Search found 634 matches

by Desperado
Tue Dec 01, 2015 8:28 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: An approach to precomputed move generation bitboards
Replies: 11
Views: 2547

Incremental bitboard attack computations

Hello, Michael, currently i try to implement bitboard attacks in an incremental way too. so, if we ignore the implementation details for the moment, we can focus on some more general questions in this context (imo). For simplicity let's say we talk about attacks for the current piece locations. 1. L...
by Desperado
Fri Nov 20, 2015 7:48 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Why computing K that minimizes the sigmoid func. value?...
Replies: 61
Views: 12371

Re: Why computing K that minimizes the sigmoid func. value?.

But currently i don't understand "Compute the K that minimizes E. K is never changed again by the algorithm " I can follow that i should: 1. map the score into the range 0...1 (eg 0.6) 2. compute a difference (eg for a a result score this might be 1.0 - 0.6) 3. square the result, which is finally t...
by Desperado
Fri Nov 20, 2015 7:41 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Why computing K that minimizes the sigmoid func. value?...
Replies: 61
Views: 12371

Re: Why computing K that minimizes the sigmoid func. value?.

Good morning, thanks for your replies, but i still don't understand it. I think it is absolutely clear to me that the absolute score of the mapped cp value does change with different K. But is it important ? The mapped value of 600cp will always be greater than for 100cp as long both are computed wi...
by Desperado
Thu Nov 19, 2015 10:09 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Why computing K that minimizes the sigmoid func. value?...
Replies: 61
Views: 12371

Re: Why computing K that minimizes the sigmoid func. value?.

Hello, i just read the chessprogramming article about Texels tuning algorithm and found the following part which i do not understand. Sigmoid(s)=1/(1+10^(-K/400)) K is a scaling constant. Compute the K that minimizes E. K is never changed again by the algorithm. Can someone expl...
by Desperado
Thu Nov 19, 2015 9:35 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Why computing K that minimizes the sigmoid func. value?...
Replies: 61
Views: 12371

Why computing K that minimizes the sigmoid func. value?...

Hello, i just read the chessprogramming article about Texels tuning algorithm and found the following part which i do not understand. Sigmoid(s)=1/(1+10^(-K/400)) K is a scaling constant. Compute the K that minimizes E. K is never changed again by the algorithm. Can someone expla...
by Desperado
Fri Sep 11, 2015 7:15 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: lazy smp questions
Replies: 72
Views: 15910

Re: lazy smp questions

Good morning, just an idea before the weekend. Although i have a ybwc like approach at hand i will definitely spent some time on lazy smp. The main purpose and advantage i see is that it would be less error-prone with better maintenance characteristics and finally a progress compared to single cpu u...
by Desperado
Thu Sep 10, 2015 7:41 am
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli 0.7.2
Replies: 43
Views: 6028

Re: cutechess-cli 0.7.2

Hi, Ferdi,

thanks for your explanations.

@Ilari

So, finally here is my wish list for a future version of cutechess-cli:

* knockout tournament mode

Regards
by Desperado
Tue Sep 08, 2015 6:39 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli 0.7.2
Replies: 43
Views: 6028

Re: cutechess-cli 0.7.2

Hello, just some easy little questions about cutechess parameters. 1. What is the trust parameter used for ? trust Trust result claims from the engine without validation. By default all claims are validated. 2. Does the st parameter accept fractions of seconds, like 0.25 ? If so, should timemargin b...
by Desperado
Tue Sep 08, 2015 6:15 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Usage sprt / cutechess-cli
Replies: 29
Views: 6010

Re: Usage sprt / cutechess-cli.

At a first glimpse I'm not completely sure why the third test was stopped early (it was not stopped by SPRT). But "unexpected move" doesn't mean that something horrible happened. Usually it means that cutechess-cli asked an engine to stop playing/thinking and then the engine responded with a move. ...
by Desperado
Tue Sep 08, 2015 7:49 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Usage sprt / cutechess-cli
Replies: 29
Views: 6010

Re: Usage sprt / cutechess-cli.

At a first glimpse I'm not completely sure why the third test was stopped early (it was not stopped by SPRT). But "unexpected move" doesn't mean that something horrible happened. Usually it means that cutechess-cli asked an engine to stop playing/thinking and then the engine responded with a move. ...