Author: Ralph Stoesser
Date: 12:07:30 07/06/03
Go up one level in this thread
On July 06, 2003 at 14:39:24, Uri Blass wrote: >On July 06, 2003 at 12:04:16, Ralph Stoesser wrote: > >>Config: >>Dual PIII-S 1266 MHz, 1GB PC-133 RAM >>Both CPU enabled (~1100 kN/s) >>64MB hash, 3 MB pawn hash >>3,4,5 men nalimov >>Opening book: remis.ctg (write protected) >>Time control: Most of the games were played in 5 1 (min: 3 1, max: 15 3) >> >>Results so far: >>Elo on Chessbase server after 60 games: 2320 (max Elo: 2458 after 35 games) >>Opponents average Elo: 2447 (taken from 42 different opponents, mostly top >>commercial engines) >> >> >>That's disappointing so far. >>I will let it play further until 100 games. > >I do not understand the meaning of the rating when I know nothing about the >opponents and their hardware. > >I changed the title because the original title may mislead people to read your >post when they can understand nothing from it(even not why are you disappointed >because you did not give rating of other programs on the same hardware for >comparison). > >Uri Hello Uri, For sure this is not an accurate test. I should have added a few more points. Maybe these points make the test somewhat more pursuable, still in a fuzzy way, of course. -My hardware is about average from the opponent ones.(This is my estimate from reading several personal info cards from the opponents) -I've done this test with Deep Fritz 7 a few days ago on the same HW and it scored >2750 after 100 games. The opponents in both tests were selected automatically from a pool of about 80 opponents total, where the possibility of getting an opponent with about the same Elo as myself's actual Elo is very high. Best regards, Ralph
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.