Search found 175 matches

by Stan Arts
Sun May 07, 2017 10:34 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Open and parse Winboard.debug file while in use by Winboard
Replies: 4
Views: 1097

Re: Open and parse Winboard.debug file while in use by Winbo

Thanks. Seems I might have a solution now and that the problem was actually my program putting on some sort of a lock when attempting to open to read, which Windows denies. So Google suggests opening files as a stream with no locks to get around this problem and in FreePascal that can be done with t...
by Stan Arts
Sat May 06, 2017 11:39 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Open and parse Winboard.debug file while in use by Winboard
Replies: 4
Views: 1097

Open and parse Winboard.debug file while in use by Winboard

So this morning I tried to create a simple Winboard.debug file parser in my Nemeton GUI so one can watch Winboard games in progress on a 3D board. Easy enough. When I feed it a finished debug file it parses the games in there to the last current position. Problem: Winboard or Windows denies acces to...
by Stan Arts
Wed Apr 12, 2017 12:42 pm
Forum: Computer Chess Club: General Topics
Topic: No one really knows how the most advanced algorithms work
Replies: 10
Views: 2481

Re: No one really knows how the most advanced algorithms wor

Yeah.. this is a bit counter-intuituve. One would say "the programmer knows it all", but the concept of training set makes stuff quite murky, as the algorithm calibrates in its own way. Still, it also happens in heuristic chess engines. Do we really know why they choose this or that move? Indeed. W...
by Stan Arts
Tue Apr 04, 2017 7:33 pm
Forum: Computer Chess Club: General Topics
Topic: Challenge: Which Side is the Human, Which Side is the Engine
Replies: 6
Views: 1096

Re: Challenge: Which Side is the Human, Which Side is the En

I guessed white engine. Reason being moves like black 14. ..cxd4 instead of Nxd4 as few engines would pass up putting a knight there. Also typical white queen moves such as 20. Qh6 what's it going to do there? And 23. Qe7 hint engine.
by Stan Arts
Thu Mar 30, 2017 1:07 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Q search explosion
Replies: 27
Views: 4951

Re: Q search explosion

Do you do "delta pruning" where you don't consider captures if the captured piece + margin does not get you close to alpha? A hard stop on Qsearch depth seems rough indeed but instead you could consider only doing recaptures on the same square after some depth. For example after 4+ ply. Will have yo...
by Stan Arts
Tue Mar 28, 2017 9:09 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: WAC 100
Replies: 22
Views: 5034

Re: WAC 100

Nemeton finds b6 on depth 24 and 3.5 million nodes. Main driver must be the hashtable as Nemeton is quite low on extentions or special search heuristics. When I disable all pruning besides alpha beta and the hashtable it finds b6 on depth 20 and 23 million nodes. Do you not take exact hashtable cuto...
by Stan Arts
Mon Mar 20, 2017 9:20 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: speed up or avoiding move sorting
Replies: 24
Views: 5470

Re: speed up or avoiding move sorting

You can always choose a faster and simpler sorting method near the leafs as that is where it hurts the most.
Makes some sense too to invest more effort sorting when there is a lot of depth left to go.
by Stan Arts
Sun Mar 12, 2017 10:49 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Time managment ?
Replies: 9
Views: 2254

Re: Time managment ?

Nemeton cuts target time in half when it tries to guess easy recaptures as following.: Search score is much better than static score at the root. (By >2.20 pawn.) Current preferred move is a capture and the position is still rather balanced. (> -2.50 pawn and < 2.50 pawn.) Slightly risky but simple ...
by Stan Arts
Mon Feb 27, 2017 12:39 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Generate Opening Book from Scratch
Replies: 10
Views: 2579

Re: Generate Opening Book from Scratch

Because humans are much stronger than "2400-2800" in the opening. Openings you can systematically study as the possibilities are limited and this is what has happened over the past centuries. Engines are different. They make less tactical mistakes in the middle and endgame giving them their huge rat...