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.
Gabor Szots
Posts: 301
Joined: Sat Jul 21, 2018 5:43 am
Location: Szentendre, Hungary
Full name: Gabor Szots

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

Post by Gabor Szots » Thu Nov 21, 2019 8:06 pm

Well, after the bug fix (released a couple of minutes ago) it did not hang but lost a piece in about 10 moves then left its queen en prise, so...
Gabor Szots
CCRL testing group

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 » Fri Nov 22, 2019 8:24 am

So what? Didn't I say it loses to Nero while having all the same features that Vice had when it trounced Nero?

Gabor Szots
Posts: 301
Joined: Sat Jul 21, 2018 5:43 am
Location: Szentendre, Hungary
Full name: Gabor Szots

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

Post by Gabor Szots » Fri Nov 22, 2019 8:30 am

shinkarom wrote:
Fri Nov 22, 2019 8:24 am
So what?
To put it frankly, if an engine blunders a queen in every game, I am not going to test it. Other testers may have less severe restrictions so there is hope from both sides.
Gabor Szots
CCRL testing group

Gabor Szots
Posts: 301
Joined: Sat Jul 21, 2018 5:43 am
Location: Szentendre, Hungary
Full name: Gabor Szots

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

Post by Gabor Szots » Fri Nov 22, 2019 8:31 am

shinkarom wrote:
Fri Nov 22, 2019 8:24 am
So what? Didn't I say it loses to Nero while having all the same features that Vice had when it trounced Nero?
I took it literally when you said it did not dominate Nero. I wanted to test it against that range of engines.
Gabor Szots
CCRL testing group

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 » Fri Nov 22, 2019 9:21 am

If by "testing" you mean CCRL, then it would be insanity to ask for that now (or maybe ever). And that was not yet the intention of my post. If you mean just playing a couple of games, then it still needs a lot of work.
After removing HashEntryFlag (again) so that it hashes only moves that are between alpha and beta; and removing returning the hash score altogether (so that it uses the table only for move ordering), here is a sample game against Nero 6.1 on 1+0 time controls:

Gabor Szots
Posts: 301
Joined: Sat Jul 21, 2018 5:43 am
Location: Szentendre, Hungary
Full name: Gabor Szots

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

Post by Gabor Szots » Fri Nov 22, 2019 10:06 am

shinkarom wrote:
Fri Nov 22, 2019 9:21 am
If by "testing" you mean CCRL, then it would be insanity to ask for that now (or maybe ever). And that was not yet the intention of my post. If you mean just playing a couple of games, then it still needs a lot of work.
OK then. I did mean testing for CCRL and I have not given up on Mussaurus for ever. From the game you posted I gather that it can play fairly reasonable chess. For your information, the weakest engine on the list that has games from me is Soberango (in fact I did play 32 games with AcquaD but I expected a stronger engine when I did that).
Gabor Szots
CCRL testing group

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 » Fri Nov 22, 2019 2:46 pm

I've converted the search to fail-soft, just because.
May I request that somebody give a cursory glance to https://github.com/shinkarom/mussaurus/ ... e/search.d and check it for blunders?

RubiChess
Posts: 119
Joined: Fri Mar 30, 2018 5:20 am

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

Post by RubiChess » Fri Nov 22, 2019 3:00 pm

shinkarom wrote:
Thu Nov 21, 2019 8:00 am
Nobody has anything to say to me at all?
All I can say is that the VICE videos on youtube were a great introduction to chess programming for me.
Keep on going!

Andreas

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 » Fri Nov 22, 2019 3:38 pm

Yes, Bluefever has done well, even about reading Polyglot books.

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 » Fri Nov 22, 2019 4:03 pm

shinkarom wrote:
Thu Nov 21, 2019 1:11 pm
@Ovyron Well, didn't you smile when you saw the poor and sloppy shape my code is in?
Ha! I doubt you can write more sloppy code than me! I'm not a fan of "for" loops, so I use the "goto" command everywhere. Actually, the reason I quit coding was because after a few years I'd check my old projects and the code was so unreadable that I couldn't know what it was doing anymore.

Whatever status you think Mussaurus is at, it definitively beats my only "from scratch" effort. I tried building a chess GUI, and it looked beautiful, the problem was that when the user moved pieces around (and I only implemented click on piece and click on destination, never dragging) randomly pieces would just move towards the destination square, miss it, and keep moving forever, and it would hang (because nothing else would be done until it reached it, but after missing it it'd move away from it), and I never knew why. I managed to reduce it to 1 in 100 attempts, but the next time it hung I gave up, I couldn't deal with it, because I always thought I had finally fixed it, only to see a piece flying off the board forever making all my fixes useless...

So yeah, don't compare to others, realize that if Mussaurus was the only engine around, it'd be the best one, and never surrender :)

Post Reply