Author: Robert Hyatt
Date: 07:17:16 12/01/03
Go up one level in this thread
On December 01, 2003 at 08:01:28, Gabor Szots wrote: >If an engine is unable to claim draw on repetition, then it is not a draw if the >opponent plays on. Not a draw even if the same position repeats 10 times. >If the GUI claims the draw, it is as if a spectator would have claimed it, >therefore invalid. No it isn't an easy case. In the case of Crafty, for example, the GUI _is_ the engine when I play in a tournament. Your understanding is wrong. Here are two examples that are exactly the opposite of what happened in this event: (1) in 1977 I was playing my program blitz vs Belle in the WCCC event in Toronto. Levy was the TD. Everytime he said something on the microphone, it zapped my modem connection to the computer I was using (this was an acoustic modem). When I restarted during this game, blitz, playing black, confused the king and queenside of the board due to a bug. At the time I used English descriptive notation (B-kb4 vs B-qb4 rather than modern algebraic Bf5 vs Bc5). After re-starting, blitz played a move B-kb4, which put its bishop on an open file with its king. This was obviously bad, and the program searched deep enough to see it. I looked at the output and noticed that the chessboard showed that the correct move was B-kb4, but since it had lost the idea of which side was K vs Q, the output said K. Belle played R-kb1, we were forced to play p-kn3 defending our bishop that was pinned on our king. Belle played p-kn4 and won the bishop. Before all that happened, Ken noticed that I had played B-qb4, which I had done because I looked at the board diagram the program displayed rather than the "my move is ...". He called David Levy over. After looking at the "My move is B-kb4" and the board showing B-Qb4, David said "the program said 'my move is B-kb4' so that is the move you have to play." We backed the game up to that point, I played that move rather than B-qb4, we lost, and I fixed the bug that night. (2) Thorsten has already posted an example where CSTAL showed one move in the analysis, displayed a different move on the board, and this time the decision was the GUI move was the one to be played, not the move shown by the engine (one showed a promotion to queen, one showed a promotion to a knight, the GUI showed the move that lost, the engine showed the move that won. Now flash forward to 2003. The GUI says "this is a 3-fold repetition". The TD says "the engine doesn't see it, we ignore what the GUI says and the game goes on. Consistency is the name of the game. There is _none_ here in this decision. Absolutely none. There are other such examples as well... These events have been about computer vs computer. We've _never_ had a fine line about what the GUI says vs what the program says. Trying to make such a line is simply trying to cover up for a poor decision. > >It seems to me that Jonny did not recognize draw because it accepted Shredder's >next move. In this case Shredder was EXTREMELY LUCKY, but nothing illegal or >unethical happened. Accepting the move does _not_ mean the engine didn't claim a draw. I play on in Crafty because the human might want to do so. Not because I change my mind and wish I hadn't claimed the draw. > >I think it is unimportant that the GUI claimed the draw only when the position >had in fact repeated 3 times. In a human tournament, if I make my move, STOP THE >CLOCK, and call the arbiter, my draw claim is still valid. But without the claim >of the player himself, NOT EVEN THE ARBITER has to right to declare a draw. This >is what saved Shredder. An unlucky blow for the Fritz team. The operator made the decision to play on. But the operator is _not_ allowed to make _any_ decisions while a game is in progress, as per the rules. Therefore this reasoning simply is unsound because it is based on rules that were not in effect. The operator is passive. He _always_ has been passive, at least when we go by the rules in force for these events. > >Gábor
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.