I'm currently working on rewriting the Board-class. It was a complete mess but is starting to shape up. The speed gain for the new Board-class is currently at a grand total of 0, but atleast the code is much better looking. :)
Hopefully the gains will come further down the chain, e.g. in evaluation.
Follow the process of creating a chess engine
4 comments:
What further serious improvements are you planning to add to your engine for increasing playing strength?
My plan is to first rewrite the messy part of the code (Board and Transposition tables) and see what gains I can get from that. Hopefully (probably) quite a bit.
When that is done I will look at the search. I'm pretty sure particulary the history heuristic and killer moves have some serious issues to improve on.
And finally I'm going to take a deep look at the evaluation, making it more customizeable so I can try out different setups.
As for completely new features I'll be looking at new ways to sort moves (particulary root moves). Also futility pruning is still missing (I could never get it to work). And perhaps some extensions are worth a look.
So a lot of things are planned, hopefully it will help quite a bit when it comes to Mediocre's strength.
Hi Jonatan - I'm glad to see Mediocre up and going again. Just for the record (and a ever-so-kindly heads up ;-)), Vicki is undergoing a complete rewrite at the moment. Mediocre will be up for a serious challenge soon! I just can't be seen in public, knowing that a Java-engine is crushing my C-engine! Good luck... (you'll need it... he-he.)
Sounds good. :) Mediocre is undergoing some serious changes as well so hopefully it should prove a bigger challenge than you think. :)
Post a Comment