Author: Tom Kerrigan
Date: 12:59:21 08/12/98
Go up one level in this thread
My reasons for saying it's not perfect are similar. Basically, you're changing the amount of time the programs spend thinking about moves to something unrealistic. My personal take on this is that it isn't a big deal, but it's still not the same as using two computers... -Tom On August 12, 1998 at 14:07:10, fca wrote: >On August 12, 1998 at 13:45:18, Tom Kerrigan wrote: > >>Basically the only reasonable way to do this is turn of "permanent brain" in >>both programs. >>This is not a perfect solution, but it's better than other things you can do, >>like trying to balance CPU time... > >I agree. > >The reasons I believe it is imperfect are that: > >(1) You are differentially penalising the program that is better at anticipating >the enemy's moves. This is not *necessarily* the program that would be better >in a 2-machine match (though it probably is), as one program may even be tuned >for a particular opponent insofar as anticipating *that one's* moves. I believe >no one does this yet, but who knows in the future what may come... > >and, less importantly > >(2) even if CPU % was being balanced with permanent brain on, the human would >not be getting the best time-value from his machine. This is because time spent >by the program with the move is "more valuable" than time taken by the second >one - after all, all of the second one's thinking may be about positions none of >which can possibly occur. So to get the "best" combined playing strength, CPU >time would need to be allocated differentially depending on who has the move. > >I have another but it is a bit far-fetched. > >Were your's the same, or do you disagree, or have any other ones, please? > >Also, what about factors other than permanent brain? > >Kind regards > >fca
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.