Author: Vincent Diepeveen
Date: 10:02:07 08/12/99
Go up one level in this thread
On August 12, 1999 at 09:32:40, Robert Hyatt wrote: >On August 12, 1999 at 06:08:00, Vincent Diepeveen wrote: > >>On August 11, 1999 at 15:56:20, Michel Langeveld wrote: >> >>>Great information but I think nobody reacts because this info is too hot >>>for this moment. As I global read your concept it seems to me some things can be >>>optimized a little more with a some MORE thinking. >>> >>>I have 1 remark: if you don't store mate in more than 50 or less than 50 this >>>means you might give away a point of you 're playing against another computer >>>with no tablebase or a human. >> >>I thought long about that yeah. >> a) it's easier to implement, that's why i have DTC right now >> b) above argument can be countered as: >> you will draw important games and a mistake i see at >> icc a lot is that a program doesn't CARE to what EGTB endgame >> he exchanges. So if it's *possible* to get to KNN KP then >> it'll do it. Even if it's a mate in 221 >> and in chess it's not smart to assume they will make a big mistake >> after which a mate in 221 becomes mate in 20. >> >> For example i've seen a lot of computers blow KBR KR, but personally >> i would not blow it, and i'm just 2275. Yeah in blitz perhaps, but >> blitz is not my designpoint. >> > >I'd bet you would blow it pretty often, because a bunch of those are nothing >more than draws anyway. indeed, that's why i try to stay out of DTM, as DTC is simply a bounded form of DTM. It's just a sure bound you have. > > >> >>>Keep doing the good work and in a few days you will get support! >> >> >> >>>Regards, >>> >>>Michel Langeveld
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.