Author: Uri Blass
Date: 22:43:27 10/01/01
Go up one level in this thread
On October 02, 2001 at 01:36:28, Dave Gomboc wrote: >On October 01, 2001 at 21:45:44, Dann Corbit wrote: > >>On October 01, 2001 at 21:39:41, Dann Corbit wrote: >> >>>On October 01, 2001 at 21:20:08, Bruce Moreland wrote: >>> >>>>On October 01, 2001 at 15:30:41, Dann Corbit wrote: >>>> >>>>>On October 01, 2001 at 15:18:44, Bruce Moreland wrote: >>>>> >>>>>>To be fair, this is because it thought g4 was also a draw. If Yace thought >>>>>>that, too, there isn't a difference. >>>>>> >>>>>>I'm not trying to apologize for Crafty, but it is dangerous to assume that just >>>>>>because one program chose the fork in the road that did not lead to an army of >>>>>>demons, it means that it could see them. >>>>>> >>>>>>Maybe it saw them, I don't know. >>>>> >>>>>Of course, every good program will shine on some position and not do as well on >>>>>others. >>>>> >>>>>How does Ferret do on this one? The machine I used was 950 MHz Athlon for >>>>>comparison. >>>> >>>>It took 35 seconds on a dual 1.2 ghz machine, with tables (but not KPP vs KP). >>>>The score was 0.00. Without tables it's been going for a while and hasn't found >>>>it yet. >> >>I am curious about why you don't have kppkp. It's only 50 megabytes: >>05/10/2000 12:00a 34,118,774 kppkp.nbb.emd >>05/10/2000 12:00a 28,762,929 kppkp.nbw.emd > >AFAIK Bruce uses his own code for tablebases. Perhaps he didn't generate KPPKP >due to en passant issues. > >Dave Maybe he did not generate tthe kpp vs kp tablebases because it is the most complicated tablebases to generate(you need all the other 5 piece tablebases except tablebases of king and 3 pieces against king in order to generate the KPP vs KP tablebases). Uri
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.