Search found 211 matches

by asanjuan
Tue Nov 25, 2014 10:07 am
Forum: Computer Chess Club: Tournaments and Matches
Topic: Is someone testing Rhetoric 1.4.1 for the CCRL 40/4 list?
Replies: 10
Views: 1622

Is someone testing Rhetoric 1.4.1 for the CCRL 40/4 list?

Last version of Rhetoric was kindly updated in the CCRL 40/40 (currently is running in a tournament), but it's still missing in the 40/4 list. I used to think that the blitz list was updated faster because it is easier to produce more games in less time. I wonder if someone has put Rhetoric in his t...
by asanjuan
Mon Nov 03, 2014 9:05 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Solving a fail low situation at the root
Replies: 47
Views: 9282

Re: Solving a fail low situation at the root

Yes, the current version returns the move from the previous iteration, but now I know that at some points, that move can be crap.
looks like we will face this problem always, as it is due to the horizont effect at the previous iteration.
by asanjuan
Mon Nov 03, 2014 7:33 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Solving a fail low situation at the root
Replies: 47
Views: 9282

Re: Solving a fail low situation at the root

Well. I know that changing only alpha or beta independently will be more efficient. That is clear. My question was related to time management regardless of the window size. I will change the question: The problem is that if in a re-search (being more or less efficient doesn't matter in this case), t...
by asanjuan
Mon Nov 03, 2014 4:51 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Solving a fail low situation at the root
Replies: 47
Views: 9282

Solving a fail low situation at the root

Hi there. I was debugging some strange moves made by the engine, and I came up debugging the very popular "fail low at the root" situation. This is, in my aspiration code, the root search returns a value <= alpha, so none of the moves can improve alpha, and then the search window is widened. Then, i...
by asanjuan
Tue Oct 28, 2014 3:05 pm
Forum: Computer Chess Club: Tournaments and Matches
Topic: Rhetoric 1.4.1 64-bit Gauntlet for CCRL 40/40
Replies: 4
Views: 758

Re: Rhetoric 1.4.1 64-bit Gauntlet for CCRL 40/40

Wow, thank you very much.

I always wanted to make something different and interesting (this is why the "!?" notation is in the logo).
by asanjuan
Tue Oct 28, 2014 2:45 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Super Fast 'Looking for magics' 1.0
Replies: 18
Views: 5082

Re: Super Fast 'Looking for magics' 1.0

It can be a good exercise, but you only need to find the magic numbers once in your life... then copy paste the magic numbers as constant in your source code and you are done. Spending hours or days trying to save 2 seconds doesn't make sense at all... Just my opinion. That's true but add it that i...
by asanjuan
Tue Oct 28, 2014 10:59 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Super Fast 'Looking for magics' 1.0
Replies: 18
Views: 5082

Re: Super Fast 'Looking for magics' 1.0

It can be a good exercise, but you only need to find the magic numbers once in your life... then copy paste the magic numbers as constant in your source code and you are done.

Spending hours or days trying to save 2 seconds doesn't make sense at all...
Just my opinion.
by asanjuan
Tue Oct 28, 2014 10:45 am
Forum: Computer Chess Club: Tournaments and Matches
Topic: Rhetoric 1.4.1 64-bit Gauntlet for CCRL 40/40
Replies: 4
Views: 758

Re: Rhetoric 1.4.1 64-bit Gauntlet for CCRL 40/40

Thanks Graham.
Rhetoric has done a good job!
by asanjuan
Wed Oct 22, 2014 6:41 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A little improvement to the Texel's tuning method
Replies: 9
Views: 3075

Re: A little improvement to the Texel's tuning method

To use games played by top programs is useless for this learning method IMO. You need suboptimal, but playable moves to get into positions where you can learn from the mistake. This is how your program can fall into the triple pawn and learn a good weight from it. Imagine that you introduce a new ev...