Search found 2903 matches

by Evert
Mon Feb 21, 2011 11:15 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Puzzle with mate scores in TT
Replies: 41
Views: 7618

Re: Puzzle with mate scores in TT

silentshark wrote:one approach is to never store mate scores in the hash table. Is this such a bad deal?
Maybe not, they're probably rare enough that it doesn't affect playing strength too much, but to me it sounds as arbitrary as saying "I don't store captures in the transposition table".
by Evert
Mon Feb 21, 2011 9:43 pm
Forum: Computer Chess Club: General Topics
Topic: Some recent lower level engines
Replies: 23
Views: 2864

Re: Some recent lower level engines

Ok, I've narrowed it down a bit, but it seems to be a problem caused by an out-of-bounds array access, meaning it could be anywhere. Running a (longer TC) match through valgrind now, hopefully that will shed some light on things. Next on the list after that is rolling back to an earlier revision. O...
by Evert
Mon Feb 21, 2011 3:10 am
Forum: Computer Chess Club: General Topics
Topic: Some recent lower level engines
Replies: 23
Views: 2864

Re: Some recent lower level engines

What can't wait though is fixing a bug I seem to have introduced recently where it will sometimes play (or falsely claim) illegal moves. And by play illegal moves, I mean things like Ka6-f6. Fails to reproduce when I load the position at the root, of course. When I get that fixed, I'll upload it an...
by Evert
Sun Feb 20, 2011 10:38 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: The "Secret" TT-move Singular Extension
Replies: 27
Views: 7529

Re: The "Secret" TT-move Singular Extension

Ralph Stoesser wrote:Here's an optimization idea. Whenever you extend an alleged forced move from TT, store back in the TT at the end of search() whether this move turned out to be the best move or not.
Not sure how that would work: if it isn't the best move, you're hardly going to store it in the TT again...
by Evert
Sun Feb 20, 2011 1:50 pm
Forum: Computer Chess Club: General Topics
Topic: Some recent lower level engines
Replies: 23
Views: 2864

Re: Some recent lower level engines

Hi Evert! Great news, I've been looking forward to your release of Jazz for a long time now. :-) Allegro regards, Kirill I've been working on it on and off, but it's improved nicely along the way. I've been close to doing a release several times, but put it off each time. Usually because of somethi...
by Evert
Sun Feb 20, 2011 12:41 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Puzzle with mate scores in TT
Replies: 41
Views: 7618

Re: Puzzle with mate scores in TT

* lookup the TT. At PV nodes only use it for move ordering, but don't return the TT score (because of draw by 3/50 move rules or mate situations). I've seen that claim before and I've never understood why it would be necessary to ignore cut-offs from the transposition table in PV nodes (other than ...
by Evert
Sat Feb 19, 2011 12:39 am
Forum: Computer Chess Club: General Topics
Topic: Some recent lower level engines
Replies: 23
Views: 2864

Re: Some recent lower level engines

With my Amateur Series, I'm always looking at new engines to see whether any are strong enough to be included. I'm working on releasing both of my engines over the weekend, Jazz and Sjaak (I may need to tweak the latter's name slightly). Jazz is considerably stronger than Sjaak, which is a more gen...
by Evert
Fri Feb 18, 2011 12:59 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: LVA MVV with relative Pin
Replies: 8
Views: 1982

Re: LVA MVV with relative Pin

i assume you didnt mean it like you wrote it :) ?! if not you should fix it in your source. only return immediatelly if staticValue >= beta, look at Robert Purves code snippet above. (what you wrote is only valid for alpha == beta-1, so called nullWindows) regards Interesting. I didn't look at code...
by Evert
Thu Feb 17, 2011 11:11 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: LVA MVV with relative Pin
Replies: 8
Views: 1982

Re: LVA MVV with relative Pin

Yes, that seems like a good idea! Atm I keep track of 'historically bad' (but only 'quiet') moves [in addition to the standard history counters], and when doing move ordering it seems to give me better move ordering when I compare the move's success rate with its blunder rate (before deciding wheth...
by Evert
Wed Feb 16, 2011 9:33 pm
Forum: Computer Chess Club: General Topics
Topic: Bitboard
Replies: 5
Views: 828

Re: Bitboard

I don't think it's difficult at all.
That said, if you already have a program based on different data structures, it may be a lot of work and not obviously worth it.