Author: Ed Schröder
Date: 00:10:55 08/17/99
Go up one level in this thread
>Posted by Stephen A. Boak on August 16, 1999 at 22:59:53: > If Ed Schroeder altered Rebel 10-5 to move faster when the Opponent is >behind on time (I think he did this, from what I've read), then this was obviously in >order to try to get an advantage (pressure the human to error) when the human >was in time trouble. If this is true, why accept an early draw when the >results of these alterations are almost paying dividends--actually have put the >pressure on the human in time trouble? Somehow this seems inconsistent. Looks inconsistent indeed but the etiquette had a higher priority for me also due to the fact the position was a dead draw for already 5 moves (or so) which explains my draw offer. Keep in mind also GM Rohde made 2 draw proposals to me which I declined. Ed Schroder > On the other hand, I understand Ed's point about not respecting the GM or IM >opponent. When moves are readily obvious to achieve time control without >losing an even (and relatively uncomplicated) position, let's give the GM the >doubt and accept the draw, instead of turning a mutually respectful game of wits into a >possibly blunderful game of blitz chess. Ed isn't trying to test the GM and >Rebel in blitz play, but rather to see if one or the other can fairly wrest >the advantage and outplay the opponent at tournament time controls [We already >know the strong chess programs can blitz better than virtually all GMs (I am not >saying the programs will win *all* the games, mind you)]. > > If a game goes to nearly the first time control and neither has an edge, and >the game is not full of interesting, non-symmetrical, dynamic possibilities, >then a draw seems a fair result to offer and to accept at that time. I >imagine such games, although they must occasionally be played out because one player >believes he must win (for prize money, to avoid a tournament knockout, >etc), are very boring to most GMs and IMs. > > If you are not seeking a win at all costs--perhaps when you face elimination >from a match or tournament if you merely draw), why not accept the draw rather >than subject the opponent to the demeaning 'prove that you are a GM and can >see the obvious under time pressure' rejection of the draw offer. > > SUGGESTION--I would notify the candidate contestants in advance (when >soliciting their challenges) that Rebel has been programmed to take advantage, >if possible, of a human opponent's time trouble and the candidate can >sometimes expect the computer team to play out the game at least until the first time >control has ended. > > One of the Rebel goals has been to make the program play like a human. It >seems to me that this also involves playing out a time trouble situation in >order to try to wrest a permanent advantage from the opponent in time trouble. > > This should not be offensive to an opponent, especially if announced in >advance. > > I can agree with Ed that winning is not everything, especially if you are >trying to encourage GMs to play your computer under slow time control, match >conditions, in order to properly develop and rate your program's play. > > However, doesn't the offer of a prize compensate the opponent for making him >play out the first time control? Should the opponent *expect* that he need >only hold an even position until just prior to the 40th move (maybe move 36-38) and >the Rebel Team will automatically accept a draw offer despite the opponent's >time trouble, and pay him $250 for a draw without battling tooth and nail >to the wire (the first time control)? > > Bottom line, if Ed wants to pay $250 for the draw, on move 38 or so, that's >his team's decision. I can't fault him for trying to maintain a good working >relationship with GMs and IMs. > > --Steve Boak
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.