Author: Dann Corbit
Date: 08:01:20 08/22/05
Go up one level in this thread
On August 22, 2005 at 05:41:04, Uri Blass wrote: >On August 22, 2005 at 04:07:18, Peter Skinner wrote: > >>On August 22, 2005 at 02:02:15, Uri Blass wrote: >> >>>I think that you do not realize that 1 free engine that is superior is enough to >>>make them worry. >> >>I don't think so. So one or two engines _might_ be able to take a few games off >>them here and there. >> >>The random factor kicks in here though. If the commercials were to top the IPCCC >>or WCCC next year, then there is nothing to talk about. >> >>It seems one or two tournaments justify allot of conjecture and hand waving.. >> >>>No way >>> >>>I expect fruit WCCC to beat the commercial programs in a match of 50 games >>>if one processor is used and there is no learning problem for fruit. >> >>I would sell my ex-wife into slavery if that happened (well I'd probably do this >>anyway.. but you get the point). You _seriously_ underestimate the Chessbase >>interface most commercials run under. http://www.husvankempen.de/nunn/rangliste.html For the individual breakdown: http://www.husvankempen.de/nunn/rangliste.html#3 Fruit has a positive score with more than 50 games against every commercial chess engine besides Shredder. These games are 40/40 but I do not expect a large difference at 40/2. This test also confirms the results: http://www.utzingerk.com/fruit21_test.htm where Fruit has a positive score against every chess engine except Chess Tiger, where Fruit scores even. Fruit's good result is not an accident. >I think that you understimate fabien. I agree. >> >>>I expect Fruit to top the SSDF if fabien only send them fruit WCCC with >>>learning. >>> >>>Uri >> >>I don't think that based on the number of opponents it would face that it would >>steal the #1 spot away from Shredder. Then again I don't even think it would >>steal the #2 stop from Junior (removing all previous Shredder versions puts >>Junior #2). > >I do not think Fruit2.1 will do it but I guess that the ssdf is going to test >better fruit that fabien will release later this year(it is going to have the >improvement in fruit WCCC and it is also going to have book learning). >> >>You also bring up learning, yet no more than a week ago you stated it wasn't >>important at all because the number of possible positions it could face. > >I do not think that I said that learning is not important for long matches like >the ssdf does but I expect fabien to implement learning before sending fruit to >the ssdf testers. When Fabien writes a SMP version, he will be world champion. I am sure of it.
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.