28

It is known that the minimum number of moves required for checkmate is 2 moves (4 half moves) resulting in a victory for Black. However, what is the minimum number of moves required for stalemate? Clearly, this would take a lot more moves since no piece from one party must be able to move, thus it must either be captured, pinned or completely blocked.

Glorfindel
  • 24,825
  • 6
  • 68
  • 114
Haran
  • 621
  • 6
  • 12
  • Another fast one is this twelve move game which ends in stalemate with all pieces still on the board. – Glorfindel Mar 09 '20 at 09:46
  • 1
    There's also the stalemate by repetition, but I think you've precluded that. – AER Mar 09 '20 at 23:41
  • 1
    @AER Wouldn't you be able to reach stalemate by repetition in the same number of turns as the number of repeated moves required? Each player would just move one of their knights back and forth. – nick012000 Mar 10 '20 at 03:24
  • 2
    @nick01200 I am not including draw by threefold repetition. Here, I refer to stalemate only as the situation where one of the players is unable to make a move in their turn. – Haran Mar 10 '20 at 05:15

1 Answers1

39

Sam Loyd, a famous puzzle composer, created this 10 move, 19 ply, stalemate game over 100 years ago. I found it via Google, in this post by AndersElborg on chess.com. Most black pieces are blocked, which saves some moves because those pieces don't need to be captured.

[FEN ""]
[StartPly "19"]

1.e3 a5 2.Qh5 Ra6 3.Qxa5 h5 4.h4 Rah6 5.Qxc7 f6 6.Qxd7+ Kf7 7.Qxb7 Qd3 8.Qxb8 Qh7 9.Qxc8 Kg6 10.Qe6
Rewan Demontay
  • 17,514
  • 4
  • 67
  • 113
Glorfindel
  • 24,825
  • 6
  • 68
  • 114
  • 1
    Can we prove that this is the fastest possible stalemate? – Haran Mar 09 '20 at 07:22
  • 4
    @Haran no idea ... but since this is a very natural chess 'record' and this solution is so old, it's unlikely that there is a faster way. – Glorfindel Mar 09 '20 at 08:28
  • 1
    I'm no expert, but if computers/databases allow to explore all possibilities up to move 10, proving this via bruteforce would be a matter of querying if a stalemate is found with less than 10 moves – fr_andres Mar 09 '20 at 11:26
  • 8
    @fr_andresSupportsMonicaCellio I thought of that, but with an estimated number of 20 possible moves per move, you're looking at 20^18 positions to analyze, which is far beyond current possibilities. Chess engines look much deeper than 10 moves, but only because they're able to ignore a whole lot of moves which don't make sense in a real chess game. – Glorfindel Mar 09 '20 at 11:29
  • If you expand the definition to include draws by repetition, you could trivially just have both players move one knight back and forth a couple times. That's probably the shortest draw you can get. – Darrel Hoffman Mar 09 '20 at 16:08
  • 5
    @DarrelHoffman I am not including draw by threefold repetition. That would make the question trivial. – Haran Mar 09 '20 at 17:15
  • @Glorfindel I'd be curious how aggressively the search tree can be pruned. The only obvious rule that come to mind (other than disallowing 3-fold repetition) is 'the game must not end in checkmate'. I'd be interested what other rules more experienced players can prove must be true to beat a 10-move stalemate. – Chuu Aug 01 '22 at 16:43