Author: Stephen Ham
Date: 14:47:10 12/20/04
Go up one level in this thread
On December 20, 2004 at 17:10:57, George Tsavdaris wrote: >On December 20, 2004 at 16:32:24, Jouni Uski wrote: > >>Shredder 8 reaches quite stunning averaga search dept. E.g. with only 15 >>second/move average with Pentium 2400 and 128MB hash it averages 15,2 ply! For >>comparison Fritz Bilbao and Gandalf 6 reach both same 12,7 ply. No wonder >>Shredder beats them 13 - 7 and 14 - 6. Of course 2,5 ply is a lot difference! >>Is there any amateur or pro engine, which can equal Shredder in depth - Junior's >>half plys can be forgotten. >> > > SOS 4 does a very deep search too. In some positions it is competitive even >with Shredder. (At number of plies not judgement ability). But i don't really >believe that this is its(Shredder) secret. Thorsten suggests that the secret >maybe the high positive scores but again i can't find a reason behind that. So >what, if Shredder evaluates high the positions? What does this has to do with >playing strength? I see this only as a feature of Shredder and not as an >advantage. Perhaps i am wrong..... > > But why Shredder should has to have a big secret? It is just the strongest >program the last years and maybe this is not due to a big secret, but just be >slightly better from the competition in many areas. Slighly better >pruning/evaluation/play at closed positions/endgame, ...etc. It doesn't have to >be an area that has a secret.......This doesn't mean of course it hasn't one:-) Dear George, When it comes to computer chess engines, then I'm the least knowledgeable person here. But one thing that I've "learned" from others here is: each engine evalutes plys differently. So just because Haircs calculates to X-plys as a function of time, while Shredder 8 calculates the same position to X+ plys in the same time, that doesn't mean that there's a true comparison in the ply count. I recall posts here where Dr. Hyatt stated that Deeper Blue's ply search was far more thorough and had more evaluations than the best engines of today. So that is further evidence of a quantity versus quality approach to position searches. Further, Hiarcs 9 has probably the lowest ply search, per unit of time, than any top engine. Yet, it's a very fast calculator and awesome at speed chess. As such, I've seen relatively weak engines spot tactical refutations in fewer ply than Shredder. In fact, that seems to be frequently the case. So my opinion is that they're measuring ply differently. Perhaps Shredder 8 is examining a greater quantity of ply, but doing so with less quality per ply. The previous Shredders displayed some ability in closed positions, compared to other engines of their day. But IMHO, Shredder 8 doesn't stand out as a superior engine in closed positions, copared to other recent engines. In fact, I can think of several better ones that are even older than Shredder 8, based upon my tests. Please read my review of Shredder at Chesscafe.com for examples of this. Since then, I've found a few more. All the very best, Stephen
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.