Mussaurus 0.1-prealpha - a poor man's Vice

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
tmokonen
Posts: 1037
Joined: Sun Mar 12, 2006 5:46 pm
Location: Vancouver

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by tmokonen » Fri Nov 22, 2019 8:40 pm

One thing I noticed when looking at your code is that the move ordering doesn't take pawn promotions into consideration. The game you posted is perfectly reasonable given what you have in your program so far, but your program did get caught by not noticing the dangerous passed pawn until it was too late. Adding promotions to your move ordering would help in situations like this, especially if passed pawn knowledge is added to the evaluation as well. I didn't notice any outright bugs in the search.

shinkarom
Posts: 19
Joined: Tue Nov 19, 2019 12:26 pm
Full name: Roman Shynkarenko

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by shinkarom » Sat Nov 23, 2019 8:32 am

Thank you for checking the code. I noticed one - the null move pruning doesn't take zugzwang into account, but until I find a satisfying condition for zugzwang it will be buggy.

Terje
Posts: 20
Joined: Tue Nov 19, 2019 3:34 am
Location: https://github.com/TerjeKir/weiss
Full name: Terje Kirstihagen

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by Terje » Sat Nov 23, 2019 3:27 pm

shinkarom wrote:
Sat Nov 23, 2019 8:32 am
Thank you for checking the code. I noticed one - the null move pruning doesn't take zugzwang into account, but until I find a satisfying condition for zugzwang it will be buggy.
If you followed VICE exactly you'll have the condition that the side to move has to have a "big piece" aka non-pawn, non-king to perform a null move, however VICE mistakenly counts the king as a big piece meaning it will always allow null moves. Fixing this, not counting the king, in my engine (also based on VICE) allowed it to solve a lot of zugzwang positions it previously could not.

shinkarom
Posts: 19
Joined: Tue Nov 19, 2019 12:26 pm
Full name: Roman Shynkarenko

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by shinkarom » Sat Nov 23, 2019 5:59 pm

That needs to have, like Vice, incrementally updated piece counter, which is work in progress.

Richard Allbert
Posts: 770
Joined: Wed Jul 19, 2006 7:58 am

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by Richard Allbert » Tue Nov 26, 2019 12:47 pm

Terje wrote:
Sat Nov 23, 2019 3:27 pm
shinkarom wrote:
Sat Nov 23, 2019 8:32 am
Thank you for checking the code. I noticed one - the null move pruning doesn't take zugzwang into account, but until I find a satisfying condition for zugzwang it will be buggy.
If you followed VICE exactly you'll have the condition that the side to move has to have a "big piece" aka non-pawn, non-king to perform a null move, however VICE mistakenly counts the king as a big piece meaning it will always allow null moves. Fixing this, not counting the king, in my engine (also based on VICE) allowed it to solve a lot of zugzwang positions it previously could not.
Yep, this was a stupid mistake, thanks for pointing it out.

User avatar
Guenther
Posts: 3108
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by Guenther » Tue Nov 26, 2019 2:17 pm

Richard Allbert wrote:
Tue Nov 26, 2019 12:47 pm
Terje wrote:
Sat Nov 23, 2019 3:27 pm
shinkarom wrote:
Sat Nov 23, 2019 8:32 am
Thank you for checking the code. I noticed one - the null move pruning doesn't take zugzwang into account, but until I find a satisfying condition for zugzwang it will be buggy.
If you followed VICE exactly you'll have the condition that the side to move has to have a "big piece" aka non-pawn, non-king to perform a null move, however VICE mistakenly counts the king as a big piece meaning it will always allow null moves. Fixing this, not counting the king, in my engine (also based on VICE) allowed it to solve a lot of zugzwang positions it previously could not.
Yep, this was a stupid mistake, thanks for pointing it out.
Richard, did you just reveal yourself as the (unknown) author of Vice, or am I just confused? :)
Current foe list count : [101]
http://rwbc-chess.de/chronology.htm

shinkarom
Posts: 19
Joined: Tue Nov 19, 2019 12:26 pm
Full name: Roman Shynkarenko

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by shinkarom » Wed Nov 27, 2019 3:29 pm

I only know that Bluefever's Arena identifies him as RICHARD2.

User avatar
Guenther
Posts: 3108
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by Guenther » Wed Nov 27, 2019 3:54 pm

shinkarom wrote:
Wed Nov 27, 2019 3:29 pm
I only know that Bluefever's Arena identifies him as RICHARD2.
viewtopic.php?f=2&t=54084&p=593625#p593625

Ouch, with that link above it means it never was a hard riddle.
At least I was in hiatus when the above quote happened and no one noticed ;-)

Here is the wayback link to Richards Lime and Jabba (chess engines) - we had lot of fun in those times,
when I was still broadcasting.

https://web.archive.org/web/20160125173 ... /index.php
Current foe list count : [101]
http://rwbc-chess.de/chronology.htm

shinkarom
Posts: 19
Joined: Tue Nov 19, 2019 12:26 pm
Full name: Roman Shynkarenko

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by shinkarom » Sat Nov 30, 2019 9:50 am

I'm deleting Mussaurus.
With my current chess skill it seems dishonest to create a chess engine; moreover, it seems not right to make such a long term project for which I don't have the stamina, and which will be drowned in bugs.

User avatar
Ovyron
Posts: 2823
Joined: Tue Jul 03, 2007 2:30 am

Re: Mussaurus 0.1-prealpha - a poor man's Vice

Post by Ovyron » Sun Dec 01, 2019 12:46 am

So you didn't enjoy working on the project?

Post Reply