Search found 2833 matches

by Rebel
Sun Aug 09, 2020 9:53 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

Uploaded the latest version, SOMU 1.7a, follow the "new" icons.

http://rebel13.nl/download/utilities.html
by Rebel
Sun Aug 09, 2020 8:19 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

Another idea is to create opening specific PST sets, here is an example for the http://rebel13.nl/text/sicilian.txt

But I am more thinking to make sets engines have trouble with such as closed position like the Dutch defence, or the Kings Indian, Beoni and where to put the pieces.
by Rebel
Sun Aug 09, 2020 7:17 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

Things are better now, maybe (in the end) it can be used for evaluation as well. PGN : pgn\elo-2600.pgn Games 82.208 Phase : Middle Game White Pawn +-----+-----+-----+-----+-----+-----+-----+-----+ | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 8 +-----+-----+-----+-----+-----+-----+-----+-----+ | 0 | 1 | 1 | 2...
by Rebel
Sat Aug 08, 2020 7:23 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

Things are better now, maybe (in the end) it can be used for evaluation as well. PGN : pgn\elo-2600.pgn Games 82.208 Phase : Middle Game White Pawn +-----+-----+-----+-----+-----+-----+-----+-----+ | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 8 +-----+-----+-----+-----+-----+-----+-----+-----+ | 0 | 1 | 1 | 2 ...
by Rebel
Sat Aug 08, 2020 5:56 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

There is no tuning, just a raw statistic from a large quality PGN, not suited for evaluation. Code: for each move : pst[piece_type][square_from]--; pst[piece_type][square_to]++; That's basically all there is except handling a few side effects . How is this possible then? A white pawn on d2, for exa...
by Rebel
Sat Aug 08, 2020 10:22 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

There is no tuning, just a raw statistic from a large quality PGN, not suited for evaluation. Code: for each move : pst[piece_type][square_from]--; pst[piece_type][square_to]++; That's basically all there is except handling a few side effects . How is this possible then? A white pawn on d2, for exa...
by Rebel
Fri Aug 07, 2020 2:07 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Re: Idea #13 [experiment]

There is no tuning, just a raw statistic from a large quality PGN, not suited for evaluation.

Code: for each move : pst[piece_type][square_from]--; pst[piece_type][square_to]++;

That's basically all there is except handling a few side effects.
by Rebel
Fri Aug 07, 2020 10:19 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea #13 [experiment]
Replies: 13
Views: 512

Idea #13 [experiment]

Creating meaningful PST's from PGN A new tool is available that creates the 12 Piece Square Tables (PST) for the MiddleGame from a PGN database. The tables can be used for 1) move ordering, 2) pruning decisions, 3) reduction decisions. Example-1 : PST made from a 82.000 human game collection between...
by Rebel
Wed Aug 05, 2020 7:55 am
Forum: Computer Chess Club: General Topics
Topic: Somu 1.6 (tool) question for Ed ...
Replies: 15
Views: 460

Re: Somu 1.6 (tool) question for Ed ...

I can confirm the error with the following small pgn-file with only 4 games. http://www.fastgm.de/schach/test.pgn Engine Depth Time Games Moves Average Forfeit Book Depth MIDG EARLY ENDG LATE Demolito 2020-01-27 25.25 4:04:17 4 250 58.63 0 32 8.00 24.61 | 26.00 | 26.04 | 26.30 Stockfish 11 37.47 3:...
by Rebel
Tue Aug 04, 2020 7:27 am
Forum: Computer Chess Club: General Topics
Topic: Somu 1.6 (tool) question for Ed ...
Replies: 15
Views: 460

Re: Somu 1.6 (tool) question for Ed ...

Hi Ed, but all are playing with the same book (openings are allways, 3 moves after ECO codes formed). Fix depths for opening books are bad because many of ECO codes never played (for an example 8 moves if ECO codes started with 9, 10, 11 moves). The average of book moves after so many games should ...