Author: Chessfun
Date: 18:32:52 11/18/00
Go up one level in this thread
On November 18, 2000 at 19:51:13, stuart taylor wrote: >Sarah, >1. Can you add another program into all this testing easily, e.g. Shredder 5 >when it is released? I have no plans at present to invest $150 Can in S5. I could change my mind in a month when christmas comes. I have Chessmaster 8000 but am not prepared at this point to play some 28 games manually...ohhh maybe I'll see. >2. Are you sure that Fritz 6b was the one to be tested, and not 6a? Isn't 6a >known to be a bit stronger? (or is it an improved version of 6b)? As far as I know both a and b are identical but b fixes minor bugs that occured with tablebases. The version playing is for certain b. >3.It is interesting that the two Tigers are quite ahead of the field, but I >don't know if that is a reliable indication that they are THAT much better. >(for various reasons, part of which, see item 4) IMHO opinion the Two tigers will be 1 and 2 on the SSDF list when they first are tested. However I don't believe they are that much better than "the field" IMO maybe 30 SSDF points or so. I don't see item 4 as bearing any relevance to whether they are stronger or not. Fritz 6 is still an excellent program that has led the SSDF for almost a year. Fritz 7 will also shortly follow, but for now IMO the Two Tigers are the strongest available chess engines. This excludes Shredder 5, Nimzo 8 and Chessmaster 8000 all of which are so new that I don't have an opinion either way. And in the case of Shredder 5 don't have it. >4.But since you are doing such intensive testing, it's a shame you didn't >incorporate my idea of including a small amount of 1)non-book openings, and >2)fischerandom openings, and an even number of 3)k-pawn and 4)Q-pawn openings >(book) and all the above (categoties of openings) the same amount of games for >each colour. > I thought such a plan would give much more food for thought and assesment of >program strength etc. But now, I suppose you've already done too much like this. >OR perhaps you disagreed with me, but hardley anyone replied to my postings of >these ideas. But I would have been happy to hear why people disagreed, OR to >actually do it according to some similar plan, if they didn't disagree. I saw your ideas when you posted the original post. And as you say it is your idea. You are welcome to play as many games as you like. I will even myself participate in the threads on the subject. However to actually play out what you suggested would take more time and is something I have little interest in, especially the fischerandoms. While I myself have played Nunn matches which are without books and have tried programs against testsuites, I feel the book is a part of the program and the program should be tested as such. Similarly with tablebases if a program uses them I see little interest in playing the program without them. While playing such games would naturally give some "food for thought" the actual testing parameters would make the task monumental. Sarah.
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.