Doch 1.3.1

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

Moderators: hgm, Rebel, chrisw

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 4:27 pm

Doch 1.3.1

Post by Don »

Doch 1.3.1 has been submitted to Jim Ablett as a bug fix release version.

It fixes the issue where the user interfaces sends <empty> if the personality is not filled in. Also, if the user tries to enter a file that does not exist, it will silently fail - so let this be a warning. I'm not sure how to transmit an error message through the UCI interface unless I force the program to exit, which is perhaps what should be done.

When it silently fails the current setting are used, whatever they may be.
Gian-Carlo Pascutto
Posts: 1243
Joined: Sat Dec 13, 2008 7:00 pm

Re: Doch 1.3.1

Post by Gian-Carlo Pascutto »

You can do "info string" but depending on the GUI it may not be prominently displayed.

FWIW I'm currently seeing about +30 ELO. The rate of improvement is impressive. I'm pretty happy at having reasonable Linux opponents to test against.
User avatar
F.Huber
Posts: 853
Joined: Thu Mar 09, 2006 4:50 pm
Location: Austria

Re: Doch 1.3.1

Post by F.Huber »

Don wrote:Doch 1.3.1 has been submitted to Jim Ablett as a bug fix release version.
Thanks for the quick fix, Don. :-)

I hope you're not offended about my "not a good programming style", I was just a bit annoyed by some dumb replies to my bugreport in the other thread. ;-)
I'm not sure how to transmit an error message through the UCI interface unless I force the program to exit, which is perhaps what should be
Well, this would be possible with "info string ...", but I don't think this is really necessary - if someone enters a nonexisting file then it's just his own fault. :P

Regards,
Franz
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 4:27 pm

Re: Doch 1.3.1

Post by Don »

Gian-Carlo Pascutto wrote:You can do "info string" but depending on the GUI it may not be prominently displayed.

FWIW I'm currently seeing about +30 ELO. The rate of improvement is impressive. I'm pretty happy at having reasonable Linux opponents to test against.
Yes, I should at least do info string. Is there a linux version of your program available? I am find it very difficult to find a variety of worthy opponents and even though I can time handicap it's a serious testing resource issue.
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 4:27 pm

Re: Doch 1.3.1

Post by Don »

Gian-Carlo Pascutto wrote:You can do "info string" but depending on the GUI it may not be prominently displayed.

FWIW I'm currently seeing about +30 ELO. The rate of improvement is impressive. I'm pretty happy at having reasonable Linux opponents to test against.
I wrote somewhere else that I'm getting 55 ELO on 1 minute + 1 second, not pondering based on a few thousands games but I'm a bit worried about how it will do at longer time controls. What time control are you testing with where you see the 30+?
Frank Quisinsky
Posts: 6808
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: Doch 1.3.1

Post by Frank Quisinsky »

Hi Don,

thanks for the new versions of Doch.

In the latest release 1.2 Doch improved around 30 ELO (x64). In the own ratinglist (schach-welt, 40 moves in 10 minutes, ponder = on, w32) 12 ELO after 680 games. With Rybka 3 and Zappa Mexiko II Doch 1.2 have the biggest advantage (around 55 ELO different between x64 and w32) with x64 from all of the top engines.

Do you think that booth, w32 and x64, are improved with around 40 ELO?

Best
Frank
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 4:27 pm

Re: Doch 1.3.1

Post by Don »

Frank Quisinsky wrote:Hi Don,

thanks for the new versions of Doch.

In the latest release 1.2 Doch improved around 30 ELO (x64). In the own ratinglist (schach-welt, 40 moves in 10 minutes, ponder = on, w32) 12 ELO after 680 games. With Rybka 3 and Zappa Mexiko II Doch 1.2 have the biggest advantage (around 55 ELO different between x64 and w32) with x64 from all of the top engines.

Do you think that booth, w32 and x64, are improved with around 40 ELO?

Best
Frank
I can only guess. I did some things going to version 1.2 that may have made the 1.2 less efficient on 32 bit systems but if that is the case we have already taken the hit for that. I can only speculate that it should be about the same going from 1.2 to 1.3
User avatar
Graham Banks
Posts: 41416
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Doch 1.3.1

Post by Graham Banks »

Don wrote:
Gian-Carlo Pascutto wrote:You can do "info string" but depending on the GUI it may not be prominently displayed.

FWIW I'm currently seeing about +30 ELO. The rate of improvement is impressive. I'm pretty happy at having reasonable Linux opponents to test against.
I wrote somewhere else that I'm getting 55 ELO on 1 minute + 1 second, not pondering based on a few thousands games but I'm a bit worried about how it will do at longer time controls. What time control are you testing with where you see the 30+?
Based on early results here, seems to be a nice improvement. :)

Cheers,
Graham.
gbanksnz at gmail.com
User avatar
Jim Ablett
Posts: 1383
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: Doch 1.3.1

Post by Jim Ablett »

Image
Doch 1.3.2 JA by Don Dailey

x64/win32 & Linux 64/32 bit compiles.

What's new:
Fixes personalities file loading bug.

Download:
http://www.mediafire.com/?lgojmvfucy2

Jim.
DomLeste
Posts: 221
Joined: Thu Mar 09, 2006 4:53 pm

Re: Doch 1.3.1

Post by DomLeste »

After each test suite and loading eng-eng game in Fritz GUI

Doch 1.32 32bit JA caused an exception..I can't play games with this version or in Shredder GUI. It doesnt move in eng v eng games.

I have 3 personalities to test but i can only analyse but cant play games with them.
Last edited by DomLeste on Wed Jan 06, 2010 5:39 am, edited 1 time in total.
Insanity: doing the same thing over and over again and expecting different results.
Albert Einstein