Author: Jeremiah Penery
Date: 16:29:31 01/23/00
Go up one level in this thread
On January 23, 2000 at 15:19:14, Tom Kerrigan wrote: >On January 23, 2000 at 14:57:16, Jeremiah Penery wrote: > >>How fast does CS-Tal run on a fast machine? Less than 20k NPS, I think. DB is >>supposed to have a bunch more evaluation, which would probably make it a bunch >>slower. > >No, the only reason you think DB has a bunch more evaluation is because Hyatt It has nothing to do with Hyatt. Please try to construct better arguments in the future. The reason I think DB has a better evaluation is that I've seen the games, and analyzed them. Kasparov and other GMs have said that DB was clearly superior to anything else they've seen. >keeps saying that it does. But none of us has any proof that DB has more >evaluation than CS Tal. And Hsu doesn't even think it has more evaluation, >because his estimate translates to at least 20k NPS. That was an estimate. He could have very well been way off, though it's not that likely. >>>I don't see why a terrific evaluation function would be right at one depth and >>>wrong at another. >>If you write an evaluation with a 5-ply search in mind, what will you do >>differently than if you have a 14 ply search with tons of extensions in mind? >>Will you really put the same knowledge with the same weights in? > >But the point is that DB's evaluation is terrific. In fact, it's so terrific >that DB running at 100k NPS can stomp all over micros. So you're right, the >evaluation function may be tuned for a 14 ply search, but it's reportedly >wonderful enough that depth doesn't make a big difference. Against the current group of micro programs, which lack a great deal of the knowledge DB has anyway, it probably didn't matter much. Against humans (GMs), it would probably matter a lot more. They could exploit any holes uncovered by an untuned evaluation much better than other programs.
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.