Search found 155 matches

by micron
Tue Jan 10, 2012 5:33 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Crafty logfile
Replies: 6
Views: 736

Re: Crafty logfile

Or send
log off
as the first command after Crafty is started.
by micron
Mon Jan 09, 2012 11:53 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Depth vs playing strength
Replies: 6
Views: 1216

Re: Depth vs playing strength

Engines normally run under time control. The depth reached is variable, depending on the position, search speed and time allocated for a move. Fixed depth has limited uses, mainly certain kinds of testing during development. FWIW, I ran a match with my engine (fixed depth 10 vs fixed depth 8), and f...
by micron
Mon Jan 09, 2012 9:51 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Iteration in pondering
Replies: 38
Views: 3664

Re: Iteration in pondering

Shouldn't really matter, since the search will burn through the lower plies very quickly because of deep entries stored in the transposition table. This is not necessarily true. In my engine, a second search (of the same position) is not significantly faster than the first. I tracked this strange d...
by micron
Thu Jan 05, 2012 5:20 am
Forum: Computer Chess Club: General Topics
Topic: Prédateur 2.2 Released
Replies: 26
Views: 3169

Re: Prédateur 2.2 Released

cutechess-cli 0.5.0 does not like Prédateur:

Invalid UCI option from "Predateur v2.2_MacIntel" : "option name UCI_EngineAbout type String default http://predateur-chess.blogspot.com"
by micron
Tue Jan 03, 2012 10:26 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SEE is too slow
Replies: 34
Views: 5140

Re: SEE is too slow

A bug in SEE could only lead to pruning too many captures or too few captures, but not to pruning "less than nothing", where "nothing" is the claimed behaviour on the "without" side. Hi Sven The paradox can be explained. Suppose that the first move in QS is truly the best, and is the only move that...
by micron
Tue Jan 03, 2012 8:56 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SEE is too slow
Replies: 34
Views: 5140

Re: SEE is too slow

But the total nodes searched is actually 2% more using SEE, than without The counter-intuitive result (that pruning increases node count instead of decreasing it markedly) suggests a bug in SEE. Here are a couple of suggestions to test against your code. [1] The order in which lowest-value attacker...
by micron
Mon Jan 02, 2012 10:01 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SEE is too slow
Replies: 34
Views: 5140

Re: SEE is too slow

Perhaps my MVVLVA is wrong, I value moves by 16*victim - attacker
where the values are K=0, P=1,N=3,B=4,R=5,Q=6
This seems ok though
No. K should be the most valuable.
by micron
Mon Jan 02, 2012 4:48 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SEE is too slow
Replies: 34
Views: 5140

Re: SEE is too slow

Those figures look reasonable. The non check-evasion part of QS should now resemble this (changes from your original pseudocode indicated ////): QS( alpha, beta ) { int val = getEval(); if ( val >= beta ) return beta; ///// if ( val > alpha ) alpha = val; ///// moves = getCaptureMoves&#...
by micron
Sat Dec 31, 2011 3:57 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SEE is too slow
Replies: 34
Views: 5140

Re: SEE is too slow

MVVLVA: score_for_sort[j] = ValOf( capturedPiece ) - (ValOf( capturingPiece ) >> 3); This works out to be as effective, for ordering 'good' captures, as score_for_sort[j] = SEE( move ); Just use SEE() for classification of doubtful moves, so you can pick up this optimization an e...
by micron
Sat Dec 31, 2011 1:27 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SEE is too slow
Replies: 34
Views: 5140

Re: SEE is too slow

1. If your SEE() is the complicated, accurate and slow kind that reveals new slider attacks during the exchange, that's overkill for move ordering. Remove the Xray attacks code for better speed. An Xray-enabled SEE is better reserved for pruning decisions in QS, where you need all the accuracy you c...