What will be the strongest engine at the end of the year?

[Frank Quisinsky]

Say the programmer which created the most interesting engine I ever saw in middlegame!

For a vote ...
Most interesting playing style ... Mikko ... you are on first position!

38 years I am sitting at home and looking computerchess.
Why I must so long wait for such an engine?

Not important ...
Now I have it.

Special thanks to yourself.
You made me really happy with Hakkapeliitta.

Best
Frank

[Frank Quisinsky]

Hi there,

more interesting as to compare current TOP engines like Komodo and Stockfish with older Commercials products (most have a boring playing style) is the playing style.

Such a vote is much more interesting as Elo.

We can't win vs. 2.700 Elo and can't win vs. 3.100 Elo.
But we can have more fun with an interesting playing style.

But to your vote.
It's a hard fight between Stockfish and Komodo.
The Stockfish Group with all the programmers working on it are very strong. An person developed an engine alone can't not hold the playing strength. It's a wonder Komodo team can hold it at the moment.

Maybe older commercial programmers can hold the style after looking in the code of Stockfish. If not we have the Situation we have so many years ... each year 30 - 50 Elo more ... if we are looking in Computer chess history.

Best
Frank

[JJJ]

Komodo and Stockfish are obvious bets. Shredder and Fritz could both easily compete for the number one spot as well. Both Alfil and Andscacs have made very fast progress recently so they have a small chance too.

What do you know about Shredder and Fritz development plz ? Let's assume Fritz 15 is Rybka 5. Rybka 4 needed something around 160 elo to equalize Komodo 9.1, but probably more until then. It is a lot.

What about Shredder now ? The last version known needed something around 300 elo. What do you know we don't to think they can make it until the end of the year ?

[Steve Maughan]

What do you know about Shredder and Fritz development plz ? Let's assume Fritz 15 is Rybka 5. Rybka 4 needed something around 160 elo to equalize Komodo 9.1, but probably more until then. It is a lot.

What about Shredder now ? The last version known needed something around 300 elo. What do you know we don't to think they can make it until the end of the year ?

I spoke with Stefan at the World Championship about Shredder. He only started this rewrite in January 2015. It's a new magic bitboard implementation - the "old" version used a letter-box structure. As many will know, a pure bitboard structure enables fast computation of complex evaluation terms. So this should give quite a boost to the engine. He said the current version is already stronger than v 12.

Stefan also told me about his new multi-threaded approach which is different to anything else. He said he got 75 ELO going from 2 to 4 processors and 50 ELO gong from 4 to 8 (I think I have the # processors right, but I'm going from memory) - which is excellent.

And of course Shredder won the Software WCC, so it's definitely no slouch.

He's planning to release a version before the end of the year.

As I said, I think this will be a milestone release.

- Steve

[Sean Evans]

"Fire, (Firebird, Fire xTreme)
an UCI compliant chess engine by Norman Schmidt [1], until version 3.0 derived from IvanHoe and the Ippolit series of programs with some help of Milos Stanisavljevic. Initially called Firebird, and later renamed to Fire due to a trademark naming conflict [2], it was released as open source, Fire licensed under the GNU GPL. The sources were later closed with Windows executables available for download for recent Intel processors [3]. Fire features magic bitboards, it can be configured with more than 70 UCI options, and applies a SMP parallel search."
source: https://chessprogramming.wikispaces.com/Fire
homepage: http://chesslogik.wix.com/fire

I see, yet another Ippo-clone So, now we have Rybka, Houdini and Fire with a myriad of other lesser known Ippo-clones, so much for originality in chess engines

Cordially,

Sean

[mar]

What about Shredder now ? The last version known needed something around 300 elo.

Well, Ginkgo on CEGT 40/20 is about as strong as Rybka 4. So Shredder should be close; either his SMP implementation is much better or he improved that much. Not bad I guess

[gerold]

I think SF will hold on and be number one by the end of this year.

[Robert Pope]

"Fire, (Firebird, Fire xTreme)
an UCI compliant chess engine by Norman Schmidt [1], until version 3.0 derived from IvanHoe and the Ippolit series of programs with some help of Milos Stanisavljevic. Initially called Firebird, and later renamed to Fire due to a trademark naming conflict [2], it was released as open source, Fire licensed under the GNU GPL. The sources were later closed with Windows executables available for download for recent Intel processors [3]. Fire features magic bitboards, it can be configured with more than 70 UCI options, and applies a SMP parallel search."
source: https://chessprogramming.wikispaces.com/Fire
homepage: http://chesslogik.wix.com/fire

How did this happen? It's my understanding that if you modify a program that has a GPL license, and distribute it in any way, you are required to release the source, as well.

[yanquis1972]

im really surprised to see komodo ahead by so much...i thought it was generally agreed that k-team will be struggling to keep up with SF? i havent paid much attention to ratings lately, is komodo significantly ahead of the SF dev version??

[Frank Quisinsky]

Hi Steve,

I know that Stefan do a lot in GUI too.
I nice update will be available.

But the biggest problem Shredder have "all the time" are all the fast lost games. From TOP-42 Shredder is here on place 38 ... means 37 other engines have clearly better results.

This one must be changed. This one is important. Also is Shredder an engine, stronger in endgames. From such engines we have really enough. Fore 10 years it was nice too see, also in combination with tablebases, today we have a lot of engines, stronger in endgames.

Shredder need a complete new face, a clearly more interesting style. But such problems a lot of commercial engines have.

I think Shredder should be improved in tactic a lot.
More as a lot ... for a commercial engines, tactic must be clearly better and for sure king safty with many pieces on board. For a Commercial engine the right word is ... disaster!

Have a look here ...
FCT League 1 statistic

The same is to see now in FCT League 2 statistic
Can be see after Arasan 18.0 POP x64 test-run

Code: Select all

     Programs                         Elo   +   - Games Score Av.Op. Draws   Won Draw Lost
 01. Komodo 9 x64                   : 3078  41 39  135  71.5%  2918  51.1%    62   69    4
 02. Stockfish 26Apr2015 BMI2 x64   : 3064  30 27  191  68.6%  2928  61.8%    72  118    1
 03. Houdini 4 STD B x64            : 3018  70 67   49  64.3%  2916  51.0%    19   25    5
 04. Hakkapeliitta 3.0 x64          : 2996  36 34  118  59.7%  2928  66.9%    31   79    8
 05. Fire 4 x64                     : 2982  28 27  157  57.0%  2933  73.2%    32  115   10
 06. Sting SF 4.8.4 x64             : 2978  28 25  140  56.8%  2930  77.9%    25  109    6
 07. Protector 1.8.0 x64            : 2975  25 23  162  55.6%  2936  79.0%    26  128    8
 08. GullChess 3.0 BMI2 x64         : 2973  28 27  140  56.1%  2931  76.4%    25  107    8
 09. Critter 1.6a x64               : 2966  26 25  141  53.5%  2941  80.1%    19  113    9
 10. Texel 1.05 x64                 : 2940  35 35  118  50.4%  2938  68.6%    19   81   18
 11. Chiron 2.0 x64                 : 2936  28 28  149  49.3%  2940  74.5%    18  111   20
 12. Nirvanachess 2.1c POP x64      : 2930  26 26  159  49.1%  2936  76.7%    17  122   20
 13. Equinox 3.30 x64               : 2928  28 28  137  47.8%  2943  76.6%    13  105   19
 14. Naum 4.6 x64                   : 2918  29 30  143  45.8%  2947  73.4%    13  105   25
 15. Andscacs 0.81 POP x64          : 2904  34 35  144  45.5%  2935  63.2%    20   91   33
 16. Rybka 4.1 SSE42 x64            : 2897  28 30  134  43.7%  2942  75.4%     8  101   25
 17. Hannibal 1.5 x64               : 2885  27 30  125  42.4%  2938  76.8%     5   96   24
 18. Sjeng c't 2010 w32             : 2883  30 32  134  41.0%  2946  71.6%     7   96   31
 19. Hiarcs 14 WCSC w32             : 2850  37 38  137  37.2%  2941  56.9%    12   78   47
 20. Senpai 1.0 SSE42 x64           : 2820  34 36  158  33.9%  2937  56.3%     9   89   60
 21. Shredder 12 x64                : 2809  33 35  181  30.4%  2953  53.0%     7   96   78

78 games??????

Best
Frank