Hiarcs is the best engine in blitz

[Tony Thomas]

It managed to become the new clear number one on my list. Note that my list is missing Rybka(commercial), Fritz 11, Loop and Deep Junior. It took away the top spot from the latest sensation Naum 3. As I have noticed Hiarcs isnt doing so well in other lists that doesnt use Own book or allow learning. So that probably means that their claim about having the best book among commercial engines might be true. Here is the list of top 10 on my computer with all the hiarcs versions included for comparison. Ratings were calculated with Bayesian Elo with offset 2300..You can see on the detailed list that the only game Hiarcs lost was against its own predecessor. Only other engine that managed to give Hiarcs some run for its money was Bright, may be it has something to do with Bright having a good book.

Code: Select all

Rank Name                           Elo    +    - games score oppo. draws 
   1 Hiarcs12SPUCI                 2958   64   59   128   82%  2708   21% 
   2 Naum 3                        2922   54   50   165   79%  2692   24% 
   3 Hiarcs11.2SPUCI               2895   51   48   186   74%  2699   16% 
   4 HiarcsX54UCI                  2850   46   44   208   71%  2685   18% 
   5 Rybka v1.0 Beta.w32           2835   37   36   281   69%  2694   25% 
   6 Fruit 2.3                     2831   42   41   227   68%  2695   25% 
   7 Shredder11UCI                 2817   43   42   211   66%  2699   21% 
   8 TogaII 1.4 beta 5c            2816   48   47   168   66%  2698   23% 
   9 Thinker 5.1c(Passive)         2810   51   50   151   64%  2700   21% 
  10 Glaurung 2.0.1 JA             2803   45   44   190   63%  2701   24% 
  11 Ktulu 8.0                     2776   37   37   276   60%  2695   20% 
  12 ChessTiger2007.1 UCI          2762   36   35   301   65%  2650   25% 
  13 Bright-0.3a                   2745   53   53   131   54%  2715   24% 

Here is a detailed report...

Code: Select all

1 Hiarcs12SPUCI             : 2958  128 (+ 91,= 27,- 10), 81.6 %

Rybka v1.0 Beta.w32           :   4 (+  3,=  0,-  1), 75.0 %
Spike 1.2 Turin               :   4 (+  2,=  2,-  0), 75.0 %
Smarthink 1.00                :   4 (+  3,=  0,-  1), 75.0 %
Trace 1.37a                   :   4 (+  2,=  2,-  0), 75.0 %
Gandalf 6.01                  :   4 (+  3,=  1,-  0), 87.5 %
Ktulu 8.0                     :   4 (+  2,=  1,-  1), 62.5 %
Pharaon 3.5.1                 :   4 (+  3,=  1,-  0), 87.5 %
Ruffian 1.0.5                 :   4 (+  2,=  2,-  0), 75.0 %
SlowChess Blitz WV 2.1        :   4 (+  3,=  1,-  0), 87.5 %
Aristarch 4.50                :   4 (+  3,=  1,-  0), 87.5 %
CM10th D2Alos                 :   4 (+  3,=  1,-  0), 87.5 %
ChessTiger2007.1 UCI          :   4 (+  3,=  1,-  0), 87.5 %
Fruit 2.3                     :   4 (+  3,=  0,-  1), 75.0 %
DeepSjeng27                   :   4 (+  3,=  1,-  0), 87.5 %
Delfi 5.2                     :   4 (+  4,=  0,-  0), 100.0 %
Movei00_8_438                 :   4 (+  4,=  0,-  0), 100.0 %
BugChess2_V1_5_2              :   4 (+  4,=  0,-  0), 100.0 %
Shredder11UCI                 :   4 (+  4,=  0,-  0), 100.0 %
Crafty 21.6 JA                :   4 (+  4,=  0,-  0), 100.0 %
Scorpio 2.0                   :   4 (+  4,=  0,-  0), 100.0 %
AlaricWB707                   :   4 (+  3,=  1,-  0), 87.5 %
Glaurung 2.0.1 JA             :   4 (+  3,=  1,-  0), 87.5 %
Hiarcs11.2SPUCI               :   4 (+  0,=  2,-  2), 25.0 %
Zappa Mexico II               :   4 (+  2,=  1,-  1), 62.5 %
Frenzee feb 08                :   4 (+  2,=  1,-  1), 62.5 %
TogaII 1.4 beta 5c            :   4 (+  4,=  0,-  0), 100.0 %
Naum 3                        :   4 (+  1,=  3,-  0), 62.5 %
Sloppy 0.2.0 X                :   4 (+  4,=  0,-  0), 100.0 %
Thinker 5.1c(Passive)         :   4 (+  3,=  0,-  1), 75.0 %
WildCat 8                     :   4 (+  4,=  0,-  0), 100.0 %
Prodeo 1.5                    :   4 (+  2,=  2,-  0), 75.0 %
Bright-0.3a                   :   4 (+  1,=  2,-  1), 50.0 %

[Dr.Wael Deeb]

Thanks Tony for the results and yes,I do agree regarding the opening book issue....

[Tony Thomas]

Thanks Tony for the results and yes,I do agree regarding the opening book issue....

As you can see, many engines that were stronger than Bright got fried by Hiarcs.. Hiarcs is indeed a fish killer on a 32 bit single CPU with own book/learning enabled under super fast time controls.

[Ovyron]

Thanks for this, it's great to see HIARCS over Naum 3

[Tony Thomas]

Thanks for this, it's great to see HIARCS over Naum 3

Are you a Hiarcshoe? I was a Zappahoe until Anthony decided to quit, I am trying to find a new bandwagon when it comes to commercial engines.

Navahoe regards
Tony

[Ovyron]

Currently I'm a Thinkerhoe, I love its playing style and that's all I care about, it has the pants to sacrifice big material if the position coming from it is not much worse than the actual position, and it's amazing at creating fortresses when down on material (then it destroys the fortresses itself because it dislikes drawing the game, and causes a loss, but that's not to blame if it tries to win.)

But HIARCS is currently my favorite commercial, as I dislike Junior (Junior is the most original but originality alone isn't enough.)

[Tony Thomas]

I dont know anything about any of the Junior versions, if I am not mistaken that's the only commercial engine that I havent owned any versions of..If it wasnt for the sucky selling policy by ICC, I would have bought it. I would hate to spend my money on an ICC account and not use it. I love thinker as well, I am pretty sure you have seen my thinker thread a while back. Hopefully Romi wont find out that I am messing with all these hoes.

[genorb]

Hi Tony,

I am also testing chess engines with two different time controls: 40/5 and 1'+1''

If I remember well you are using 1'+1'' too. Within few days (weeks) I will publish a new rating list with Hiarcs 12. But I can already give here a preliminary one. On the rating list I will publish soon, Hiarcs will have played around 400 games and the ratings will be more reliable.

As one can see below Hiarcs 12 performs very well with that fast time control (I do NOT use own book, but I use a common neutral book for all engines). His score (in %) is for the moment better than the one of Rybka 2.3.2a, but the average rating of the opponents is smaller, so the rating of Hiarcs 12 is also smaller. Of course to increase the average rating of the opponent, Hiarcs will need to play against stronger opponents and thus his score could (should?) decrease. So for the moment it is still difficult to know if Hiarcs 12 will beat Rybka 2.3.2a under the condition I test the engines (more details about the condition on my website http://www.ordichec.net in french...).

For your information here is the results of Hiarcs 11.1 and 11.2

Code: Select all

Name                       Elo    +    - games score oppo. draws
Hiarcs 11.1               2723   26   25   602   68%  2584   26%
Hiarcs 11.2               2702   29   28   482   71%  2541   26%

The preliminary rating list

Code: Select all

Rank Name                       Elo    +    - games score oppo. draws 
   1 Rybka 2.3.2a              2791   28   27   575   77%  2574   24% 
   2 Hiarcs 12.0               2778   57   53   152   81%  2529   21% 
   3 Shredder 11               2732   42   41   245   74%  2541   21% 
   4 Naum 3.0                  2707   31   30   444   74%  2523   26% 
   5 Loop 13.6                 2673   20   20   936   61%  2584   27% 
   6 Fruit 2.3.1               2669   29   29   436   67%  2542   25% 
   7 Rybka 1.0                 2667   24   24   649   65%  2554   27% 
   8 Toga II 1.3.4             2635   32   32   347   64%  2534   28% 
   9 Thinker 5.1c              2616   48   47   160   62%  2533   27% 
  10 Zappa Mexico II           2601   29   28   428   59%  2533   32% 
  11 Glaurung 2.0.1            2596   38   38   246   58%  2536   25% 
  12 Deep Sjeng 2.7            2579   29   29   428   54%  2552   23% 
  13 Spike 1.2                 2555   23   23   649   51%  2546   30% 
  14 Alaric 707                2553   28   28   444   51%  2544   25% 
  15 Naum 2.0                  2552   23   23   656   52%  2540   30% 
  16 Bright 0.2c               2539   39   39   242   52%  2523   21% 
  17 Wildcat 8.0               2537   48   48   160   50%  2535   20% 
  18 Slow Chess Blitz WV2.1    2494   23   23   655   47%  2517   30% 
  19 The King 3.33             2490   28   28   471   41%  2557   23% 
  20 Colossus 2007d            2488   35   35   290   44%  2528   24% 
  21 Alfil 8.1.1               2487   39   39   226   46%  2516   30% 
  22 Pharaon 3.5.1             2482   23   23   655   47%  2506   27% 
  23 List 5.12                 2474   23   23   656   47%  2501   24% 
  24 Ruffian 1.0.5             2472   23   23   655   43%  2523   26% 
  25 SmarThink 0.17a           2457   24   24   655   43%  2511   22% 
  26 Pro Deo 1.5               2456   24   24   655   42%  2518   20% 
  27 Scorpio 2.0               2454   39   39   241   41%  2518   23% 
  28 Booot 4.14                2445   40   40   226   41%  2514   24% 
  29 Hamsters 0.6              2444   40   41   226   41%  2514   23% 
  30 SOS 5.1                   2437   24   24   655   42%  2504   23% 
  31 Pseudo 0.7c               2437   23   24   655   41%  2505   25% 
  32 Deep Frenzee 3.0          2426   23   24   656   41%  2494   24% 
  33 Aristarch 4.50            2424   23   23   656   41%  2494   26% 
  34 Delfi 5.2                 2423   35   35   294   38%  2518   24% 
  35 Movei 0.8.438             2410   33   33   344   38%  2505   22% 
  36 Jonny 2.83                2402   24   24   650   38%  2496   20% 
  37 Zappa 1.1                 2399   24   24   649   38%  2488   22% 
  38 Baron 1.8.1               2399   24   25   656   37%  2501   18% 

[AdminX]

It managed to become the new clear number one on my list. Note that my list is missing Rybka(commercial), Fritz 11, Loop and Deep Junior. It took away the top spot from the latest sensation Naum 3. As I have noticed Hiarcs isnt doing so well in other lists that doesnt use Own book or allow learning. So that probably means that their claim about having the best book among commercial engines might be true. Here is the list of top 10 on my computer with all the hiarcs versions included for comparison. Ratings were calculated with Bayesian Elo with offset 2300..You can see on the detailed list that the only game Hiarcs lost was against its own predecessor. Only other engine that managed to give Hiarcs some run for its money was Bright, may be it has something to do with Bright having a good book.

Tony,

I wonder how Hiarcs 11.2 would do in these test if allowed to use Hiarcs 12 opening book? Any ideas?

Ted

[Tony Thomas]

It managed to become the new clear number one on my list. Note that my list is missing Rybka(commercial), Fritz 11, Loop and Deep Junior. It took away the top spot from the latest sensation Naum 3. As I have noticed Hiarcs isnt doing so well in other lists that doesnt use Own book or allow learning. So that probably means that their claim about having the best book among commercial engines might be true. Here is the list of top 10 on my computer with all the hiarcs versions included for comparison. Ratings were calculated with Bayesian Elo with offset 2300..You can see on the detailed list that the only game Hiarcs lost was against its own predecessor. Only other engine that managed to give Hiarcs some run for its money was Bright, may be it has something to do with Bright having a good book.

I wonder how Hiarcs 11.2 would do in these test if allowed to use Hiarcs 12 opening book? Any ideas?

Ted

Not sure, since my computer has an overheating problem, I wont even make any promises that I will give it a try. However, its an interesting idea.

Tony