TalkChess.com

Look at the learning graphs on page 4: https://arxiv.org/pdf/1712.01815.pdf Yes, it is the normal behavior and most NN learning graphs look like that. However, that is for the particular net parameters of that training run. There are many hyper-parameters that change from one net to another. A some...

PPS If you would enjoy more chess and less period drama, you may like reading the book also.

*** Warning: Spoilers Alert *** I have also been enjoying QG very much, and have watched it 5+ times. As to it being realistic about actual chess, while it may not be completely accurate (which would be intolerably boring for all but a few diehards), many have said it is in fact far more realistic t...

Your approach is theoretically sound, but not practical for human "hand crafted" changes.

It is pretty much what net training is doing with now into the billions of positions, using automated processes of course.

The idea is that the test suite is not going to be WAC 300 but based on the positions that the engine went wrong... I think you missed my point. Test suites generally are minimally helpful; it does not matter how many positions there are. Actual game play against several opponents (in addition to m...

permutations??

Suggest: https://www.chessprogramming.org/Zobrist_Hashing

A FEN string would be fine as the side to move, ep square (if any) and castle status should also be encoded.

Your approach has indeed been done for several decades.
It works at first, but then it becomes clearer that tuning for a test set is not the same as actual game strength.
Recall the early focus on WAC 300 scores...

Works just fine for me from a cmd window...

Ras wrote: ↑

Thu Nov 26, 2020 7:35 pm

brianr wrote: ↑

Thu Nov 26, 2020 7:07 pm

Is there a tourney manager that uses the higher precision timers?

I'm not on Windows anymore, but under Linux, c-chess-cli has very precise timing (via clock_gettime with monotonic clock).

Thanks, I have not tried that one yet.