Crystal 3.1

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

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Vinvin
Posts: 4911
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Crystal 3.1

Post by Vinvin » Fri Jan 15, 2021 3:24 pm

jhellis3 wrote:
Thu Jan 14, 2021 5:19 pm
Release highlights:

Not too long after 3.0 was released there was an important false mate fix upstream, which has now been merged.
...
jshriver wrote:
Thu Jan 14, 2021 10:54 pm
I built it in various formats for Windows and put on my CDN. Hope it works well for you.

https://chess.nyc3.cdn.digitaloceanspac ... .1.tar.bz2
Thanks jhellis3 and jshriver !!

Vinvin
Posts: 4911
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Crystal 3.1

Post by Vinvin » Sat Jan 16, 2021 11:42 pm

New record in solving the Hard-Talkchess-2020 set : http://talkchess.com/forum3/viewtopic.p ... 04#p879604

Code: Select all

SugaR.AI.ICCF.1.00 : 166 solved
ShashChess15.1     : 166 solved
Bluefish-12-W      : 162 solved
Stockfish_20121407 : 153 solved
Eman 6.61          : 153 solved
SugaR.AI.1.00      : 149 solved
Crystal-291220     : 169 solved
Crystal V3.1       : 173 solved

Paloma
Posts: 1007
Joined: Thu Dec 25, 2008 8:07 pm

Re: Crystal 3.1

Post by Paloma » Sun Jan 17, 2021 3:34 am

Why is Bluefish-FD-XI with 181/213 not included in the calculation?

Dann Corbit
Posts: 11989
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Crystal 3.1

Post by Dann Corbit » Mon Jan 18, 2021 3:02 am

Crystal 3.1 takes no prisoners (doing some O. Blathy hard ones):
OK, total miss. Right move, but where is the checkmate (Sting found it in 5 minutes):
8/p1p1P3/p7/5p2/p3n3/2n2pp1/4prrk/4Kbbq w - - acd 49; ce 0; acs 900; acn 81648000; pv e8=Q Ng5 Qh8+ Nh3 Qxc3 a3 Qxc7 Ng5 Qe5 Kh3 Qxf5+ Kh4 Qe4+ Kh5 Qe8+ Kh4 ; bm e8=Q;
BAM:
1kb4K/1p1p4/1PpPp3/4Pp2/p1P1pP2/2N1Pp2/P4P2/2b5 w - - acd 38; ce 14387; acs 900; acn 108052000; pv Nb1 ; bm Nb1;
BAM:
2b4b/Pr1p3P/qP1p2Q1/1R1n4/8/n6r/7P/k6K w - - acd 45; ce 9759; acs 900; acn 83153000; pv Qg1+ Ka2 ; bm Qg1+;
Bam:
1N1Brkbn/5p2/6pr/4P1Pp/4P3/7K/8/7N w - - acd 38; ce 9847; acs 900; acn 54497000; pv Nd7+ Kg7 Bf6+ Kh7 Kh4 Rc8 Nf2 Ra8 Nd3 Re8 Ne1 Rc8 Nc2 Re8 Nd4 Rb8 Nf3 Re8 Nd2 Rb8 Nc4 Re8 Ncb6 Rd8 Bxd8 f5 Nf6+ Kg7 gxh6+ Kf7 Nbd7 Ke6 exf5+ gxf5 Nc5+ Kxe5 Nxg8 Kd6 Bf6 Ng6+ Kxh5 Nf4+ Kg5 Nd5 Kxf5 Ne3+ Kg5 Kc6 Ne7+ Kb5 h7 Nf1 h8=Q ; bm Nd7+;
Bam:
7K/1P1p1kpP/n1pPp3/1pP1Pp2/5Pp1/1P4P1/8/6N1 w - - acd 39; ce 9827; acs 900; acn 83763000; pv Ne2 ; bm Ne2;
Cannot wait to see how the rest turn out.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Dann Corbit
Posts: 11989
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Crystal 3.1

Post by Dann Corbit » Mon Jan 18, 2021 3:07 am

Paloma wrote:
Sun Jan 17, 2021 3:34 am
Why is Bluefish-FD-XI with 181/213 not included in the calculation?
I guess he looked at the thread count and saw something nefarious.
If I ask for 1 thread and 4 threads are pounding like mad, then something is up.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Jouni
Posts: 2386
Joined: Wed Mar 08, 2006 7:15 pm

Re: Crystal 3.1

Post by Jouni » Mon Jan 18, 2021 12:59 pm

Nice progress with 3.1 and default net! What is "Clean Search" meaning?
Jouni

Vinvin
Posts: 4911
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Crystal 3.1

Post by Vinvin » Mon Jan 18, 2021 1:26 pm

Dann Corbit wrote:
Mon Jan 18, 2021 3:07 am
Paloma wrote:
Sun Jan 17, 2021 3:34 am
Why is Bluefish-FD-XI with 181/213 not included in the calculation?
I guess he looked at the thread count and saw something nefarious.
If I ask for 1 thread and 4 threads are pounding like mad, then something is up.
It's more the fact that Bluefish-FD-XI is very weak compare to other engines : http://talkchess.com/forum3/viewtopic.p ... 56#p844056
I don't want to promote it.

Dann Corbit
Posts: 11989
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Crystal 3.1

Post by Dann Corbit » Mon Jan 18, 2021 1:37 pm

Why do we care how strong a tactical solver is for playing games?
I do not know for sure that Bluefish uses too many threads, but I do know it uses multi-pv under the covers (which is OK in my opinion).

To me, if a program has 6 Elo compared to the latest SF at 3800 and it can solve all of Blathy's problems it is still the greatest solver of all time.
That program has not been invented yet, but I hope that it does get invented

Chest does not even play chess. But it is a great chess solver for checkmate problems of 12 plies or less.
Same for ChestUCI for deeper problems (yes, Chest can do it too, but it is too complicated to set up, so thanks for writing ChestUCI)
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Jouni
Posts: 2386
Joined: Wed Mar 08, 2006 7:15 pm

Re: Crystal 3.1

Post by Jouni » Mon Jan 18, 2021 2:41 pm

Fully agree Dann. There should be warning in Crystal page: NOT FOR PLAYING!
Jouni

Jouni
Posts: 2386
Joined: Wed Mar 08, 2006 7:15 pm

Re: Crystal 3.1

Post by Jouni » Mon Jan 18, 2021 4:33 pm

"Clean Search" makes engine to crash in Fritz GUI.
Jouni

Post Reply