Author: David Eppstein
Date: 11:34:00 08/23/98
Go up one level in this thread
On August 23, 1998 at 11:31:54, Thorsten Czub wrote: > We found that maybe fritz extends some check-check lines very very deep, > that deep that it believes the game is DRAW and the extentions give back a > draw score and the main-line has repetition moves... But the opponent has > POSITIVE score and will not play for a draw of course, although he maybe > COULD do so... In these situations you not only see fritz having these > draw-behaviour but also it comes not much deep. This reminds me of a bug I had at one point in my program, where it had repetition detection causing some evals to be zero, and hashed them indiscriminately so they would be looked up as zero even in situations where the same position might not be a repetition. In positions that were even or slightly disadvantageous to the program, it would make horrendous blunders with nearly zero score, only to have the score suddenly drop a few moves later. This wouldn't happen when the program was ahead because it wouldn't like the draw scores... We discussed this in another thread here a few weeks back -- I think the conclusion was that it was relatively easy to eliminate such "false positives" where you were seeing a repetition that isn't really there, but much harder to avoid "false negatives" in which you should see a repetition but don't (the famous Deep Blue game being an example of such a missed repetition, although maybe not one that anyone should have seen). Anyway, if this is really a bug in Fritz (not just a bad evaluation as Amir suggests) this could be one possible explanation.
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.