Author: Mike S.
Date: 13:51:50 06/16/02
Go up one level in this thread
On June 16, 2002 at 14:12:43, Christophe Theron wrote: >On June 16, 2002 at 13:26:42, Mike S. wrote: > >>(...) >>But sooner or later, I think there will be (or should be) a "settings switch" >>normal gameplay / analysis in chess programs, to change several parameters to >>adapt to these to major ways engines are used. For example, more "test >>suite-like" focus on tactics for the analysis mode. (...) >>So the best engine might be the one which analysis best, not the one which is >>best in comp-comp games. I'm beginning to see comp-comp games as a sport (only), >>but people usually don't buy cars to compete in car races. >The problem is that in "combination find" modes the chess engines are notably >weaker in positional play (what to do when there is no combination to find). > >So I'm not sure that people using chess programs would be more happy with such >modes. I don't assume a club player (or better) will use engines for positional analysis. I may be wrong. But people could switch back to gameplay mode anyway, i.e. to get an evaluation for a "silent" position or the like. My idea was a settings switch, independant from the operation mode. Might even be useful to play a more tactical orientated blitz game against humans (play using the analysis setting...). This is a developement which is yet to come IMO. Maybe it's too soon. For example, I think that Nimzo 7.32 / 8 are most probably still among the best 2 or 3 engines for quick analysis (because I usually get excellent tactical test results from a similar nimzo program). But in gameplay against programs, it's a bit behind now. Which means, rating lists don't give best advise to people who (mostly) want to analyse their own games with their chess programs. SSDF rank is easier to communicate, but is not identical to the performance relations in analysis IMO, or to other not directly success-related things (i.e. "beauty" of play from a human viewpoint, if you think of 150+ moves avoiding 50 move draws for example). This *could* become a problem if people start to feel there's too much optimization for comp-comp results (only). We'll see. Regards, M.Scheidl
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.