Author: Don Dailey
Date: 13:47:43 01/17/98
Go up one level in this thread
Hi Stuart, You did not include any information telling us if the program plays any stronger. In the past I've discovered incredible algorithms that help just about any problem sets but actually make the program weaker. I'm not saying this is happening here, just be careful. - Don On January 17, 1998 at 15:45:59, Stuart Cracraft wrote: >I did a trivial experiment to see the effect of extensions >on a combination/mate-based suite like Win-at-Chess. It shows >the power of extensions and that you shouldn't ignore them. >They are incredibly helpful for guiding the search. > >At the time the test was run, my program (with extensions) >solves (at 5 seconds per move on a 25mhz 486) 192 out of 300 of >the Win-at-Chess problems. For comparison, the latest >Crafty gets a tad over 200 problems correct with a 5 >second limit on the same machine. > >The extensions that were included in the above result were: > >Extension Set "A" > > getting out of check > up to 2 recaptures beyond the 2nd ply > pawn advances to 7th rank > pawn advances to 6th rank that are captures > >In the past, I've also played with but did not include: > >Extension Set "B" > > threat extensions > passed pawn extensions > one legal move only extensions > mate-threat-in-null-move extension. > >The reason is that I've not had enough time to debug these and >make them efficient or even determine their efficacy. When I took out >extension set "A" above, and reran on Win-at-Chess with no extensions at >all, the score on Win-at-Chess dropped to 123 solved out of 300, a drop >of 69 problems solved (from 192 out of 300 with extensions). This >is a drop in total WAC correct from 64% to 41% using my >limited time (5 seconds) and limited hardware (486) experiment. > >So I guess I'd have to say that extensions are damned helpful! > >fyi -- my program does not yet use fractional extensions and >if it awards an extension at a certain depth, it does not consider >others and will just increase the depth by one ply only for all >successors to this position that are searched. > >--Stuart
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.