardBlindThat’s wild, I definitely would have guessed it going the other way too, now I kinda wanna buy some hangarback walkers, an elixir, and some temple of the false gods to redact and play it out
I think the situation is that after a certain point you don't want to grow your hangarbacks, because you don't want to be forced to kill the enemy hangarback. For instance, by the time all the hangarbacks are 10/10, attacking is losing. But if there's such a point, then the elixir gains life
Second Sunrise was banned because it was good enough to be played in tournaments, and it extended the game waaaay too long.
Yes, this is presicely stated in the video as well as the official article written by Wizards of the Cost that I have linked in the description.
Makes me wonder how many board states are “unsolvable”…
Currently the 3 Card Blind database has 33,682 matches saved, and only this one raised this level of concern. Assuming it is a somewhat representative sample, that means only like 0.003% of matches are like this. That is likely an oversimplification, but it is an indicator that these events are probably very rare (unless you are actively trying to cause them to happen).
Did anyone try to use Combinatorial Game Theory and Surreal Numbers? This nerd sniped me hard and I'm going to be thinking about it for a while.
Short answer is that I am not sure. We have some pretty talent folks in our discord, so it would not suprise me, but these are subjects I am personally not familiar with.
I say ban hangarback. Elixir is a nice hedge against thoughtsieze
Kyle hill type beat
Subbing just to get the eventual answer vid
1:53 i think you mean strong
Couldn't tiebreak be... Let them play it out? 😅
This would certainly be the funniest way to handle this, but it is not particularly practical. I joked about it quite a bit in our discord though.
The card to ban may actually be City of Traitors, given its balancing downside is null in the format
There are a lot of other lands with a similar ability.
So.. was it solved?
Sort of? There is concensus among folks who spent the most time working on it, and there was someone who created code to try to solve it that comes to the same conclusion, but no one has put together a comprehensive proof.
Allegedly, the Elixir deck wins the game when it is on the play, and the game ends in a stalemate when it is on the draw, but we don't have conclusive proof that this is correct.
@@MTG3CardBlind are there people still trying to solve it?
@@enugie No, I think the folks who devoted the most energy to this feel pretty confident about the conclusion, but no one wants to actually go through and create the comprehensive proof
Why not ban city of traitors?
City might get banned anyways, but the reason not to ban it due to this situation is because City supports many strategies that do not generate these sort of issues, such as Isochron Scepter decks.
What do you mean, what do you do!? They play the game. Time out rules may come into play. If it goes to extra turns, the rules are already stated (I believe it’s higher life total at end of extra turns wins, if there still is no winner). Players will be given the opportunity to play optimally and potentially time out or misplay and actually win or lose.
It sounds like you might not be familiar with this specific format. Frankly, it might not be your cup of tea, and that is okay, but the comment you made is descriptive of "normal Magic" not 3 Card Blind.
Intro to the format: ua-cam.com/video/29P243ggtU8/v-deo.htmlsi=_YB32bVBAwDGKfj5