Search found 112 matches

by Andrew
Tue Nov 24, 2020 7:54 am
Forum: Computer Chess Club: General Topics
Topic: Crystal 3.0
Replies: 32
Views: 6000

Re: Crystal 3.0

Decided it was an appropriate time for release. Changes since 2.0: Various bugfixes mostly due to members of the community pointing out usage issues to me. Some minor eval smoothing (simple linear interpolation between previous value). Introduction of NNUE. For now, NNUE is either on or off, there ...
by Andrew
Wed Nov 18, 2020 7:05 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: NNUE accessible explanation
Replies: 16
Views: 6026

Re: NNUE accessible explanation

Thanks Rein I hadn't seen that. Having a look now!

Andrew
by Andrew
Tue Nov 17, 2020 9:17 am
Forum: Computer Chess Club: General Topics
Topic: Chessbase patent
Replies: 12
Views: 1345

Re: Chessbase patent

Maybe this is a patent troll just waiting to make money because they applied for the patent. Should not be allowed because this has been in the open domain for a very long time. THey were also the same inventors for the https://patents.google.com/patent/US5678001A/en?oq=5%2c678%2c001 patent i.e. th...
by Andrew
Tue Nov 17, 2020 8:21 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: NNUE accessible explanation
Replies: 16
Views: 6026

Re: NNUE accessible explanation

3) When it comes to training, all that I know so far is Texel's tuning method which I recently used to improve my checkers engine (http://www.fierz.ch/cake186.php) by using logistic regression on win-loss-draw information on a few million (N) positions to improve the weights of my handwritten eval ...
by Andrew
Tue Nov 17, 2020 7:51 am
Forum: Computer Chess Club: General Topics
Topic: Chessbase patent
Replies: 12
Views: 1345

Re: Chessbase patent

I'll just add that as far as I can tell from the inventor/applicant and other data on this patent it has no relationship to Chessbase
the company. But correct me if I'm wrong!

Andrew
by Andrew
Tue Nov 17, 2020 6:42 am
Forum: Computer Chess Club: General Topics
Topic: Chessbase patent
Replies: 12
Views: 1345

Re: Chessbase patent

Check claim 1 and the abstract. They're getting an engine to analyze a set of game moves, storing top ranked moves (whatever that means) in a database, some more steps, generating from the database "instructional" puzzles. Not sure if it would be patentable subject matter here in Aus. At the start t...
by Andrew
Fri Nov 13, 2020 6:42 am
Forum: Computer Chess Club: General Topics
Topic: A new website for Shredder ...
Replies: 39
Views: 3477

Re: A new website for Shredder ...

I also think it was Shredder 12 that had position learning (done at every move) which was removed
in 13. Hopefully any new versions will have this restored!

Andrew
by Andrew
Fri Nov 13, 2020 6:26 am
Forum: Computer Chess Club: General Topics
Topic: A new website for Shredder ...
Replies: 39
Views: 3477

Re: A new website for Shredder ...

If you're ever searching for a mate score from a position, Shredder 13 is a great program to have in your toolkit. Sometimes it will perform much worse than other recent engines, but quite often it will perform much better. Not sure why this is, if there is something its search code that is digging ...
by Andrew
Tue Nov 03, 2020 8:43 am
Forum: Computer Chess Club: General Topics
Topic: Komodo 14.1 Release and Dragon Announcement
Replies: 114
Views: 11088

Re: Komodo 14.1 Release and Dragon Announcement

I'd love to see these engines train networks specifically for KP and KPR endings. They are the most common in normal play. However the big question is how big they would need to be useful. 1Mb? 10 Mb? ELO advantage, tiny or reasonable? Maybe useful studying certain middle or endgame positions/studie...