Millennium Chess Genius and The King tests

[spacious_mind]

Here are some tests that I completed with Millennium computers:

Millennium Chess Geniuses




Millenium Chess Genius Exclusive clearly showed that it is the best Millennium Genius scoring 171 ELO better than Chess Genius. Surprisingly Chess Genius Pro did not do as well as Chess Genius even though its 120 MHz it is much faster than the 48 MHz of Chess Genius.

This is how they compare with other Lang programs that i have tested in the past:

Richard Lang Comparison




Only Richard Lang;s Champ Classic running in DOSBox on an I7 at full speed which is estimated at Pentium with 374 MHz speed slightly outperformed MCGE.

So how do the Millennium Lang's compare with Millenniun The King Element?

Millennium The King Element



Well to me it is pretty clear that The King is rated higher than 2500 ELO it more in the 2600 ELO range performing at Grandmaster level.

Here is how The King Element compares with other tested Johan de Koning Kings:

The King Comparison



Surprisingly although the scores varied slightly it scored exactly the same as XBox Chessmaster 10000. Only my Mysticum Chessmaster 11 personalty and DOSBox King 2.55 running at Pentium 374 MHz scored higher than Millennium The King Element playing at 300 MHz.

I am planning to test all of the 5 play styles and also if I can find my CM11K personality I will see if I can add that as well to the Millennium The King Element

Best regards

[spacious_mind]

Well I just had to play one game with The King against one of my favorites Amiga Sargon 3 68060 using WINUae on an I7 Pentium. This is the same Sargon 3 that is playing in Division 2 Tournament with DOS programs and dedicated computers.

[Event "The King Test Match"]
[Site "https://lichess.org/GGNmutFx"]
[Date "12/22/2018"]
[Round "1"]
[White "I7-WinUAE Amiga 68060 Sargon 3 (2600)"]
[Black "Millennium 300 MHz The King Element Active (2600)"]
[Result "1/2-1/2"]
[WhiteElo "?"]
[BlackElo "?"]
[Variant "Standard"]
[TimeControl "-"]
[ECO "D46"]
[Opening "Semi-Slav Defense: Bogoljubov Variation"]
[Termination "Normal"]
[Annotator "lichess.org"]

1. c4 c6 2. Nf3 Nf6 3. Nc3 d5 4. e3 e6 5. d4 Nbd7 6. Bd3 Be7 { D46 Semi-Slav Defense: Bogoljubov Variation } 7. O-O O-O 8. Bd2 dxc4 9. Bxc4 b5 10. Bd3 Bb7 11. Qc2 b4 12. Ne2 c5 13. Ng5 Rc8 14. Nxh7 cxd4 15. Qb1 dxe3 16. Bxe3 Re8 17. Nxf6+ Nxf6 18. Bxa7 Qd5 19. f3 Red8 20. Rd1 Qg5 21. Ng3 Bc5+ 22. Bxc5 Qxc5+ 23. Kh1 Qf2 24. a4 Bxf3 25. gxf3 Qxf3+ 26. Kg1 Ng4 27. Bh7+ Kh8 28. Rxd8+ Rxd8 29. Qe1 Kxh7 30. Qe4+ Qxe4 31. Nxe4 Rd4 32. a5 Rxe4 33. a6 Rd4 34. a7 Rd8 35. a8=Q Rxa8 36. Rxa8 Nf6 37. Ra7 Kg6 38. Ra6 Nd5 39. Kf2 Kf6 40. Kf3 g5 41. Ke4 Kg7 42. Rc6 Kg6 43. Ra6 Kh5 44. Ra7 f5+ 45. Ke5 f4 46. Kxe6 Ne3 47. Ra8 Kg4 48. Ke5 Kf3 49. Kd4 Nf1 50. Rg8 g4 51. Rf8 Nxh2 52. Ke5 g3 53. Rxf4+ Ke2 54. Re4+ Kd1 55. Rd4+ Kc1 56. Rd3 Nf1 57. Rd8 Kxb2 58. Rf8 Ne3 59. Ke4 Nd1 60. Rg8 Kc3 61. Rxg3+ Kc4 62. Rg7 Kc3 63. Rc7+ Kd2 64. Rb7 Kc3 65. Ke5 b3 66. Rc7+ Kb2 67. Kd4 Nf2 68. Kc4 Nd1 69. Re7 Ka2 70. Ra7+ Kb2 71. Rb7 Ne3+ 72. Kd3 Nd1 73. Rb6 Ka2 74. Ra6+ Kb2 75. Kd2 Nc3 76. Ra7 { The game is a draw. } 1/2-1/2

Wow this was a tough and interesting game where it felt as if both programs were playing cat and mouse against each other.

[fen]2rr2k1/1b3pp1/4pn2/8/Pp6/3B1PN1/1P3qPP/RQ1R3K w - a3 0 24[/fen]

In this position King Element sacrifices his Bishop of a pawn with 24. … Bxf3

[fen]2rr2k1/5pp1/4p3/8/Pp4n1/3B1qN1/1P5P/RQ1R2K1 w - - 2 27[/fen]

A few moves later due to all the pressure Sargon 3 had to do the same giving up its Bishop with the move 27. Bh7+

[fen]8/5ppk/4p3/8/Pp1rN1n1/8/1P5P/R5K1 w - - 1 32[/fen]

In this position Sargon 3 plays 32. a5! and The King has no choice but to take the Knight and then quickly retreat back to sacrifice its Rook for the otherwise unstoppable pawn.

[fen]R7/5ppk/4pn2/8/1p6/8/1P5P/6K1 w - - 1 37[/fen]

So now you have an endgame with Sargon 3 left with Rook and 2 pawns against The King's Knight and 4 pawns.

Tpically with dedicated chess computers and old chess programs the Knight and 2 pawns end up winning the game usually because of the drop off in dedicated computer endgame skills.

However not with Sargon 3, who usually plays quite a strong endgame, both of them in fact played a strong cat and mouse endgame with neither of them making any errors.

Final Position

[fen]8/R7/8/8/8/1pn5/1k1K4/8 w - - 21 76[/fen]

Lichess Evaluation

https://lichess.org/GGNmutFx

I7-WinUAE Amiga 68060 Sargon 3 (2600)
1 Inaccuracies
1 Mistakes
0 Blunders
6 Average centipawn loss

Millennium 300 MHz The King Element Active (2600)
2 Inaccuracies
1 Mistakes
0 Blunders
6 Average centipawn loss

A Grandmasterly performance. It looks like these two programs will play some interesting games against each other. Who knows perhaps Sargon 3 might have finally met its match against a dedicated computer chess program

Best regards

[dedicate computers]

Hi Nick

Even program with different machines, different processors, does not make sense the comparison in my opinion. Openings book should also be the object of analysis, since it corresponds to 60% of the performance of the machine. Strange table.
Regards
Oswald

[spacious_mind]

Hi Nick

Even program with different machines, different processors, does not make sense the comparison in my opinion. Openings book should also be the object of analysis, since it corresponds to 60% of the performance of the machine. Strange table.
Regards
Oswald

Hi Oswaldo,

Thats the whole point to compare different processors and different machines under the same test conditions. Thats the only way to to get a pretty accurate comparison between different machines.

By your argument a BT tests and Colditz test would have to be even stranger?

Besides it doesnt make sense right that as result of these tests I can pick a program like Sargon 3 and randomly play it with its 1988 small opening book against Millennium with both book 1 and book 2 on and used against Sargon 3, yet Sargon 3 has no problems surviving the opening.

BTW all the above programs have pretty good opening books. Why do you think I did not choose Chessmachine to play against The King? Well its obvious as a result of those tests that it would be a waste of time playing King at 300 MHz the gap is too big.

So what part of the test seems strange to you? Also what dedicated computer are you going use to play against The King at 300 MHz? The answer is zero unless you only use Resurrection and Revelations engines like Rybka, Shredder and Hiarcs.

Best regards

[Wardy]

2600 is where I thought the King would land judging by some of the early games against the R30. That's why I thought Steve's WunderMachine would struggle, however that's not the case so far at least.....

Looks like a great opponent for club players, the sacrificial style probably hurts the rating overall but must be great fun to play against.

[spacious_mind]

2600 is where I thought the King would land judging by some of the early games against the R30. That's why I thought Steve's WunderMachine would struggle, however that's not the case so far at least.....

Looks like a great opponent for club players, the sacrificial style probably hurts the rating overall but must be great fun to play against.

Yes the styles are interesting, reminds me of the great CM9000 days where you spent ages messing around creating your own styles. I might still have the styles from those days on an old hard drive somewhere. Just need to search and hopefully find them.

Currently testing Normal style. Active style that I just tested and played is a lot of fun.

Best regards

[Yarc]

2600 is where I thought the King would land judging by some of the early games against the R30. That's why I thought Steve's WunderMachine would struggle, however that's not the case so far at least.....

Looks like a great opponent for club players, the sacrificial style probably hurts the rating overall but must be great fun to play against.

Yes the styles are interesting, reminds me of the great CM9000 days where you spent ages messing around creating your own styles. I might still have the styles from those days on an old hard drive somewhere. Just need to search and hopefully find them.

Currently testing Normal style. Active style that I just tested and played is a lot of fun.

Best regards

Hi Nick,

Very interesting comparisons! TheKing module is certainly an excellent machine to have. I'm impressed by the rating 2608, that is very high. With an ELO of 2500+ that's Grandmaster rating. I wonder how well it would fair against a human player of that rating? Nonetheless, it's still impressive.

I think the engine was written to be fun. No doubt a personality could be created that would defeat one machine but then it may not do so well against others. I think one of the good reasons to have customizable personalities is to have an almost endless supply of chess opponents to play against.

Regards
Ray

[lexman]

Extremely interesting tests and the game against Sargon 3 is very impressive for a dedicated, and a good game to.
As regards playing humans Topschach are saying they will play test games with the King module agianst a GM and an IM towards the end of Jauary
regards Lex

[Yarc]

As regards playing humans Topschach are saying they will play test games with the King module agianst a GM and an IM towards the end of Jauary
regards Lex

Yes, that will also be very interesting to see. I believe Topschach are of the thinking that against Human players with ELO of 2500+/- that The King module would not do so well. Only games will reveal the truth.

Regards Ray

[spacious_mind]

2600 is where I thought the King would land judging by some of the early games against the R30. That's why I thought Steve's WunderMachine would struggle, however that's not the case so far at least.....

Looks like a great opponent for club players, the sacrificial style probably hurts the rating overall but must be great fun to play against.

Yes the styles are interesting, reminds me of the great CM9000 days where you spent ages messing around creating your own styles. I might still have the styles from those days on an old hard drive somewhere. Just need to search and hopefully find them.

Currently testing Normal style. Active style that I just tested and played is a lot of fun.

Best regards

Hi Nick,

Very interesting comparisons! TheKing module is certainly an excellent machine to have. I'm impressed by the rating 2608, that is very high. With an ELO of 2500+ that's Grandmaster rating. I wonder how well it would fair against a human player of that rating? Nonetheless, it's still impressive.

I think the engine was written to be fun. No doubt a personality could be created that would defeat one machine but then it may not do so well against others. I think one of the good reasons to have customizable personalities is to have an almost endless supply of chess opponents to play against.

Regards
Ray

Hi Ray,

Yes it would be interesting to see some 2500-2650 rated GMs play against it. In the example game I posted I would suspect a GM would have difficulty to win in that particular game as both programs played almost faultlessly if you look at the evaluations from Lichess. Perhaps the sacrifices might even give the GM's some fits as nowadays people just don't play like that as often anymore after having been brainwashed by Stockfish.

Regards

[lexman]

IM Roman Vidonyak is 2442 fide so on paper not out of the ballpark for the King. However I think the issue is that stronger humans can use anti computer chess strategies quite successfully against older programs when they know they are playing a machine and this can distort the result. On the other hand The King in active style does sacrifice and randomize quite a bit so there is at least some interest in the match

[Yarc]

Hi Ray,

Yes it would be interesting to see some 2500-2650 rated GMs play against it. In the example game I posted I would suspect a GM would have difficulty to win in that particular game as both programs played almost faultlessly if you look at the evaluations from Lichess. Perhaps the sacrifices might even give the GM's some fits as nowadays people just don't play like that as often anymore after having been brainwashed by Stockfish.

Regards

Hi Nick,

I use stockfish to help me find blunders in these dedicated matches, but as clever as the programming is behind SF, I agree that computer chess at this level is diluting the game. This is one of the reasons I prefer dedicated computer chess and especially liking TheKing's style of play. The other reason is that I don't stand a chance of understanding what SF is doing!

Regards Ray

[spacious_mind]

Extremely interesting tests and the game against Sargon 3 is very impressive for a dedicated, and a good game to.
As regards playing humans Topschach are saying they will play test games with the King module agianst a GM and an IM towards the end of Jauary
regards Lex

That is good to hear. It is long past due for top players to play computers again. It does irritate me though that players should feel that have to play boring anti computer chess to win. Just wish they would sit down and play their normal game and just enjoy the experience.

Personally I would have more satisfaction at winning playing my game rather than preparing some repetitive stuff in order to win. For me that smacks of lack of creativity if one has to resort to parrot fashon repetitiveness. If you have to resort to that then whats the point of even bothering to play it.

But that's me, I enjoy losing magnificently and occasionally winning even more magnificently Those are moments to treasure.

Best regards

[Steve B]

Hi Nick

nice testing as usual
remind me again..these tests are done at 30 sec a move(active chess) or longer?

Forgetful Regards
Steve

[spacious_mind]

Hi Nick

nice testing as usual
remind me again..these tests are done at 30 sec a move(active chess) or longer?

Forgetful Regards
Steve

Hi Steve,

I am too lazy to do 3 minute per game so I always do the tests at 30 seconds per move. 3 minutes per move just takes far too long if you have so many to do.


Regards