Search found 299 matches

by Fabien Letouzey
Thu Jul 09, 2015 20:08
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Why? Is your definition of "quiescent" that no exchanges are possible? I would expect that to give very large QS trees, especially as you apply this recursively. It's probably only a vocabulary issue but I don't think in terms of what is quiescent or not. Instead I focus on what I can do to improve...
by Fabien Letouzey
Thu Jul 09, 2015 17:24
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Anyway i will have another quick look, if i can make it ( a little) faster. I have a feeling that a branchless version would be significantly "less slow", perhaps "only" 10-20% slower. Imagine that W and B are 0/1 booleans indicating a white (resp. black) piece on a given square. Then B - W + 1 is ...
by Fabien Letouzey
Thu Jul 09, 2015 16:30
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

And a inline definition for bdtrit Of course, bdtrit is rigorously equivalent to my trit array. It was actually on my Todo list for a while. But I assumed it would be too slow, given the large number of times I need it. Do you have any hard figure? If not fast enough, it's also possible to make it ...
by Fabien Letouzey
Thu Jul 09, 2015 15:51
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Rein Halbersma wrote:So at depth=0, what do you do here? Return eval() or try 27-21 and 27-22?
I return eval() if it fails high, otherwise try the two moves.
by Fabien Letouzey
Thu Jul 09, 2015 14:50
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Why the exclamation mark? Because it was first tried in the Chinook checkers program? I don't recall many other programs using ETC successfully: did you use this in your other draughts/chess/Othello programs? Interesting. For some reason I put ETC into the draughts-only category, perhaps because of...
by Fabien Letouzey
Thu Jul 09, 2015 12:53
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Catherine wrote:Please, can you post for us, all the games played by Scan during this tournament, or if you have some test games against others programs. Thank.
You will find them in the Olympiad thread.
by Fabien Letouzey
Thu Jul 09, 2015 12:04
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

A remaining question. in the bd.trit[square] which you use to calculate the index for a specific region , I assume the value is 1 for empty, 0 for a white man, and 2 for a black man. In case of white and black king is the value than also 1 (as empty) ? Yes, king = empty. I didn't try king = man, as...
by Fabien Letouzey
Thu Jul 09, 2015 08:23
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Now technical details for the programmers. Search Algorithm: Safe PVS from Fruit (no TT cutoff and less pruning at PV nodes). Not important but I use it as a habit (signature algorithm). No aspiration. SMP: "modern" (recursive) form of YBWC. Designed for 4 cores; probably bad above that. One impleme...
by Fabien Letouzey
Wed Jul 08, 2015 17:11
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Re: Scan

Fabien, if you want I can integrate Scan into my Engine framework. This way it is able to communicate with the Damage GUI (Version 2015) via Guide. For all interested I will share the Damage GUI. I have no idea what that is. I guess you mean to "link" the Scan source code to some code of yours that...
by Fabien Letouzey
Wed Jul 08, 2015 08:34
Forum: Draughts, Computer, Internet
Topic: Scan
Replies: 164
Views: 86395

Scan

Hi all, In this post I want to explain Scan a little bit for everybody. A technical description for programmers will follow. I would say Scan's search is like Damage's: deep but full of holes. It's different though, and relies mostly on evaluation to make pruning decisions. Evaluation is similar in ...
by Fabien Letouzey
Tue Jul 07, 2015 17:26
Forum: Draughts, Computer, Internet
Topic: Computer Olympiad 2015 Leiden
Replies: 36
Views: 17727

Re: Computer Olympiad 2015 Leiden

Krzysztof Grzelak wrote:Fabien and on what hardware you have played the tournament.
It was a quad i7 2.5 GHz (actual clock speed with 4 threads). There was 16 GB RAM but Scan only used about 2GB.
by Fabien Letouzey
Tue Jul 07, 2015 17:14
Forum: Draughts, Computer, Internet
Topic: Computer Olympiad 2015 Leiden
Replies: 36
Views: 17727

Re: Computer Olympiad 2015 Leiden

According to Damage the TD King move 50. 39-34 against Scan was wrong, whereas 50. 41-36 is a draw. Yes, 39-34 seems a mistake. Scan expected 50-44 instead which it thought was equal. I vaguely remember discussing it with Ton during the game. Scan believed TD King made positional mistakes earlier, ...
by Fabien Letouzey
Tue Jul 07, 2015 10:15
Forum: Draughts, Computer, Internet
Topic: Computer Olympiad 2015 Leiden
Replies: 36
Views: 17727

Re: Computer Olympiad 2015 Leiden

BTW, is your engine Scan going to be open source? I'd love to learn your approach to the eval in full detail. Scan will be open source but this will only include code directly related to game playing. The development version is more like a library of modules to allow me to run "lab" experiments. Fu...
by Fabien Letouzey
Tue Jul 07, 2015 08:14
Forum: Draughts, Computer, Internet
Topic: Computer Olympiad 2015 Leiden
Replies: 36
Views: 17727

Re: Computer Olympiad 2015 Leiden

It is a good lesson for all draughts programmers out there, and it is taking us 1 step closer to near-perfect play in 10x10 draughts. Hi all, I will try to address topics separately. Maybe we can start by having a look at Scan's games. I add that Scan did not play any serious game prior to the tour...