5

This question already has an answer here:

Historically stalemate has been handled in different ways. I wonder how different chess would be if it was a win for the stalemating player.

Has enough thought been put into this so that it's possible to give a clear answer? Would there be significantly less draws and more wins for white? More aggressive play? Would changes be seen already in the opening?

A related question that already has been asked is why the actual rule is as it is. The accepted answer to that question talks mostly about history and the logic that has been given for the rule, but another answer goes into how endgame would be different with this variant instead. There it is a side issue for explaining something else, which has made someone suggest this is a duplicate question. I hope that the actual what-if question can get more thorough answers for itself than when it is a side issue of a related question. For example my issue about if changes would be seen already in the opening have not been brought up.

The ideal answer for me would be if something like AlphaZero (without preconceived notions) had spent some time playing this variant, but I guess I'm out of luck on that?

marked as duplicate by Brian Towers, Glorfindel, Phonon, itub, GloriaVictis Jun 8 '18 at 9:44

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 1
  • The answers to chess.stackexchange.com/questions/1412/why-is-stalemate-a-draw answer your question. Please learn to use the search function. – Brian Towers Jun 7 '18 at 15:50
  • No, that's a different question, even though some answers and comments there (mostly the answer by Jonathan Garber are of interest here). – pst Jun 7 '18 at 19:52
  • 2
    More aggressive play? I doubt it. Common sense says there would be less aggressive play. Players will be less willing to sacrifice a pawn for an attack if you make it too easy to win the endgame with an extra pawn. Would changes be seen already in the opening? No more gambits. Dull maerialistic chess. – bof Jun 8 '18 at 2:30
  • @bof I agree. There maaaay be some more "aggressive play" at amateur level along the lines of "oh, I just lost a pawn, and since my chances to defend are smaller without stalemate draws, well, YOLO sac sac", but I doubt that is the type of aggressive play that we actually want to see. Masters will certainly become more prudent and will simply resign earlier/more often when behind in material. Not very exciting either. – Annatar Jun 8 '18 at 8:13
3

This would definitely give less draws, and more wins for both colours. The following very common draw themes no longer possible:

  • Bishop of wrong colour
  • KP+K, the losing K is in corner, pawn is on the same file
  • KP+K, the losing K controls the queening square (very common in tournament play)
  • KQ+KP, where the pawn is a promoting bishop pawn or rook pawn

Less margin for the losing side to come back for a draw.

I'm very convinced Google won't waste money on a variant like this.

  • Would it lead to more wins for both colors, or would it simply increase White's level of dominance? White has a clear advantage in the game; the only question is whether it can leverage that advantage enough to achieve better than a draw. While stalemates are rare in actual play, many games are recognized as draws because the side that's behind could prevent the leading side by forcing anything better than a stalemate. – supercat Jun 7 '18 at 16:15
  • 4
    As for the value of using AI to explore chess variants, I think that could be interesting for purposes of finding a better balance. If there were a rule that failure by White to win within N moves would count as a victory for Black, that rule by itself would shift the game excessively in Black's favor, but if combined with a rule making stalemate a win could shift the balance back toward White. It would be interesting to know the value of N for which perfect play by White could force a win in N moves, but perfect play by Black could prevent a loss in N-1 or fewer moves. – supercat Jun 7 '18 at 16:22
  • @supercat Interesting idea Perhaps let N = MX + 2018 - year and start MX at 150 or so today, decreasing N each year until the chess world says stop? Certainly this is too volatile for many, yet your basic idea is intriguing. – chux Jun 7 '18 at 17:22
  • @supercat " It would be interesting to know the value of N for which perfect play by White could force a win in N moves, but perfect play by Black could prevent a loss in N-1 or fewer moves. " You're pretty much asking for chess to be weakly solved. – Acccumulation Jun 7 '18 at 18:44
  • @Acccumulation: Finding an exact value of N would probably be intractable, but it would probably be possible to come up with estimated upper and lower bounds that are pretty good (e.g. say that anything over NH moves would likely be forced win for white, and anything below NL would be a forced win for black, for some NH and NL that aren't too outrageously far apart. – supercat Jun 7 '18 at 19:44
3

Building on SmallChess' answer, KP vs K endgames that are normally drawn would be lost. This has very big ramifications, as there are many rook endgames where the side up a pawn can easily transpose into a (formerly drawn) KP vs K endgame. For example:

  • The Philidor Endgame
  • A normal K+R+3P vs K+R+2P on the same side of the board.
  • A K+R+4P vs K+R+3P where the extra pawn is a passed a-pawn. If the defending rook is behind the pawn, the endgame is considered barely drawn. However, the attacking rook might be able to exchange the extra a-pawn for a pawn on the kingside, reaching the aforementioend K+R+3P vs K+R+2P endgame.

This logic can also be applied to other endgames, such as K+Q+P vs K+Q. However, Rook endgames are the most common, and by making stalement = a win, many of these fundamental endgames are ruined.

Once that has happened, it's often not difficult for the better side to transpose down into one of them (especially the Philidor).

So making Stalement a win would drastically lower the number of draws that happen when one side is defending an endgame. However, there are many cases where no one was ever winning in an endgame (i.e., even material throughout), and most of these would probably result in draws still. The reason is that many draws in equal endgames arise from an "impasse" both sides create, where neither can make progress.

Not the answer you're looking for? Browse other questions tagged or ask your own question.