Author: Bruce Moreland
Date: 23:08:00 02/04/00
Go up one level in this thread
On February 04, 2000 at 21:25:07, Marc Plum wrote: >I've just been reading the "drawn position" thread relating to a game between >Nimzo and Shredder in Enrique Irazoqui's tournament. One sub-thread concerned a >possible KPvsknn ending. This position was not forced, I believe, but was a >plausible variation. Although tablebases showed it to be a mate in 62, >application of the 50 move rule would make it a draw. > >The rules of chess, however, do not require the 50 move rule to be applied in >all situations, although this has become the norm. Rule 10.12 of the FIDE laws >of chess permits the number of moves to be increased for certain positions, if >these positions and the extra number of moves required has been announced in >advance of the tournament. I have not heard of this being done in practice, but >the possibility is there. > >It seems to me that a sensible special rule for computer tournaments would be to >extend the 50 move rule for any position where a program's tablebases showed a >clear win. The reason for the 50 move rule is most likely to keep a player from >torturing his opponent endlessly in a drawn positions. Ideally, however, this >should not affect the outcome of a provably winning position. > >The main problem I see with this is that some programs would simply refuse to >continue past the 50 move mark. I have looked at some tablebase endings with >Shredder 4.0, and I have to go into the options and increase the number of moves >until a draw if I want to see a really long mate. Other programs cannot be set >to recognize exceptions. Still, such games could be adjudicated in accordance >with common-sense rules. If a program has a winning position, tablebases, and >the ability to use them, then adjudicate the game as a win. > >The use of tablebases may also have implications for the 50 move rule in human >vs computer contests, but that is a whole different topic. I think there are some interesting issues here. I don't favor extending the 50-move rule under any circumstances. If I am not using endgame databases, or if there are some databases that I don't choose to use or don't happen to have, I should not have to modify my program so that it understands which endings have been determined to require more than 50 moves to mate. It is ludicrous to expect that I should have to modify my program in such a manner that it does not play according to FIDE rules, so that some other program can more conveniently play out of accord with FIDE rules. If someone generates a 6-man table and finds a mate in 200, how am I supposed to deal with this if I don't happen to have the table? Imagine this case: I'm not using any tables, and I'm on the weaker side of a 6-man case that is drawn according to my opponent's table. At move 47 I make a move that allows a conversion 4 moves later, into a 5-man ending that is won for my opponent. My score is "draw" because conversion is not possible before the 50th move. But my opponent now says "mate in 35" or some such, since I have transferred to a 6-man case that is in fact lost. Should I lose this game? I don't think so. These games should be played in accord with FIDE rules, to avoid problems such as these. If FIDE wants to specify that some endings provide for extension of the 50-move rule, then I should have to modify my program to be in accordance with the rules, but if they don't do this, I should be able to run according to the rules that are in force. This is inconvenient to those who want to use big tables, but I say that this is a down side of using those tables, and you have to think about this when you install them. bruce
This page took 0 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.