Author: Robert Hyatt
Date: 20:05:23 07/29/99
Go up one level in this thread
On July 29, 1999 at 18:58:12, Ian Osgood wrote: >Do other program authors curtail the search when there is a forced move at the >root? > >How do you detect that a root move may be forced? > >Could you compare the values of the best and second-best root moves after a >search iteration to detect a forced root move? (Granted, the second-best score >won't be accurate due to alpha-beta, but I figure that if the difference was >greater than a queen's value, you could still conclude that the best move was >forced.) > >Thanks, > >Ian I do two kinds of forced moves... (1) getting out of check, and there is only one legal move. I play such a move instantly (after a brief search to get a move to ponder). (2) obvious recapture. I limit this extensively... there has to be one capture that is far better than the rest (with static analysis). This capture has to maintain material balance (and not appear to win material so that it won't accept a sacrifice after a short search). It must capture apiece that just captured one of my pieces. If all of these are true, and the search has not failed low during the time this move was searched, I stop after 1/3 of the total target time has been used. This is _very_ dangerous. One of the test suites has a position from Cray Blitz vs Belle, 1981, the position where Bxh6 leads to a draw, Qxb6 leads to losing. We were running on a batch machine, using 1 minute per move to have time to get the move, and make it on the board, then edit a batch file to make the move when the opponent made his move. In this game, with a target of 60 seconds, Cray Blitz (this was a 10K node per second Cray Blitz, not parallel, not optimized as this was the 'beginning of the story') used 20 seconds, decided that Qxb6 was an 'obvious' move, and played it quickly. And then realized that it had royally screwed up. During the game, a couple of us started looking at Bxh6 when we saw that Qxb6 was bad... but CB didn't find it. It would have needed about 2 minutes to see that Bxh6 was a draw, but it didn't have anywhere near that...
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.