In my chess engine I have some language/platform limitations. I've been using a TT table in the past but it didn't work very well, now i want to re-implement it ...
Given that I can only have 32bit hashes, what would be the best TT size? I assume it depends on the search depth as well, which is in ...
Search found 5 matches
- Thu Nov 11, 2010 9:16 am
- Forum: Programming and Technical Discussions
- Topic: What would be the optimal TT size (given some constraints?)
- Replies: 0
- Views: 1652
- Thu Nov 11, 2010 9:10 am
- Forum: Programming and Technical Discussions
- Topic: Move ordering improvements
- Replies: 13
- Views: 11388
Re: Move ordering improvements
I was wondering about the History Heuristics as well, so rather than opening a new topic, I'd like to ask a few questions here:
- HH seems a more general case of Killer Moves, wouldn't they be preferred over KM?
- at what ply depth do they start to become irrelevant? 6? 8?
- should they be ...
- HH seems a more general case of Killer Moves, wouldn't they be preferred over KM?
- at what ply depth do they start to become irrelevant? 6? 8?
- should they be ...
- Mon Nov 08, 2010 3:12 pm
- Forum: Programming and Technical Discussions
- Topic: Iterative negamax?
- Replies: 5
- Views: 3500
Re: Iterative negamax?
Thanks a lot for the code, I'm "decoding" now 

- Sun Nov 07, 2010 6:53 pm
- Forum: Programming and Technical Discussions
- Topic: Iterative negamax?
- Replies: 5
- Views: 3500
Re: Iterative negamax?
Just to be clearer, I'm looking for iterative alpha-beta pruning. I can use conditionals and loops, but no gotos.
- Sat Nov 06, 2010 9:53 am
- Forum: Programming and Technical Discussions
- Topic: Iterative negamax?
- Replies: 5
- Views: 3500
Iterative negamax?
I want to change the usual recursive negamax search into an iterative one. I am implementing chess into a rather primitive environment (if not downright hostile) and the iterative approach would provide benefits.
I've been looking but I couldn't find one semi-legible working code. I'd love to see ...
I've been looking but I couldn't find one semi-legible working code. I'd love to see ...