Author: Omid David Tabibi
Date: 18:08:57 12/09/03
Go up one level in this thread
On December 09, 2003 at 06:47:33, Uri Blass wrote: >On December 08, 2003 at 13:52:27, Omid David Tabibi wrote: > >>On December 08, 2003 at 11:35:03, Robert Hyatt wrote: >> >>>On December 08, 2003 at 11:07:25, Omid David Tabibi wrote: >>> >>>>On December 08, 2003 at 10:56:56, Robert Hyatt wrote: >>>> >>>>>On December 08, 2003 at 10:42:01, Omid David Tabibi wrote: >>>>> >>>>>>On December 08, 2003 at 01:35:02, Slater Wold wrote: >>>>>> >>>>>>>On December 08, 2003 at 01:24:46, Steve Lim wrote: >>>>>>> >>>>>>>He's on vacation. Won't be anymore games for like another week. >>>>>> >>>>>>What's the result so far? (and what exactly was the time control?) >>>>> >>>>> >>>>>The first test-match was at roughly 10:1 time odds. I don't remember >>>>>the draws (I think 5) but the win/lose score is 4:2 so far in Crafty's >>>>>favor, if I didn't miss a result. >>>> >>>>1 minute is just a pure tactical combat. A fair 10 to 1 time control would be >>>>100 minutes to 10 minutes, or even 50 minutes to 5 minutes... >>> >>> >>>The games are 40 moves in 2 hours. DJ is simply running on a machine 10x >>>slower. >> >>OK. Thanks for the info. With that regard, 4:2 for Crafty is very impressive. >>Had I had money on the bet, I would be on the losing side now. >> >>BTW, is Junior using its tournament book? >> >>Without its own book, Junior seems to be far weaker. For example, before Graz I >>conducted 6 matches of 1 hour per game between Falcon and Junior 8, using >>fritz8.ctg for both sides. The result was 4.5 - 1.5 for Falcon. >> >>Others results were: >> >>1 hour per game >>both sides using fritz8.ctg up to move 10 >> >>Falcon - Shredder 7.04 3.5 - 2.5 >>Falcon - Fritz 8 3.0 - 3.0 >>Falcon - Junior 8 4.5 - 1.5 >>Falcon - Deep Sjeng 1.0 4.0 - 2.0 >>Falcon - Hiarcs 9 4.5 - 1.5 >>Falcon - Chess Tiger 15.0 3.0 - 3.0 >>Falcon - Comet B60 5.5 - 0.5 >>Falcon - Crafty 19.03 5.0 - 1.0 >>Falcon - Ruffian 1.0.1 4.0 - 2.0 >>Falcon - SOS.2 for Arena 6.0 - 0.0 >>Falcon - SOS.3 for Arena 3.5 - 2.5 > >excellent result for falcon. >No losses in your matches when both programs use the same Fritz8.ctg. In Graz I had the problem of having an automatically generated book versus engines with specially prepared books (for example, in the game against Fritz, Falcon was practically lost after the opening phase). But most of all, I think Falcon's poor hardware contributed to the bad performance. Falcon's style is very risky and aggressive. On the same hardware, Falcon outsearches all the engines I tried, so complicating the matters always pays off. In Graz however, Falcon complicated the matters, and then got outsearched by its opponent's processing power. For example, on the same processor Falcon's nps is higher than that of Shredder. But in Graz Shredder was running at several million nps while Falcon was at about 400k nps... > >Uri
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.