Search found 269 matches

Sun Aug 16, 2009 4:47 am
Forum: Computer Chess Club: General Topics
Topic: Announcing Daydreamer 1.0
Replies: 15
Views: 2883

Re: Announcing Daydreamer 1.0

Hi Leo, Daydreamer and I are from the USA. Thanks for adding Daydreamer to your page, I appreciate the publicity.
Sun Aug 16, 2009 12:40 am
Forum: Computer Chess Club: General Topics
Topic: Announcing Daydreamer 1.0
Replies: 15
Views: 2883

Announcing Daydreamer 1.0

http://cloud.github.com/downloads/AaronBecker/daydreamer/daydreamer.jpg Daydreamer 1.0 (UCI), by Aaron Becker http://github.com/AaronBecker/daydreamer/tree/master I've just released the first version of Daydreamer, the chess engine I've been working on this summer. It started as a side effort in th...
Thu Aug 13, 2009 8:08 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Proving the correctness of the algorithm
Replies: 30
Views: 3743

Re: Proving the correctness of the algorithm

The 291/300 score was 30s per move on a 1.8 GHz Core Duo chip (the older 32 bit variety). With 5s per move it scored 276/300. edit: I would speculate that a sophisticated eval might actually hurt you on WAC. The problems are tactical, so the extra time spent in evaluation might be profitably spent s...
Wed Aug 12, 2009 11:04 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Proving the correctness of the algorithm
Replies: 30
Views: 3743

Re: Proving the correctness of the algorithm

You don't need positional knowledge to solve any of those problems. As long as you're counting material correctly and searching deeply enough (apparently 6 plies should suffice for those problems), you will find the solutions. You really don't need a sophisticated evaluation function to make a reaso...
Wed Jul 22, 2009 2:56 am
Forum: Computer Chess Club: General Topics
Topic: Diminishing returns of increasing search depth
Replies: 43
Views: 7641

Re: Diminishing returns of increasing search depth

Throw out the low data points and a linear fit doesn't look too bad. Whether the curve is linear, logarithmic, or something else, obviously there will be a point of diminishing returns eventually, because the game tree will bottom out. I just wonder how deep you have to go before it becomes impossib...
Tue Jul 21, 2009 6:05 pm
Forum: Computer Chess Club: General Topics
Topic: Diminishing returns of increasing search depth
Replies: 43
Views: 7641

Re: Diminishing returns of increasing search depth

There's enough noise in the data that I would be concerned about overfitting by adding more parameters. For example, I'm not sure what phenomenon accounts for this part of the data: 4/3 224 5/4 152 6/5 224 but I wouldn't want to model it without understanding what's going on. In any case, I think ul...
Tue Jul 21, 2009 5:30 am
Forum: Computer Chess Club: General Topics
Topic: Diminishing returns of increasing search depth
Replies: 43
Views: 7641

Re: Diminishing returns of increasing search depth

I calculated the aggregate elo at each depth, arbitrarily setting the strength at 1 ply to 0. The result certainly looks logarithmic to me, although this is hardly a rigorous test. Fit Curve, A*log(B*x+D)+C (A=3846.48, B=0.349316, C=-603.192, D=1.11704) rms = 26.8582, chi-square = 2.81317, sigma = 5...
Fri Jul 10, 2009 10:28 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Test suites for specific engine features
Replies: 5
Views: 1103

Re: Test suites for specific engine features

Dann, thanks for the link to such a massive collection of epds. I'm sure I'll find tons of useful stuff in there once I start sorting through it. The themed tests might come in handy once I have a more sophisticated eval as well. Matthew, I found your blog quite interesting. I think once my engine i...
Fri Jul 10, 2009 7:16 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Test suites for specific engine features
Replies: 5
Views: 1103

Test suites for specific engine features

I'm currently in the process of writing a new chess engine, and in an attempt to make things go as smoothly as possible, I'm trying to test each feature I add as thoroughly as possible before moving on. In some areas, this is quite simple. For move generation, for example, perft testing easily uncov...