Author: Robert Hyatt
Date: 13:13:04 11/03/05
Go up one level in this thread
On November 03, 2005 at 03:18:40, Mark Mason wrote: >Plesae make allowances if this is a well acknowledged observation, but I am >relatively new to Comp chess and so have jst experienced it. > >In some of my engine to engine games (5 min blitz) on playchess server - we >get to an ending and because my engine can "see" dozens of moves ahead accessing >the EGTBs that I am drawn (or even lost) it will offer a draw or resignation >straight away. > >In several cases, my opponent is clearly not using EGTB's and is calculating >each move. He has had only a few seconds left and I have lots of time, but >because I am using EGTBs I offer draw or resignation - but he would obviously >have lost on time had we continued !! > >In these cases, it seems to me that there is a risk that if you are using EGTBs >and opponent is not, you lose or deaw games that could otherwise have been won. >It made me wonder if using then for blitz play is actually worth the gain, when >measured against the risk I have described. > >I know that it is not all about winning and ELO, but I found this an interesting >observation and wouldwelcome other views/experiences. > >Mark The idea is not new. :) Crafty's "swindle mode" is based on that very observation. All moves lead to EGTB draws, except for maybe a few that lead to losses. Crafty generates a list of all legal moves, excludes any move that doesn't lead to a forced draw, then it does a normal search with zero EGTB probes, and plays the best move the search can find, which stretches the draw out as far as possible. But since we know that we are only searching moves that lead to a draw, we can't possibly lose...
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.