Author: Aaron Gordon
Date: 23:02:50 06/25/05
Go up one level in this thread
On June 25, 2005 at 19:14:00, Arnon Yogev wrote: >Im planning to run some tests with long time controls, and i've seen in the >Readme file of Fruit that History Pruning may actually hurt play in longer >timed games. Is it safer to disable this option or should i just leave it as >Deafult?. >Thanks, Arnon. Much to my surprise, it seems to help in both blitz and standard games (not sure if you'd call 30 minutes per side standard time control though). I have two small matches running, here are the results thus far: Pentium 3 1.2GHz, 30 minutes + 1 second per side, ponder off, nunn2 opening set, 64mb hash per engine. Current score is: Fruit-DF 2.1 vs Shredder 9 (30.5-23.5, +24 =13 -17) (Fruit-DF is just a compile with Delta and Futility pruning auto-enabled) Athlon XP 2.7GHz, 3 minutes + 1 second per side, ponder off, nunn2 opening set, 384mb hash per engine. Current score is: Fruit-DF 2.1 vs Fruit 2.1 (40.0-28.0, +20 =40 -8) (For the people that are going to message me about same engine match ups, I already know this. This is an experiment I'm doing anyway..) I've run other small matchups and so far the Delta and Futility pruning has won every one of them.. including against Shredder 9. It also won a small tourney I ran that included Junior 9, Shredder 9, and Fritz 8 Bilbao.
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.