Search found 825 matches

by wgarvin
Tue Jul 28, 2015 6:02 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Time assignment to children
Replies: 58
Views: 11136

Re: Time assignment to children

Depth is still of great importance as it covers up the weaknesses of the evaluation function (some of which are not remotely easily repairable). Let's take an imaginary tree with two interesting root moves, white to move. Interesting root move 1 pushes a pawn and closes the position and leads to a ...
by wgarvin
Tue Jul 28, 2015 5:56 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Time assignment to children
Replies: 58
Views: 11136

Re: Time assignment to children

I have provided justification for spending equal time on each branch (vs equal depth). What are the justifications for equal depth? 1. Scores from equal depths are more likely to be directly comparable. 2. The cost of searching to depth X+1 instead of X, usually exceeds the savings from searching t...
by wgarvin
Mon Jul 27, 2015 10:07 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Time assignment to children
Replies: 58
Views: 11136

Re: Time assignment to children

Now obviously, we don't actually do uniform-depth searches anymore. We do singular extensions, check extensions, recapture extension, one-reply extension, etc, and LMR. All of those techniques have been successfully applied in some program at some point. What do they all have in common? They all fo...
by wgarvin
Fri Jul 10, 2015 4:16 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Small bug in Stockfish passed pawn evaluation?
Replies: 10
Views: 2070

Re: Small bug in Stockfish passed pawn evaluation?

I dislike this trick too. It violates the "Principle of Least Surprise". Programmers who haven't read this bitboard class should expect & to be a bitwise-and of the two values. They should always be able to safely assume that--this trick would surprise them, and is therefore bad. :)
by wgarvin
Tue Jun 09, 2015 2:03 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Annual reminder
Replies: 38
Views: 7321

Re: Annual reminder

As far as reluctance to make a strange pledge, I can see that point too. E.g.: "I promise not to beat my wife." "I promise not to to rob a bank." "I promise not to kick a puppy." Do these things really even need to be spoken? If they do, it is something really sad. We're setting the bar awfully low...
by wgarvin
Tue Jun 09, 2015 1:22 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Annual reminder
Replies: 38
Views: 7321

Re: Annual reminder

See, this reminds me of an incident a couple of years ago, when we were living in North-America. We were following pre-natal classes and in one of them they covered the topic of shaking babies and how you should then never, ever do that. This was followed by a little ceremony where you could write ...
by wgarvin
Sat May 23, 2015 1:39 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: 500mhz Atom compute module for $50 ?
Replies: 3
Views: 750

500mhz Atom compute module for $50 ?

I noticed the other day that Adafruit sells an Intel Edison Compute Module . Its a 500mhz intel Atom with 1GB of memory in a relatively small form factor, meant for use in embedded projects that require more compute power than the usual microcontrollers can provide. It has wireless, but I'm not sure...
by wgarvin
Fri May 22, 2015 12:57 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: You can help me specify a new computer
Replies: 60
Views: 10827

Re: You can help me specify a new computer

For general purpose computing you could put together a very reasonable system for under $2000. You only need an expensive graphics card if you are into PC gaming. I would recommend something like the i7-4790k, which is 4 cores at 4GHz, and overclockable: http://ark.intel.com/products/80807/Intel-Co...
by wgarvin
Fri May 01, 2015 5:52 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Piece weights with regression analysis (in Russian)
Replies: 54
Views: 13930

Re: Piece weights with regression analysis (in Russian).

I found that it auto-translates fairly well (though elephant for bishop is a little distracting). Google Translate made me laugh out loud with this one: The main components of virtually any protein-free chess are searching and assessment of the position . But still, its extremely readable. Its amaz...
by wgarvin
Tue Mar 31, 2015 6:45 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Using Stockfish as a true random number generator, possible?
Replies: 16
Views: 3105

Re: Using Stockfish as a true random number generator, possi

What do you mean when you say "truly random"? The number of nodes searched using bench 32 2 20 default depth is a random variable . I assume it's normally distributed, and you could estimate the mean and variance. What exactly do you mean when you say "true random number generator"? I'm not sure yo...