Page 1 of 1
Is there a fen suite to ...
Posted: Fri Oct 09, 2020 10:00 pm
by Mike Sherwin
... check the accuracy of a material only search at depth 1 to n when engine has no extensions (except for qsearch that is captures only) or reductions?
Re: Is there a fen suite to ...
Posted: Fri Oct 09, 2020 10:08 pm
by Dann Corbit
I don't think such a thing would make sense except for a minimax version which would not be useful.
Even alpha-beta would be hard to make repeatable unless the move ordering was somehow idealized.
And if you have pruning like null move, then the results lose their meaning.
Out of curiosity, what is it that you want to measure with such a test suite?
Re: Is there a fen suite to ...
Posted: Fri Oct 09, 2020 10:22 pm
by Mike Sherwin
Dann Corbit wrote: ↑Fri Oct 09, 2020 10:08 pm
I don't think such a thing would make sense except for a minimax version which would not be useful.
Even alpha-beta would be hard to make repeatable unless the move ordering was somehow idealized.
And if you have pruning like null move, then the results lose their meaning.
Out of curiosity, what is it that you want to measure with such a test suite?
I have a new engine, Bricabrac, that is currently a material only searcher. It does not have null move, yet. All I want to verify is at depth n does it find material truth for depth n. Thanks.
Re: Is there a fen suite to ...
Posted: Fri Oct 09, 2020 10:25 pm
by Dann Corbit
Is this pure wood count, or do you take material imbalance into consideration also?
Re: Is there a fen suite to ...
Posted: Fri Oct 09, 2020 10:30 pm
by Mike Sherwin
Dann Corbit wrote: ↑Fri Oct 09, 2020 10:25 pm
Is this pure wood count, or do you take material imbalance into consideration also?
Pure wood count. 100,300,300,500,900
Re: Is there a fen suite to ...
Posted: Sun Oct 11, 2020 10:46 pm
by Ferdy
Mike Sherwin wrote: ↑Fri Oct 09, 2020 10:30 pm
Dann Corbit wrote: ↑Fri Oct 09, 2020 10:25 pm
Is this pure wood count, or do you take material imbalance into consideration also?
Pure wood count. 100,300,300,500,900
Uploaded
material epds in github.
pawns.epd
Code: Select all
r4rk1/p2b2pp/1q2pb2/3p3Q/3p4/3B4/PP3PPP/R1B1R1K1 b - - ce 100;
r7/p2q1rk1/2p4p/4P1p1/1BQ1Ppb1/1P6/P1P3PP/4RRK1 w - - ce 200;
ce values are side pov. First position black is ahead by 1 pawn. In the second, white is ahead by 2 pawns.
Re: Is there a fen suite to ...
Posted: Sun Oct 11, 2020 11:51 pm
by Mike Sherwin
Ferdy wrote: ↑Sun Oct 11, 2020 10:46 pm
Mike Sherwin wrote: ↑Fri Oct 09, 2020 10:30 pm
Dann Corbit wrote: ↑Fri Oct 09, 2020 10:25 pm
Is this pure wood count, or do you take material imbalance into consideration also?
Pure wood count. 100,300,300,500,900
Uploaded
material epds in github.
pawns.epd
Code: Select all
r4rk1/p2b2pp/1q2pb2/3p3Q/3p4/3B4/PP3PPP/R1B1R1K1 b - - ce 100;
r7/p2q1rk1/2p4p/4P1p1/1BQ1Ppb1/1P6/P1P3PP/4RRK1 w - - ce 200;
ce values are side pov. First position black is ahead by 1 pawn. In the second, white is ahead by 2 pawns.
Thank you Ferdy! This should be very helpful.
