Author: Nobuhiro Yoshimura
Date: 18:45:43 06/08/98
Go up one level in this thread
On June 08, 1998 at 11:55:17, Bruce Moreland wrote:
>
>On June 08, 1998 at 05:18:16, Jouni Uski wrote:
>
>>If we compare programs A and B in tactical test suite (WAC) and A solves
>>293 positions with average solution time 5 sec, but B solves 298 with
>>average speed 60 sec - so which one is better tactically? I think
>>it's quite a problem to define.
>
>Maybe the same, since you get almost all of the WAC problems in a few
>seconds.
>
>Even if one gets a few more in the same amount of time it's no *strong*
>indication of anything, really.
>
>Mine gets 286 in 5 seconds on a P6/200, but takes 53 seconds to get to
>295.
>
>That is "find and hold", so if I'd stopped at 5 seconds rather than
>running a whole minute, it might have gotten a couple more, I don't
>know.
>
>I can make versions that get more right, but which I don't like as much.
>
>bruce
When you modify your search engine or evaluation functions and test
aginst
many test suites, can you tell me how to decide whether you are going
to
adpot or reject the modification.
For example:
before) 5sec 90% correct and 10sec 95%correct
after) 5sec 85% correct and 10sec 98%correct
The above are just a made up number so that you donot have to
answer specificly, but what kind of guidelines do you use to
determine whether you adpot or reject a new modification?
Thank you in advance
Nobuhiro
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.