Best tactical engine?

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

Moderators: hgm, Rebel, chrisw

ArmyBridge

Best tactical engine?

Post by ArmyBridge »

Hi chess friends
I can´t remember if some one have tested the top engines (Rybka 3, Deep Fritz 11, Naum 4, Thinker, Hiarcs !")in tactical set to know ,who is the best tactical engine?, I guess it would be Rybka 3, but I have curiosity to know if some test has taken place

Regards 8-)
Jouni
Posts: 3315
Joined: Wed Mar 08, 2006 8:15 pm

Re: Best tactical engine?

Post by Jouni »

bright 0.4a!

Jouni
User avatar
Eraserheads
Posts: 235
Joined: Fri Mar 10, 2006 9:19 am
Location: Quezon City, Philippines

Re: Best tactical engine?

Post by Eraserheads »

Hi Jouni,

I just got Bright 0.4a

As usual, I let new engines play against the "sparring partners" Genius 7, WChess 2000 and Zarkov. By this method I get to know basically how an engines plays - whether they play positionally, or in an attacking fashion.

Some engines stand out above the others in their distinctly unique playing styles. Engines like Chess Tiger 14 (specifically version 14) plays in the most "Karpov" style; Shredder 5 has a knack to sacrifice pawns, Thinker really plays the most exciting chess now and HIARCS has a pronounced tendency to play in unusual material imbalances, to name a few.

In Bright's case, the first thing I noticed is that it plays in a very unorthodox way. I see that it plays along unusual paths even when employing classical openings like the Queen's Gambit. It makes me think it's playing Chess on another planet. Reminds me of Morozevich and of the "Contemporary Style" Kasparov spoke about.

See how Bright steers the Nimzo-Indian into chaos.

[Event "?"]
[Site "?"]
[Date "2009.04.23"]
[Round "?"]
[White "Genius 7"]
[Black "Bright 0.4a"]
[Result "0-1"]

1. d4 {0s} Nf6 {book 0s} 2. c4 {4s (Nf3)} e6 {book 0s}
3. Nc3 {3s (Nf3)} Bb4 {book 0s} 4. Qc2 {4s} c5 {book 0s}
5. dxc5 {3s} O-O {book 0s} 6. a3 {3s} Bxc5 {book 0s} 7. Nf3
{3s} b6 {book 0s} 8. Bg5 {3s (Bf4)} Bb7 {book 0s} 9. e4
{2s} h6 {+0.14/9 13s} 10. Bh4 {8s} g5 {+0.37/9 5s} 11. Bg3
{13s (b4)} g4 {+0.28/8 9s} 12. Ng1 {3s (b4)} Nc6 {-0.01/8
11s} 13. Qd2 {14s (Nge2)} Nd4 {-0.22/7 8s} 14. Ra2 {48s
(Be5)} a5 {-0.23/7 7s} 15. Be5 {37s (Bd3)} Nc6 {-0.85/6 5s}
16. Bf4 {0s} d5 {-0.61/7 8s} 17. Bxh6 {48s (exd5)} Nxe4
{-0.66/8 6s} 18. Nxe4 {3s} dxe4 {-0.56/9 7s} 19. Qf4 {0s
(Bxf8)} f5 {-0.89/6 7s} 20. Bxf8 {59s} Bxf8 {-0.85/13 0s}
21. Ne2 {52s (Be2)} Bg7 {-0.94/7 4s} 22. Nc1 {5s (h3)} Nd4
{-1.39/7 4s} 23. b3 {15s} e5 {-1.38/9 0s} 24. Qd2 {34s
(Qg3)} a4 {-1.15/7 5s} 25. bxa4 {37s} Kh7 {-0.77/10 0s}
26. Rb2 {10s (h3)} Bh6 {-0.81/8 8s} 27. Qc3 {0s} Rxa4
{-0.81/6 3s} 28. h3 {33s (c5)} g3 {-0.51/5 4s} 29. fxg3 {4s
(c5)} Qg5 {-0.33/4 4s} 30. Ne2 {12s (Rxb6)} Ba6 {-0.18/5
6s} 31. h4 {4s (Nxd4)} Qg7 {-0.77/6 5s} 32. Nc1 {17s
(Nxd4)} Bxc1 {-1.87/6 3s} 33. Qxc1 {4s} Qxg3+ {-1.46/8 1s}
34. Kd1 {4s} Rxa3 {-2.13/8 1s} 35. Rxb6 {3s} Qg4+ {-1.80/7
2s} 36. Ke1 {14s} Re3+ {-2.48/9 0s} 37. Qxe3 {12s} Nc2+
{-2.58/9 11s} 38. Kf2 {0s} Nxe3 {-2.31/1 6s} 39. Rxa6 {6s}
Qf4+ {-3.58/9 0s} 40. Ke2 {4s} Ng4 {-5.57/9 2s} 41. Ra7+
{2s (Ra3)} Kh6 {-6.07/8 3s} 42. Ra6+ {5s} Kh5 {-7.45/11 0s}
43. Kd1 {6s} Nf2+ {-7.50/10 0s} 44. Kc2 {3s} Nxh1 {-7.83/9
0s} 45. Be2+ {5s} Kxh4 {-8.43/12 0s} 46. Ra8 {5s} Ng3
{-8.42/10 0s} 47. Rh8+ {3s} Kg5 {-9.14/11 2s} 48. Rg8+ {2s
(Bd1)} Kf6 {-9.30/10 3s} 49. Bd1 {0s} e3 {-9.30/8 4s}
50. Kb3 {9s} Qd4 {-9.98/9 0s} 51. Bf3 {3s} e2 {-10.36/8 1s}
52. Bxe2 {9s} Nxe2 {-11.25/10 0s} 0-1

What a position!

[D] 3q4/1b5k/1p5b/4pp2/r1Pnp1p1/P1Q5/1R3PPP/2N1KB1R w K -

===============================

A Slav game becomes a wild goose chase here:


[Event "?"]
[Site "?"]
[Date "2009.04.23"]
[Round "?"]
[White "Bright 0.4a"]
[Black "Genius 7"]
[Result "1/2-1/2"]

1. d4 {book 0s} d5 {4s (Nf6)} 2. c4 {book 0s} c6 {2s}
3. Nf3 {book 0s} Nf6 {2s} 4. Nc3 {book 0s} dxc4 {2s (e6)}
5. Ne5 {book 0s} b5 {22s} 6. a4 {+0.21/9 15s} Qb6 {16s
(Bd7)} 7. e3 {+0.32/8 14s} e6 {16s (Nbd7)} 8. Bd2 {+0.22/7
9s} Nbd7 {54s (Bd6)} 9. axb5 {+0.55/8 6s} cxb5 {0s} 10. b3
{+0.10/9 10s} Nxe5 {3s} 11. dxe5 {+0.33/9 3s} Nd7 {4s}
12. bxc4 {+0.25/8 3s} b4 {2s} 13. Na4 {+0.15/8 2s} Qc6 {31s
(Qc7)} 14. f4 {-0.07/6 5s} Bb7 {4s (Ba6)} 15. Qg4 {-0.27/6
7s} Qe4 {2s} 16. Be2 {-0.79/10 11s} Qxg2 {16s} 17. Qxg2
{-0.91/12 0s} Bxg2 {5s} 18. Rg1 {-0.97/12 1s} Bc6 {0s
(Be4)} 19. c5 {-0.71/9 5s} b3 {25s} 20. Bc4 {-1.08/12 0s}
Bxa4 {6s (Rb8)} 21. Rxa4 {-0.28/8 5s} Nxc5 {1s (b2)}
22. Bb5+ {-0.12/8 6s} Kd8 {12s (Nd7)} 23. Rd4+ {-0.40/8 6s}
Kc8 {26s} 24. Be8 {-0.62/10 0s} b2 {21s (Be7)} 25. Ke2
{-0.68/7 4s} Be7 {5s} 26. Bxf7 {-0.68/9 2s} Rf8 {14s}
27. Bh5 {+0.15/12 0s} Rb8 {4s} 28. Rb1 {-0.52/11 3s} Rb3
{4s} 29. Rc4 {-0.14/9 0s} Kd7 {2s} 30. Bc3 {-0.35/7 4s} g6
{3s (Rfb8)} 31. Bf3 {+0.26/7 3s} Rfb8 {2s} 32. Kd2 {+0.27/7
2s} a5 {12s (g5)} 33. Kc2 {+0.20/6 5s} Ra3 {7s (a4)} 1/2-1/2

Check out this position!

[D] r1k1Bb1r/p4ppp/4p3/2n1P3/3R1P2/4P3/1p1BK2P/6R1 b - -
Werewolf
Posts: 1807
Joined: Thu Sep 18, 2008 10:24 pm

Re: Best tactical engine?

Post by Werewolf »

I can answer your question definatively because I spend a great deal of time testing engines on tactical test suits. I have tested engines through 100's of positions and my results are as follows. The following test is a good representative of all the others.

Rybka 3 (45/50)
Bright 0.4a (43/50)
Zappa Mexico II (41/50)
Deep Fritz 10 (40/50)
HIARCS 12.1 (40/50)
Naum 4 (38/50)
Crafty 23.0 (27/50)
Fritz 11 (25/50)

Rybka 3 is the strongest engine tactically. On average it not only solves more positions than Bright 0.4a but solves them quicker too. All this assuming you have 64 bit windows of course!
govert
Posts: 270
Joined: Thu Jan 15, 2009 12:52 pm

Re: Best tactical engine?

Post by govert »

Carl.
I notice that Fruit - and its derivatives - is absent from your list.
Is that because it is bad tactically, or because you did not test it?
ArmyBridge

Re: Best tactical engine?

Post by ArmyBridge »

Werewolf wrote:I can answer your question definatively because I spend a great deal of time testing engines on tactical test suits. I have tested engines through 100's of positions and my results are as follows. The following test is a good representative of all the others.

Rybka 3 (45/50)
Bright 0.4a (43/50)
Zappa Mexico II (41/50)
Deep Fritz 10 (40/50)
HIARCS 12.1 (40/50)
Naum 4 (38/50)
Crafty 23.0 (27/50)
Fritz 11 (25/50)

Rybka 3 is the strongest engine tactically. On average it not only solves more positions than Bright 0.4a but solves them quicker too. All this assuming you have 64 bit windows of course!
Wow very fine!! but I think your miss to test DF11, acording to ChessBase Deep Fritz 11 1CPU is stronger than Fritz 11, so it would be interesting test DF11 instead of Fritz 11, your list bringme me some very news about some engine. For explample once A. Cozzie said that Zappa is a very agressive engine but not too good for tactics!!, I guessed that Hiarcs 12.1 was stronger and I was wrong.. what about Thinker? have you test it?

Regards
gerold
Posts: 10121
Joined: Thu Mar 09, 2006 12:57 am
Location: van buren,missouri

Re: Best tactical engine?

Post by gerold »

govert wrote:Carl.
I notice that Fruit - and its derivatives - is absent from your list.
Is that because it is bad tactically, or because you did not test it?
Fruit has not been worked on much lately so has fallen
behind the other top engines. Speaking for myself
i do not care about testing derivatives or clones.

Best to you,

Gerold.
Dann Corbit
Posts: 12564
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Best tactical engine?

Post by Dann Corbit »

gerold wrote:
govert wrote:Carl.
I notice that Fruit - and its derivatives - is absent from your list.
Is that because it is bad tactically, or because you did not test it?
Fruit has not been worked on much lately so has fallen
behind the other top engines.
According to this list:
http://www.computerchess.org.uk/ccrl/4040/
Toga has the same strength as Hiarcs 12 and Bright 4a.

And this list:
http://www.husvankempen.de/nunn/40_40%2 ... liste.html
shows Fruit 2.4 Beta A to be about the same strength as Deep Sjeng 3.

Both the derivatives and native versions are among the strongest engines in the world.
Speaking for myself
i do not care about testing derivatives or clones.

Best to you,

Gerold.
Whether anyone cares to test them or not is another matter.

The strongest open source engine appears to be Glaurung (specifically the Stockfish derivative).

The strongest free engine is Thinker 5.4 D.
Why is it that Lance can pack so much power into such a tiny can?
Reminds me of something Inigo Montoya once said:
"I wonder if he is using the same wind that we are using?"
Werewolf
Posts: 1807
Joined: Thu Sep 18, 2008 10:24 pm

Re: Best tactical engine?

Post by Werewolf »

govert wrote:Carl.
I notice that Fruit - and its derivatives - is absent from your list.
Is that because it is bad tactically, or because you did not test it?
Tactically Fruit & co are pretty weak. On the test above they are typically in the high 20's.

Also Deep Shredder 11 is weak. The only other notable engine is Deep Junior 10 which is on a par with HIARCS 12.1
Werewolf
Posts: 1807
Joined: Thu Sep 18, 2008 10:24 pm

Re: Best tactical engine?

Post by Werewolf »

ArmyBridge wrote:
Werewolf wrote:I can answer your question definatively because I spend a great deal of time testing engines on tactical test suits. I have tested engines through 100's of positions and my results are as follows. The following test is a good representative of all the others.

Rybka 3 (45/50)
Bright 0.4a (43/50)
Zappa Mexico II (41/50)
Deep Fritz 10 (40/50)
HIARCS 12.1 (40/50)
Naum 4 (38/50)
Crafty 23.0 (27/50)
Fritz 11 (25/50)

Rybka 3 is the strongest engine tactically. On average it not only solves more positions than Bright 0.4a but solves them quicker too. All this assuming you have 64 bit windows of course!
Wow very fine!! but I think your miss to test DF11, acording to ChessBase Deep Fritz 11 1CPU is stronger than Fritz 11, so it would be interesting test DF11 instead of Fritz 11, your list bringme me some very news about some engine. For explample once A. Cozzie said that Zappa is a very agressive engine but not too good for tactics!!, I guessed that Hiarcs 12.1 was stronger and I was wrong.. what about Thinker? have you test it?

Regards
I do not have Deep Fritz 11. It is true it is 80 elo above Deep Fritz 10.1 but tactically Chessbase themselves said the two were roughly equal. I will try to test Thinker in due course but it's lacvk of output in chessbase's GUI makes it less appealing.