12

As I write this, game 73 of the TCEC Season 18 superfinal has just ended. What was most "interesting" about this game is the sheer amount of shuffling both sides did: starting from around move 30 to move 200+, effectively nothing happened with Leela pushing a pawn only to avoid 50-move draws. It had three such moves to reset the 50-move counter, which naturally led to 150+ moves of shuffling. On move 220, Leela finally played 220. e4, which led to some trades and a more equal-ish position, but this also reset the 50-move counter and the shuffling started anew. The game went past move 300 with opposite color bishops on the board, enough for humans to say it's a clear draw, but Leela still saw a 40% chance for White to win, which led to more shuffling. Finally, on move 357 and faced with yet another 50-move draw, Leela's eval dropped and the game was adjudicated as drawn.

(This description makes it sound like it's Leela's fault the shuffling happened, but in fairness, Stockfish also saw a substantial White advantage for much of the game.)

This kind of scenario happens pretty often in computer chess, but I'm wondering what would've happened if this were a human game. Can the player with Black claim that White is being unsporting by shuffling and refusing to accept a draw? Are arbiters permitted to adjudicate if one player isn't making an active attempt to win the game, but seems to be merely prolonging it?

  • 3
    A similar example from human chess would be Howell vs Harikrishna, 2019. In the 236 move long game Howell refused to accept a draw in a dead drawn endgame and later claimed that arbiters were 'breathing down my neck (literally)'. – AKP2002 Jun 29 at 8:20
11

Can the player with Black claim that White is being unsporting by shuffling and refusing to accept a draw?

No. White may be quite legitimately trying to win on time. In blitz this is always allowed. In longer formats it depends on the format.

Are arbiters permitted to adjudicate if one player isn't making an active attempt to win the game, but seems to be merely prolonging it?

Only if there is no increment and the other player is down to less than 2 minutes on the clock. This is dealt with in the section called "Guidelines III. Games without increment including Quickplay Finishes" of the FIDE Laws of Chess and applies only if this was announced beforehand.

Guidelines III. Games without increment including Quickplay Finishes

III.1 A ‘quickplay finish’ is the phase of a game when all the remaining moves must be completed in a finite time.

III.2.1 The Guidelines below concerning the final period of the game including Quickplay Finishes, shall only be used at an event if their use has been announced beforehand.

III.2.2 These Guidelines shall apply only to standard chess and rapid chess games without increment and not to blitz games.

...

III.4 If the player having the move has less than two minutes left on his clock, he may request that an increment extra five seconds be introduced for both players. This constitutes the offer of a draw. If the offer refused, and the arbiter agrees to the request, the clocks shall then be set with the extra time; the opponent shall be awarded two extra minutes and the game shall continue.

III.5 If Article III.4 does not apply and the player having the move has less than two minutes left on his clock, he may claim a draw before his flag falls. He shall summon the arbiter and may stop the chessclock (see Article 6.12.2). He may claim on the basis that his opponent cannot win by normal means, and/or that his opponent has been making no effort to win by normal means:

III.5.1 If the arbiter agrees that the opponent cannot win by normal means, or that the opponent has been making no effort to win the game by normal means, he shall declare the game drawn. Otherwise he shall postpone his decision or reject the claim.

III.5.2 If the arbiter postpones his decision, the opponent may be awarded two extra minutes and the game shall continue, if possible, in the presence of an arbiter. The arbiter shall declare the final result later in the game or as soon as possible after the flag of either player has fallen. He shall declare the game drawn if he agrees that the opponent of the player whose flag has fallen cannot win by normal means, or that he was not making sufficient attempts to win by normal means.

| improve this answer | |
  • Who is "the opponent" referring to? Is it the player who is shuffling, or the opponent of that player? – Allure Jun 30 at 5:39
  • The opponent is the opponent of the player claiming the draw. i.e. the player who is supposedly not trying to win. That is clear if you read section III.5 above – Brian Towers Jun 30 at 9:34
4

In fairness , it seems Leela zero could have won that game by NOT shuffling on move 78. The pawn break e4!was decisive , and Leela missed it. Stockfish own PV lead to an ending with Rook and Bishop for Leela , Rook and Knight for Stockfish, with Leela a pawn up. I analyzed with Houdini 6.02 with TWO iterations of twenty minutes each , going in the first so far at depth 36 (Eval was stuck at +1.62). I put up the second iteration of analysis as soon as there were a diverging ply , with Black to move. After another 20 minutes d= 31/84 the lines indicated a rook endgame , with eval +4.11.It seems Leela zero missed a win, and not always shuffling pay off.

| improve this answer | |
  • I'd be careful with relying on Houdini's eval in these positions, because 1) Houdini is significantly weaker than Stockfish, and 2) the Stockfish playing was on insane hardware. You ran Houdini for 20 minutes and reached depth 31/84 in the position on move 78, but the Stockfish playing reached depth 41/72 in only 13 seconds. – Allure Jul 1 at 1:27
  • 78e4 Nh7 79 exf5 Bxf5 80 Bh4 Cf6 81 g4 Bh7 82 Bg3 Dd8 83 bxe6 Rfe7 84 Te5 Ag8 85 Bf7 Bh7 86 Bxh7 Kxh7 87 Qe2 Kg8 88 gxh5 Qd5 89 Kh2 Qf7 90 h6 Nd7 91 Rxe7 Rxe7 92 Dd3 g6 93 d5 cxd5 94 Qxd5 Kh7 95 Qxf7+ Rxf7 96 Kg2 Kxh6. From Stockfish own PV .Now it comes the endgame I suspected is winning. WHITE King in g2 Rook in d1 Bishop in g3 pawns in a5 b4 f3 h3 . BLACK King in h6 Rook in f7 Knight in d7 pawns in a6 b7 g6 WHITE to move. – Stefano Oct 13 at 18:18
  • 80 Nf6 82 Qd8 84 Re5 Bg8 92 Qd3. Sorry for the above errors. – Stefano Oct 13 at 18:24
  • The above ending d = 42/87 eval +1.85. Second iteration d= 27 /65 eval +3.74. It seems winning , albeit slowly. – Stefano Oct 13 at 19:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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