Stockfish 1.5

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

zamar
Posts: 613
Joined: Sun Jan 18, 2009 7:03 am

Re: Stockfish 1.5

Post by zamar »

[D]2r1r1k1/1pq2pp1/p2p2b1/P2P4/2P5/3p1P1P/P2Q1BP1/2R1R1K1 w - - 0 26
Uri Blass wrote: The following position happened in one of my games and I believe it is probably a draw so I would like to see computers evaluate it closer to draw inspite of the fact that white is a pawn up in case that I use a special personality.

The game is already at move 39 of white and I do not ask for advice in a specific game but only for personality that is going to be against opposite color bishops if I try to win.

Uri
Stockfish doesn't apply any scaling to position you posted. The problem of evaluating the opposite colored bishops is that in endgame they certainly reduce winning chances, but in midgame positions opposite colored bishops often increase attacking chances.

In position you posted white could speculate on using a1-h8 diagonal for attacking with bishop + queen and at the same time speculate playing c5 for creating passed pawn. If he had time to accomplish this he would likely be winning. Deep search will probably reveal that in this specific position this plan won't work (because of open e-file and strong black passer). But it's impossible to identify this by simple static evaluation() call. So I'm afraid that although drawish heuristic you suggest might improve play in this specific position, it could do harm in others very similar positions.
Joona Kiiski
Alexander Schmidt
Posts: 1203
Joined: Thu May 10, 2007 2:49 pm

Re: Stockfish 1.5

Post by Alexander Schmidt »

Hmmm, it crashes here in Chess960, just in games where kings and rooks are on their usual position it works..
Tord Romstad
Posts: 1808
Joined: Wed Mar 08, 2006 9:19 pm
Location: Oslo, Norway

Re: Stockfish 1.5

Post by Tord Romstad »

Alexander Schmidt wrote:Hmmm, it crashes here in Chess960, just in games where kings and rooks are on their usual position it works..
You are right.

This is an extremely stupid and embarrassing bug. Chess960 did work some time back, but apparently we broke it later. I've fixed the bug now, but will run some tests to make sure there are no other bugs before we release a new version.

Thanks for the bug report.

:oops:
Alexander Schmidt
Posts: 1203
Joined: Thu May 10, 2007 2:49 pm

Re: Stockfish 1.5

Post by Alexander Schmidt »

Thanks a lot :)

I really can't wait to see StockFish playing FRC :D
rainhaus
Posts: 185
Joined: Sun Feb 01, 2009 7:26 pm
Location: Germany
Full name: Rainer Neuhäusler

Re: Stockfish 1.5

Post by rainhaus »

Rush hour of the chess engines! Fritz 12, Shredder 12, Stockfish 1.5. The Stockfish Triad wrote a new source, Jim Ablett compiles at lightning speed and lo and behold, a new engine was born, well bedded into the world wide standard of SMKs UCI and into a smart GUI of your choice. In some way I feel almost guilty, because I've got a very nice present from hard working people putted into my empty hands. Thank you very much. Here fits Franz Beckenbauer's (a German socker legend) slogan: is it already Christmas?
Cheerio Rainer
mcostalba
Posts: 2684
Joined: Sat Jun 14, 2008 9:17 pm

Re: Stockfish 1.5

Post by mcostalba »

Tord Romstad wrote:
Alexander Schmidt wrote:Hmmm, it crashes here in Chess960, just in games where kings and rooks are on their usual position it works..
You are right.

This is an extremely stupid and embarrassing bug. Chess960 did work some time back, but apparently we broke it later. I've fixed the bug now, but will run some tests to make sure there are no other bugs before we release a new version.

Thanks for the bug report.

:oops:
I made the buggie change :oops: , but is not very new, probably few months ago (I am not at home and I cannot browse the revision's history now), perhaps even earlier then 1.4 that didn't work for other reasons.

I agree we really need to release a manteinance release when 960 works for sure of course.

Alexander are you able to compile sources yourself ? I could send you tomorrow the new sources with the proposed fix and so you can test and confirm bug is fixed for you. In this case is better a test more (from the bug reporter) then less so we are really sure we don't have other surprises...
Tord Romstad
Posts: 1808
Joined: Wed Mar 08, 2006 9:19 pm
Location: Oslo, Norway

Re: Stockfish 1.5

Post by Tord Romstad »

mcostalba wrote:
Tord Romstad wrote:
Alexander Schmidt wrote:Hmmm, it crashes here in Chess960, just in games where kings and rooks are on their usual position it works..
You are right.

This is an extremely stupid and embarrassing bug. Chess960 did work some time back, but apparently we broke it later. I've fixed the bug now, but will run some tests to make sure there are no other bugs before we release a new version.

Thanks for the bug report.

:oops:
I made the buggie change :oops: , but is not very new, probably few months ago (I am not at home and I cannot browse the revision's history now), perhaps even earlier then 1.4 that didn't work for other reasons.
No, it's later than 1.4. Stockfish 1.4, where Chess960 didn't work, was released on July 4. Chess960 support was fixed on July 10. The new bug was introduced on July 23.

Clearly, we should run Chess960 test games a little more often.
;)
jpqy
Posts: 550
Joined: Thu Apr 24, 2008 9:31 am
Location: Belgium

Re: Stockfish 1.5

Post by jpqy »

Here a great win against Bright..very nice to follow these games!!

Blitz 5min. Core i7 920 @3.87Ghz 2009 4cores HT OFF

1 Stockfish_15_ja 2800 ½11111½010 7.0/10
2 Bright-0.4a 2800 ½00000½101 3.0/10

Results i already get:
Blitz 5min. Core i7 920 @3.87Ghz 2009 4cores HT OFF

Stockfish_15_ja 2800 - Cyclone xTreme Fear 2800 7.5 - 2.5 +5/-0/=5 75.00%
Stockfish_15_ja 2800 - MP-x86-Inert---Thinker 5.4C 2800 6.0 - 4.0 +4/-2/=4 60.00%
Stockfish_15_ja 2800 - Bright-0.4a 2800 7.0 - 3.0 +6/-2/=2 70.00%

Great improvement,Stockfish 1.5 plays very strong!!

JP.
Alexander Schmidt
Posts: 1203
Joined: Thu May 10, 2007 2:49 pm

Re: Stockfish 1.5

Post by Alexander Schmidt »

mcostalba wrote:Alexander are you able to compile sources yourself ?
No, I am sorry. But if you can to send an executeable for testing I would be glad if I can help. I will pm my email adress.
User avatar
Leto
Posts: 2071
Joined: Thu May 04, 2006 3:40 am
Location: Dune

Re: Stockfish 1.5

Post by Leto »

Just a quick test but at 2 minutes blitz Stockfish 1.5 x64 4CPU beat Naum 4 4CPU in a 10 game match with a score of 6.5-3.5, impressive.