Search found 538 matches

by chrisw
Sun Jun 16, 2019 9:47 am
Forum: Computer Chess Club: General Topics
Topic: WCCC vs TCEC Chess Engine Competitions..
Replies: 33
Views: 1608

Re: WCCC vs TCEC Chess Engine Competitions..

Way past time to get off the old anti-ACM bandwagon, Way past time to get off the old ICGA bandwagon also unless they reinvent themselves. Pretty sad comment. The main thrust of the ICCA (then ICGA) was to foster interest in computer chess. They STILL provide that service via the annual mind games ...
by chrisw
Tue Jun 11, 2019 11:07 am
Forum: Computer Chess Club: General Topics
Topic: WCCC vs TCEC Chess Engine Competitions..
Replies: 33
Views: 1608

Re: WCCC vs TCEC Chess Engine Competitions..

Wccc is and has long been utterly irrelevant. A few Europeans will disagree. Their opinions are irrelevant also. Wccc is watched by a tiny smattering of European chess nerds. Tcec is watched by hoards of enthusiasts. Wccc does not even follow their own rules. I find them utterly repulsive, despite ...
by chrisw
Tue Jun 11, 2019 11:02 am
Forum: Computer Chess Club: General Topics
Topic: WCCC vs TCEC Chess Engine Competitions..
Replies: 33
Views: 1608

Re: WCCC vs TCEC Chess Engine Competitions..

FIDE and WCCC have a cosy relationship, I’m all for WCCC but to claim the title of World Champion engine seems a bit misleading to the average punter. When most people hear of World Champion in any endeavour they assume that somehow that is the best in the world. FIDE should just stick with Human c...
by chrisw
Sun Jun 09, 2019 8:29 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 1410

Re: Hash collision?

Although the likelihood of an incorrect cutoff affecting the search in a meaningful way seems small, would it be worth verifying other aspects of the hash entry to reduce collisions? For example, checking that the stored move is pseudo-legal. In the cases where the hash entry wasn’t treated as term...
by chrisw
Sun Jun 09, 2019 4:58 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: max number of pseudo legal move
Replies: 22
Views: 798

Re: max number of pseudo legal move

is there a worst position than this one [d] 3Q4/1Q4Q1/4Q3/2Q4R/Q4Q2/3Q4/1Q4Rp/1K1BBNNk w - - 0 1 218 pseudo legal moves ? I have this comment in chess source: // max width found so far in testing is 100 #define MAXWIDTH_FULLMOVELIST 128 that would have been counting every pseudo_move_generator coun...
by chrisw
Sun Jun 09, 2019 4:08 pm
Forum: Computer Chess Club: General Topics
Topic: Leela, such a great engine (and innovation)?
Replies: 36
Views: 1993

Re: Leela, such a great engine (and innovation)?

I see you only test on very fast hardware. Ok. But it gets really funny if you use weak hardware. I have a tablet where LC0 makes 1-15 NPS. I am not speaking of KNodes but Nodes. It plays very strong chess although the Android tablet is weak. I first thought it would give a good opponent for my old...
by chrisw
Sun Jun 09, 2019 12:13 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 1410

Re: Hash collision?

Indeed, score mutations would only propagate as far towards the root as the cut-moves are singular (by definition, I would say). So we live by the grace of the fact that in chess long series of singular moves are quite rare, and although they might occur in the tree, there will almost never be one ...
by chrisw
Sun Jun 09, 2019 11:29 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 1410

Re: Hash collision?

well, speaking as someone whose program(s) always contained hard to imagine bugs, it seems that any evaluation bug (for example, getting mate at a glance wrong in some rare situation, like en passent involved) will find its way back to the root and cause a loss, when evaluation thought it had a win...
by chrisw
Sat Jun 08, 2019 6:04 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 1410

Re: Hash collision?

It is not only that there are relatively few nodes close to the root, but also that these need comparatively high depth. So even if you are so extremely unlucky that you get a key collision in a node close to the root, the entry you retrieve instead will have virtually zero chance to have a usable ...
by chrisw
Sat Jun 08, 2019 5:58 pm
Forum: Computer Chess Club: General Topics
Topic: The mystery of Alex Bernstein
Replies: 25
Views: 1010

Re: The mystery of Alex Bernstein

Thank you both Roman and Sergei for providing that information on Shura-Bura's Program ! Thank you, Gerd! BTW, some small corrections on the role of Smilga. He wasn't a programmer and he never worked in Steklov's institute. He was a physicist and science journalist [1]. He wrote a paper with popula...