Well I have been tinkering recently to see if I can get more opponents for my dedicated computers other then playing dedicated computer against dedicated computer. Therefore I decided to look at engines and see if there was an equal way of playing them against dedicated computers.
Well the nice thing about the Mysticum is that you also have the ability to slow it down and then set it into the Start Config to play at this speed when you turn it on.
And that is exactly what I did, I played around with the configs and got the Mysticum to play at a speed of 30 MHz.
Now to get an idea of the MIPS speed of Mysticum I found that at normal 1 GHz speed it runs at about 1384 MIPS. Therefore I figured at 30 MHz it should be running at about 41.5 MIPS (Million Instructions per second)
41.5 MIPS is equivalent of a:
- 6502 running at 130 MHz. and since a 6502 with 5 MHz chess computer ran at around 1.6 MIPS this is still 26 times faster than a 6502 with 5 MHz.
- A Mephisto TM at 36 MHz runs at about 13 MIPS (yes a 6502 was faster per MHz) therefore Mysticum at 30 MHz is the equivalent of a 68030 with 132 MHz.
A 486DX66MHz computer ran at 54 MIPS therefore the Mysticum at 30 MHz is equivalent to a 486 with 50 MHz computer.
For my test matches I am going to use Tasc Chessmachine Gideon 3.1 which uses a 32 MHz Arm 2 processor. I am guessing that Mysticum at 30 Mhz is about 3-4 times faster in its MIPS.
Gideon 3.1 has approximately 2350 ELO. His opponent Deep Sjeng ct 2010 has a massive 3048 ELO at CCRL 40/40 when running on a Athlon 64 X2 4600+ (2.4 GHz) which has a massive 174767 MIPS.
Therefore I was now really curious to see how Deep Sjeng ct 2010 would play at 30 MHz. The games are played at 30 seconds per move for Gideon 3.1 and 40 moves in 20 minutes for Deep Sjeng ct 2010. Deep Sjeng ct 2010 has 1 MB Hash. Gideon 3.1 512K Hash.
Game 1
[Event "Computer Test Match"]
[Site "Pelham, Alabama, USA 30:1"]
[Date "2011.12.26"]
[Round "?"]
[White "Mysticum 30 MHz DS ct 2010"]
[Black "Tasc CM 32 512K Gideon 3.1"]
[Result "1/2-1/2"]
[ECO "D47"]
[Annotator ",Microsoft"]
[PlyCount "117"]
[EventDate "2011.12.26"]
[EventType "match (rapid)"]
1. c4 e6 2. d4 d5 3. Nc3 c6 4. e3 Nf6 5. Nf3 Nbd7 6. Bd3 dxc4 7. Bxc4 b5 8. Bd3
Bb7 9. e4 b4 10. Na4 c5 11. e5 Nd5 12. O-O cxd4 13. Re1 Be7 14. Nxd4 O-O 15.
Qg4 {Mysticum 30 MHz DS ct 2010 out of book} Re8 {
Mysticum 30 MHz DS ct 2010 out of book} 16. Bh6 Bf8 17. Rac1 Rc8 18. Bg5 Rxc1
19. Rxc1 Be7 20. Bxe7 Nxe7 21. Nc5 Nxc5 22. Rxc5 a6 23. Rc4 Qa5 24. Qe2 Rd8 25.
g3 Bd5 26. Rc1 Ba8 27. Rc4 Qxa2 28. f3 Qa5 29. Qe1 Ng6 30. Bxg6 hxg6 31. b3 Rb8
32. Kg2 Bd5 33. Rc2 Qb6 34. Qf2 f6 35. Qe3 Rd8 36. Kf2 fxe5 37. Nc6 Qc7 38.
Nxb4 Qd6 39. Qc5 Bxb3 40. Qxd6 Rxd6 41. Rc8+ Kf7 42. Nc6 Rd2+ 43. Ke3 Rxh2 44.
Nxe5+ Ke7 45. Ra8 Ra2 46. Ra7+ Kd6 47. f4 g5 48. Rd7+ Kc5 49. fxg5 Bd5 50. Nd3+
Kb5 51. Rxg7 Rg2 52. Kf4 Re2 53. g6 Re4+ 54. Kg5 Re3 55. Nf4 Rxg3+ 56. Kf6 Be4
57. Rg8 e5 58. Kxe5 Bxg6 59. Rxg6 1/2-1/2
Sucess! Gideon 3.1 didn't get killed and the game was played well from both sides. Gideon 3.1 had a pawn advantage for most of the game but could not capitalize. DS 2010 even at 30 MHz still amazingly calculated most moves 10 or 11 ply deep and still went as far 13-15 ply in the endgame when there were only a few pieces left. In comparison Gideon 3.1 mostly had 6 to 7 plys and in the endgame went as far as 9 ply.
Another thing to note is that DS 2010 playing normally with 1 Ghz calcluates about 100,000 k/s and upwards as the game progresses. At 30 MHz it calculated at 3000-4000 k/s so the speed reduction works as expected.
Game 2
The second game was a Caro-Kann Panov-Botvinnik Attack in which DS 2010 sacrificed a minor piece for 3 pawns. But this was not enough compensation and DS 2010 ends up losing the game.
[Event "Computer Test Match"]
[Site "Pelham, Alabama, USA 30:1"]
[Date "2011.12.26"]
[Round "?"]
[White "Tasc CM 32 512K Gideon 3.1"]
[Black "Mysticum 30 MHz DS ct 2010"]
[Result "1-0"]
[ECO "B14"]
[Annotator ",Microsoft"]
[PlyCount "151"]
[EventDate "2011.12.26"]
[EventType "match (rapid)"]
1. e4 c6 2. c4 d5 3. exd5 cxd5 4. d4 Nf6 5. Nc3 g6 6. Qb3 Bg7 7. cxd5 O-O 8.
Be2 Nbd7 9. Bf3 Nb6 10. Nge2 {Mysticum 30 MHz DS ct 2010 out of book} Bg4 11.
Bxg4 Nxg4 12. a4 Qc8 {Tasc CM 32 512K Gideon 3.1 out of book} 13. Bf4 Qf5 14.
a5 Bxd4 15. axb6 Bxf2+ 16. Kd2 g5 17. h3 gxf4 18. hxg4 Be3+ 19. Ke1 Qxg4 20.
Rh2 axb6 21. Ra4 Rfc8 22. Nd1 Rxa4 23. Qxa4 Qg6 24. Nxe3 fxe3 25. Rh3 b5 26.
Qd1 Kf8 27. d6 exd6 28. Rxe3 Rc4 29. Qd5 Rg4 30. g3 Rg5 31. Qd4 Qb1+ 32. Kd2
Qg6 33. Qh8+ Qg8 34. Qf6 Qg6 35. Qe7+ Kg8 36. Nf4 Qh6 37. Qxb7 Rc5 38. Qd7 b4
39. b3 Qh2+ 40. Re2 Qh6 41. Qb7 Qf8 42. Qxb4 Qa8 43. Qa4 Qxa4 44. bxa4 Kg7 45.
Re4 Kf6 46. Rb4 Ra5 47. Ne2 Ke6 48. Nd4+ Kf6 49. Nb5 Ke5 50. Rh4 f5 51. Rf4 Ra8
52. Nd4 Rg8 53. Nxf5 d5 54. Kd3 Ra8 55. Nd4 Rg8 56. Rf3 Rg4 57. Re3+ Kd6 58.
Kc3 Kc5 59. Rd3 Re4 60. a5 h5 61. Nb3+ Kc6 62. Kc2 Re2+ 63. Kb1 Re5 64. Kb2 Kb5
65. Kc3 Rg5 66. Kd4 Kc6 67. Re3 Rg4+ 68. Ke5 Rg5+ 69. Kf6 Rg4 70. Ke6 d4 71.
Rf3 Re4+ 72. Kf7 Rg4 73. Rf6+ Kc7 74. Rg6 d3 75. Rxg4 hxg4 76. Ke6 1-0
Game 3
This time Gideon 3.1 makes a hash of things shortly after the opening and could never really recover and DS 2010 wins.
[Event "Computer Test Match"]
[Site "Pelham, Alabama, USA 30:1"]
[Date "2011.12.26"]
[Round "?"]
[White "Mysticum 30 MHz DS ct 2010"]
[Black "Tasc CM 32 512K Gideon 3.1"]
[Result "1-0"]
[ECO "D39"]
[Annotator ",Microsoft"]
[PlyCount "117"]
[EventDate "2011.12.26"]
[EventType "match (rapid)"]
1. d4 Nf6 2. Nf3 d5 3. c4 e6 4. Bg5 dxc4 5. Nc3 Bb4 6. e4 c5 7. e5 h6 8. exf6
hxg5 9. fxg7 Rg8 10. a3 {Tasc CM 32 512K Gideon 3.1 out of book} Bxc3+ 11. bxc3
cxd4 {Mysticum 30 MHz DS ct 2010 out of book} 12. Qxd4 Qxd4 13. Nxd4 Nd7 14.
Bxc4 Rxg7 15. Kd2 Rh7 16. Rae1 a5 17. Nf3 Rg7 18. Re3 g4 19. Nd4 Nb6 20. Bd3
Rg5 21. Rg3 e5 22. Re1 Rh5 23. f4 Rxh2 24. Rxe5+ Kf8 25. Rb5 Rh6 26. f5 Bd7 27.
Rb1 Rd8 28. Rb3 Ba4 29. Rb2 Rdd6 30. Rxg4 Rh1 31. Nc2 Bxc2 32. Kxc2 Rhh6 33.
Rd4 Kg7 34. Rb5 Rxd4 35. cxd4 Rc6+ 36. Kd2 Rd6 37. g4 Kf6 38. Be4 Rxd4+ 39. Ke3
Rd6 40. Kf4 Kg7 41. g5 a4 42. Rb4 Kf8 43. Bxb7 Ke8 44. Ke5 Ke7 45. f6+ Kd7 46.
Bf3 Kc7 47. Bh5 Rd5+ 48. Kf4 Rd7 49. Bxf7 Rxf7 50. Kf5 Nd7 51. Rc4+ Kb7 52.
Rxa4 Rh7 53. Rf4 Nf8 54. g6 Rh5+ 55. Kg4 Rh8 56. g7 Rg8 57. Kf5 Nd7 58. Rg4 Kc6
59. f7 1-0
Score after 3 games:
Mysticum 30 MHz DS ct 2010 - 1.5
Tasc CM 32 512K Gideon 3.1 - 1.5
Well as a dedicated chess computer fan, I always suspected that modern engines would find that their Grandfathers would still be tough opponents to beat if they played under similar conditions. I think this great match shows this to be the case. There is still lots of life left in the old dogs.
Now this has opened up all sorts of future great matches between dedicated engines and dedicated chess computers
But not only this, a normal human being can now also have great matches again with engines!
I will still try to tinker around with Mysticum to see if further slowdown is possible, but for now 30 MHz seems to be perfect competition for dedicateds with all those ancient engines and new engines at my disposal. And all this without a keyboard or monitor! WOW!
Match to be continued...............
Best regards,
Nick