Author: Thorsten Czub
Date: 14:57:49 01/02/03
Go up one level in this thread
On January 02, 2003 at 16:54:31, Ed Schröder wrote: >>It's one of those interesting extensions that solve tactical problems like >>crazy, but when you play two versions against each other, they score 50%. > >Same here, exactly 50%. So in my case it isn't used as default setting because >against humans I think it's better to have some more depth than even more >tactical power. The main question i do have is: what do you conclude from : "when you play two versions against each other, they score 50%" ??? that the programs are as strong ? how do you know ? When engine 1 gets 50% against engine 1+X, does this say that engine 1+X is as strong ?? IMO your conclusion is WRONG. So your measement of strength is wrong. you believe that 50% against older versions of the same engine says: it is as strong. but this is IMO nonsense. Only when you get 50 % against ALL OTHER ENGINES you can IMO conclude that the new engine is as strong as the old was. But you don't have the time to test it out. And when you had the time, the others have new engines ! It's a dilemma without way out. >>Singular extension is another one. I can double tactical speed with singular >>extension, at the cost of a ply or sometimes two, but the program plays almost >>exactly the "well" as it did before. HOW do you measure "the program plays almost >>exactly the "well" as it did before" ??? You can only measure it by playing against Fritz8 Fritz7 Shredder7 Tiger15 ... and relate this with the score you had with the version before. It will not work if you let it run through a test-suite. It will not work if you test it against your own engine. thats not possible. >Same here, how coincidental. > >Ed yes.
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.