Author: Dave Gomboc
Date: 20:58:06 04/23/99
Go up one level in this thread
On April 23, 1999 at 16:19:17, KarinsDad wrote: >On April 23, 1999 at 14:43:33, José de Jesús García Ruvalcaba wrote: > >[snip] >> >> And even more interesting would be a comp-human match with adjournments! >>Imagine the machine analyising for 18 hours and the human (I am thinking of a >>grandmaster) working with her/his team of analysts. >> Most programs should be modified to handle adjournments; because when you >>adjourn a game there is always the other player's turn for your perspective. I >>think current programs choose a move to "ponder" on when it is the opponents >>turn, and that would be disastrous if the opponent actually makes another move >>(imagine wasting 18 hours of computer time). >> > >My program is being written to not only ponder on the PV, but also on other >extremely likely responses. So effectively, it is almost like it changes sides >to move and has a PV, but it also saves other moves and their PVs (and some of >their more likely subPVs) so as to minimize "recalculating nodes" when the >opponent chooses a move other than the PV. > >I do not know how other programs handle pondering. > >KarinsDad :) Is this some sort of best-first ponder search, then? :) Dave
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.