Author: Vasik Rajlich
Date: 04:50:35 02/16/04
Go up one level in this thread
On February 15, 2004 at 17:44:05, Drexel,Michael wrote: >Why do many (all?) chess engines continue to ponder a move after a huge fail >high? >They found most likely the refutation of the ponder move. >Why don't they start to ponder another move instead? >That does not make any sense to me. > >Michael In the case of UCI engines, the problem is with the UCI protocol. You must send the ponder move with the best move, and it would be quite a hack to start ignoring the "go XXX ponder" command and start doing some homework on some new ponder move. In addition, the work would have to be done in secret, since the GUI couldn't report it. In theory, I think the "bestmove XXX" and "ponder YYY" should be separate commands. At the moment something like 5% of my engine's pondermoves are simple blunders, there was no reasonable-depth fail-high response to the played move in the HT. I don't think there is a way to fix this which is completely free computationally. Vas
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.