Search found 4003 matches

by syzygy
Sun Jun 09, 2019 11:11 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 2468

Re: Hash collision?

Collisions in the sense of two positions mapping to the same entry are unavoidable. They cannot cause strang evals and moves as long as the engine detects them as different. So what is important is the ability of the engine to detect that two different positions mapping to the same entry/bucket are...
by syzygy
Sun Jun 09, 2019 11:08 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 2468

Re: Hash collision?

In Stockfish, the ratio of false hits is one in every 2^16/3 = 21845 probes, so one in every 16384 is unlikely to do a lot of damage. Maybe Stockfish doesn't probe on every node (position) it generates, I don't know. But if it does, and you run on a fairly modern PC it can have like 1000 false hits...
by syzygy
Sat Jun 08, 2019 12:51 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Hash collision?
Replies: 42
Views: 2468

Re: Hash collision?

So.. going back to my original posting, here's my data. What do people think about the probability of hash collisions causing strange evals and moves? 1. Time control 40 moves in 25 minutes 2. Searching approx 2 million NPS 3. Size of hash table, 2097152, with 4 buckets 4. Using a 64 bit int for th...
by syzygy
Sat Jun 01, 2019 7:15 pm
Forum: Computer Chess Club: General Topics
Topic: for kai laskos (about doubling speed elo gain)
Replies: 14
Views: 2032

Re: for kai laskos (about doubling speed elo gain)

Isn't it to be expected that the draw rate will be lower (and so the W/L will be higher) between the same two engines as TC increases? Both engines will make fewer mistakes and there will therefore be less noise. At STC, mistakes by both engines cancel each other out so far as they don't represent ...
by syzygy
Sat Jun 01, 2019 2:38 pm
Forum: Computer Chess Club: General Topics
Topic: for kai laskos (about doubling speed elo gain)
Replies: 14
Views: 2032

Re: for kai laskos (about doubling speed elo gain)

... which is relevant to the W/L question Kai raised (though it's only Komodo 9.3). This is not that hard. The harder thing seems to be to get the behavior of Win/Loss ratio to LTC. From the latest months' data of Andreas, I seem to get that Win/Loss ratio increases from rapid to LTC, almost consis...
by syzygy
Thu May 30, 2019 4:04 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One idea to solve chess?
Replies: 35
Views: 6582

Re: One idea to solve chess?

I think that if you want to solve chess then you should try to solve. chess engines of today do not try to solve the game and when they say 0.00 people have no idea if they found a forced draw or simply found a line that they evaluate as equal. That’s easily dealt with. Use bit zero to distinguish ...
by syzygy
Sun May 26, 2019 9:18 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One idea to solve chess?
Replies: 35
Views: 6582

Re: One idea to solve chess?

Well that’s a straw man, bizarre also because they call magic number move generator systems magic bitboards over on the chess programming wiki last time I looked. OK, so you have understood my statement that "a neural network is not a piece of magic" differently from what I was trying to say with i...
by syzygy
Sat May 25, 2019 8:40 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One idea to solve chess?
Replies: 35
Views: 6582

Re: One idea to solve chess?

A neural network is not a piece of magic. there probably is a magic neural network for 8,9,10 and so on egtbs, with the weights set to just the right magic values. There might even be an infinite number of such networks. Problem actually is that (most likely) none can be found by error back propaga...
by syzygy
Sat May 25, 2019 3:15 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One idea to solve chess?
Replies: 35
Views: 6582

Re: One idea to solve chess?

I think that if you want to solve chess then you should try to solve. chess engines of today do not try to solve the game and when they say 0.00 people have no idea if they found a forced draw or simply found a line that they evaluate as equal. That’s easily dealt with. Use bit zero to distinguish ...
by syzygy
Sat May 25, 2019 3:11 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One idea to solve chess?
Replies: 35
Views: 6582

Re: One idea to solve chess?

Let's say it was possible. I doubt that one could store the number of weights needed for that network because I dont think current networks arelarge enough to give a perfect answer to any chess position Has anyone tried to compress the data of a 3 men tablebase in a NN and get the same result? That...