Stuck on engine development

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

oriyonay
Posts: 32
Joined: Tue Jun 01, 2021 5:46 am
Full name: ori yonay

Stuck on engine development

Post by oriyonay »

Hi there, my beautiful friends :)
I've been trying to make some progress with my chess engine, but everything I've been trying seems to hurt my engine's performance. I'd love it if a few of you experts could have a look at my code and provide any helpful tips/feedback you can give, as simply as possible :)

Also - how should I go about testing my engine? I've been running self-play matches using the following command on cutechess-cli:

Code: Select all

cutechess-cli -engine conf=lunatest2 tc=0/10+0.1 -engine conf=lunatest tc=0/10+0.1 -tournament round-robin -rounds 1000 -sprt elo0=0 elo1=5 alpha=0.05 beta=0.05 -resign movecount=6 score=1000 -draw movenumber=45 movecount=5 score=15 -concurrency 8
I've heard some things about the strategic test suite being good at estimating engines' elo but am not sure how to run it on Mac.

Any help is very appreciated!!
Have a wonderful rest of your day,
- Ori :)
gflohr
Posts: 57
Joined: Fri Jul 23, 2021 5:24 pm
Location: Elin Pelin
Full name: Guido Flohr

Re: Stuck on engine development

Post by gflohr »

Not that I consider myself an expert but maybe search for the open todos in your code. For example, it looks as if you would currently ignore the en passant square in your FEN parser. In other words, your engine does not know all legal moves.

I think that unit tests are also helpful and would probably discover obvious problems like the en passant thing in the FEN parser.
Henk
Posts: 7216
Joined: Mon May 27, 2013 10:31 am

Re: Stuck on engine development

Post by Henk »

Yes its a hopeless project. No easy way to determine if one version is a bit better than the other.
Maybe solve only end game positions with less pieces if that would be easier. Solving tactical puzzles. I don't know.

Simplifying code might generate ideas. Or do nothing and wait for a next (stupid) idea comes into your mind.
I haven't made much elo progress for 5 years or maybe it plays even worse.

Not many like (Stockfish) clones or not.
R. Tomasi
Posts: 307
Joined: Wed Sep 01, 2021 4:08 pm
Location: Germany
Full name: Roland Tomasi

Re: Stuck on engine development

Post by R. Tomasi »

My recommendation would be to only focus on one thing at a time. Only move to the next "todo" if you are really convinced that the current version of your engine does exactly behave the way you would expect and has no bugs. If you work on multiple issues simultaneously you are bound to be living in a world of pain.