Collector's Corner..On Collecting Dedicated Chess Computers

[mychess]

Hello,

A game between L'Empereur and Novag Savant I (ie not II).
L'Empereur was at level 4, 30 seconds per move, with pondering (impossible to deactivate).
Novag Savant was at level 3, 45 to 90 seconds per move, without pondering (impossible to activate ).

Code: Select all

 1. d4   d5
 2. c4   e6
 3. Nc3  Nf6
 4. cxd5 exd5
 5. Bf4  Nc6
 6. Nf3  Bd7
 7. e3   Bb4
 8. Bd3  O-O
 9. O-O  Re8
10. Qb3  Bg4
11. a3   Bxf3
12. axb4 Be4
13. Nxe4 dxe4
14. Bc4  Qe7
15. b5   Nd8
16. Ra3  c5
17. Rfa1 cxd4
18. Rxa7 Rxa7
19. Rxa7 d3
20. Qa3  d2
21. Be2  Qxa3
22. bxa3 Nd5
23. Bg4  Nc3
24. b6   f5
25. Bd1  Nxd1
26. Ra8  Nc3
27. g3   d1=Q+
28. Kg2  Qf3+
29. Kg1  Nd1
30. h4   Qxf2+
31. Kh1  Nxe3
32. Bxe3 Qxe3
33. Kh2  Qxb6
34. Kg2  Re5
35. Rc8  Ra5
36. a4   Rxa4
37. Rc2  Qe6
38. Rd2  Ra2
39. Rxa2 Qxa2+
40. Kg1  Nc6
41. Kf1  e3
42. Kg1  Qf2+
43. Kh1  e2     and Savant saw and announced chekmate.

Cordialy.

3563391568F2D393

[Steve B]

Hi Mychess

And yet another L'emp game makes the record books!
we went from hardly any published games by L'emp to now an avalanche of games
Still ...the number of actual wins remains anemic...

L'emp On The March Regards
Steve

[paulwise3]

A game between L'Empereur and Novag Savant I (ie not II).
L'Empereur was at level 4, 30 seconds per move, with pondering (impossible to deactivate).
Novag Savant was at level 3, 45 to 90 seconds per move, without pondering (impossible to activate ).

Hi, thanx for the game,
but it seems a bit unfair that L'Empereur played at 30 sec/move instead of 45 secs/move.

Regards, Paul

[mychess]

Hi Steve,

The revenge of L'Empereur :
Novag Savant was at level 3, 45 to 90 seconds per move, without pondering (impossible to activate ).
L'Empereur was at level 4, 30 seconds per move, with pondering (impossible to deactivate).

Code: Select all

 1. e4    e6
 2. d4    d5
 3. Nc3   Bb4
 4. e5    c5
 5. a3    Bxc3+
 6. bxc3  Qc7
 7. Nf3   Ne7
 8. Be3   c4
 9. Be2   O-O
10. O-O   Bd7
11. Qd2   Nf5
12. Rad1  Nxe3
13. Qxe3  Ba4
14. Qd2   Nc6
15. Ng5   Qe7
16. Ra1   f6
17. exf6  Rxf6
18. Rfe1  h6
19. Nh3   Qd6
20. Rab1  b6
21. f4    Re8
22. Qc1   Kh7
23. Bf3   Ref8
24. g3    b5
25. Nf2   R6f7
26. Bg4   Re7
27. Bh3   Rf6
28. Ng4   Rff7
29. Bg2   Rb7
30. Ne5   Nxe5
31. Rxe5  Rf6
32. Qb2   Rbf7
33. Rbe1  g6
34. Bh3   Re7
35. R5e2  a5
36. Qa1   Re8
37. Qb2   h5
38. Re3   Kh6
39. R3e2  Qc6
40. Re5   Re7
41. R5e2  Qd6
42. Re5   Re8
43. R5e3  Kh7
44. R3e2  Kg7
45. Re3   Kh6
46. Re5   Qc6
47. R5e2  Re7
48. Re5   Qd6
49. R5e2  Re8
50. Bg2   Ref8
51. h3    h4
52. gxh4  Kh5
53. Re5+  Kxh4
54. Kh1   Rc8
55. Rg5   Qf8
56. Ree5  Rxf4
57. Rxg6  Re8
58. Rgxe6 Rxe6
59. Rxe6  Rf1+
60. Bxf1  Qxf1+
61. Kh2   Qxh3+
62. Kg1   Qxe6
63. Kf2   Qf5+
64. Ke3   Qxc2
65. Qa1   Qe4+
66. Kf2   Qf5+
67. Ke3   Qg5+
68. Kf3   Bd1+
69. Kf2   Qd2+
70. Kf1   Kg3
71. Qb2   Qxb2
72. Ke1   Qe2#            0-1

Honour is safe !

To Paulwise3 : Yes, Savant doesn't have a fixed time level, but in the instruction manual of l'Empereur, it is wrtten : AVERAGE DURATION BETWEEN EACH MOVE .
And L'Empereur use pondering, Savant doesn't.

Cordialy.

[Tibono2]

Hi gents,
thanks for sharing L'Empereur games. Suggest this might deserve a dedicated thread?
All the best,
Tibono

[Steve B]

In Game 2...SII opens with the rare 1.f4
both computers were out of book at move 3
the game quickly dissolved into a closed position with each side shuffling rooks on the back ranks
i decided to adjudicate the game a draw

[Date "2024.3.20"]
[Round "2"]
[White "Savant II"]
[Black "L'empereur"]
[TimeControl "45 sec./avg."]
[Result "1/2-1/2"]
1.f4 d5 2.Nf3 Nf6 3.e3 Bg4 4.h3 Bxf3 5.Qxf3 Nc6 6.Bb5 e6 7.Bxc6+ bxc6 8.O-O Rb8 9.Nc3 Bd6 10.b3 O-O 11.d4 c5 12.Bd2 cxd4 13.exd4 c6 14.Rfe1 Qb6 15.Be3 Rb7 16.Bf2 Rfb8 17.Rad1 Qb4 18.Bh4 Be7 19.a4 a5 20.g4 Ne4 21.Nxe4 Bxh4 22.Nf2 Bf6 23.Qe3 g6 24.g5 Be7 25.Nd3 Qb6 26.Re2 Bd6 27.Kh1 Ra7 28.Rg1 Raa8 29.Ra1 Ra6 30.Rf1 Ra7 31.Rb1 Rab7 1/2-1/2

FINAL POSITION


I decided to adjudicate the game a draw here
the position is closed with material even and both sides shuffling rooks along the back ranks for the last 5 moves
i guess i could have let this play on for more moves but i was afraid i would die of old age before they moved a different piece


L'emp leads 1.5-.5 Regards
Steve

[Steve B]

Hi gents,
thanks for sharing L'Empereur games. Suggest this might deserve a dedicated thread?
All the best,
Tibono

Hi Eric
Glad you are followng the L'emp games
thank you for your suggestion
my "Collectors Corner" thread has been on going since 2007 with close to 686k views and 3500 posts
it contains many small matches like this with rare chess computers
perform a word search on any old dedicated chess computer and there's a good chance there will be a post about it in this thread
im comfortable keeping the games here
anyway...with the upcoming Summer Olympic games in Paris this year it seems fitting that L'emp would make its return now after all these years

L'empereur And The Games Of The XXXIII Olympiad Regards
Steve

[paulwise3]

Hi Steve,

As promised I imported this second game too in Lichess,
See https://lichess.org/jtMxUnCt
The Lichess SF-engine agreed more or less with your adjucation being it a draw.
But in a number of positions it suggested that L'Empereur could get a slight advantage by moving it's c-pawn to c5.

Awaiting the next game regards,
Paul

[paulwise3]

Hi mychess,

Here your games in pgn:

[Event "Testmatch mychess"] [Site "Hiarcs.net"] [Date "2024.03.19"] [Round "1"] [White "CLJ Industries L'Empereur, 30 sec./avg"] [Black "Novag Savant II, 30-60 sec/avg"] [Result "0-1"] [PlyCount "88"] 1. d4 d5 2. c4 e6 3. Nc3 Nf6 4. cxd5 exd5 5. Bf4 Nc6 6. Nf3 Bd7 7. e3 Bb4 8. Bd3 O-O 9. O-O Re8 10. Qb3 Bg4 11. a3 Bxf3 12. axb4 Be4 13. Nxe4 dxe4 14. Bc4 Qe7 15. b5 Nd8 16. Ra3 c5 17. Rfa1 cxd4 18. Rxa7 Rxa7 19. Rxa7 d3 20. Qa3 d2 21. Be2 Qxa3 22. bxa3 Nd5 23. Bg4 Nc3 24. b6 f5 25. Bd1 Nxd1 26. Ra8 Nc3 27. g3 d1=Q+ 28. Kg2 Qf3+ 29. Kg1 Nd1 30. h4 Qxf2+ 31. Kh1 Nxe3 32. Bxe3 Qxe3 33. Kh2 Qxb6 34. Kg2 Re5 35. Rc8 Ra5 36. a4 Rxa4 37. Rc2 Qe6 38. Rd2 Ra2 39. Rxa2 Qxa2+ 40. Kg1 Nc6 41. Kf1 e3 42. Kg1 Qf2+ 43. Kh1 e2 { black announces mate in 1 (44... e1=Q#)} 0-1
[Event "Testmatch mychess"] [Site "Hiarcs.net"] [Date "2024.03.20"] [Round "2"] [White "Novag Savant II, 30-60 sec/avg"] [Black "CLJ Industries L'Empereur, 30 sec./avg"] [Result "0-1"] [ECO "C19"] [Annotator "Paul,Wiselius"] [PlyCount "144"] [EventDate "2024.??.??"] [TimeControl "1800"] 1. e4 e6 2. d4 d5 3. Nc3 Bb4 4. e5 c5 5. a3 Bxc3+ 6. bxc3 Qc7 7. Nf3 Ne7 8. Be3 c4 9. Be2 O-O 10. O-O Bd7 11. Qd2 Nf5 12. Rad1 Nxe3 13. Qxe3 Ba4 14. Qd2 Nc6 15. Ng5 Qe7 16. Ra1 f6 17. exf6 Rxf6 18. Rfe1 h6 19. Nh3 Qd6 20. Rab1 b6 21. f4 Re8 22. Qc1 Kh7 23. Bf3 Ref8 24. g3 b5 25. Nf2 R6f7 26. Bg4 Re7 27. Bh3 Rf6 28. Ng4 Rff7 29. Bg2 Rb7 30. Ne5 Nxe5 31. Rxe5 Rf6 32. Qb2 Rbf7 33. Rbe1 g6 {?} 34. Bh3 Re7 35. R5e2 (35. Qb4 {!} Qd7 36. Rxd5) 35... a5 36. Qa1 Re8 37. Qb2 h5 38. Re3 Kh6 39. R3e2 Qc6 40. Re5 Re7 41. R5e2 Qd6 42. Re5 Re8 43. R5e3 Kh7 44. R3e2 Kg7 45. Re3 Kh6 46. Re5 Qc6 47. R5e2 Re7 48. Re5 Qd6 49. R5e2 Re8 50. Bg2 Ref8 51. h3 h4 52. gxh4 Kh5 53. Re5+ Kxh4 54. Kh1 Rc8 55. Rg5 Qf8 56. Ree5 Rxf4 57. Rxg6 Re8 58. Rgxe6 {??} (58. Qc1 {! with mate in 5}) 58... Rxe6 59. Rxe6 Rf1+ 60. Bxf1 Qxf1+ 61. Kh2 Qxh3+ 62. Kg1 Qxe6 63. Kf2 Qf5+ 64. Ke3 Qxc2 65. Qa1 Qe4+ 66. Kf2 Qf5+ 67. Ke3 Qg5+ 68. Kf3 Bd1+ 69. Kf2 Qd2+ 70. Kf1 Kg3 71. Qb2 Qxb2 72. Ke1 Qe2# 0-1

To Paulwise3 : Yes, Savant doesn't have a fixed time level, but in the instruction manual of l'Empereur, it is wrtten : AVERAGE DURATION BETWEEN EACH MOVE .
And L'Empereur use pondering, Savant doesn't.

About the pondering ability: When rating the strength of a computer, the ability to ponder is irrelevant. Of course it plays better when it can. That adds to the strength of play. But you cannot decide to give it a shorter computing time to compensate that. Because if pondering results in an expected move that is not the played move by the opponent, the pondering has been useless, and the computer has less time to compute the answer to this unexpected move. There are a lot of good chess computers that cannot ponder, see for example the Excalibur Alexandra.

To ponder or not to ponder regards,
Paul

[Steve B]

Hi Steve,

As promised I imported this second game too in Lichess,
See https://lichess.org/jtMxUnCt
The Lichess SF-engine agreed more or less with your adjucation being it a draw.
But in a number of positions it suggested that L'Empereur could get a slight advantage by moving it's c-pawn to c5.

Awaiting the next game regards,
Paul

Thanks again Paul for the Lichess port
Replaying the games on Lichess is very entertaining!

Mychess Playing On Lichess Regards

Steve

[mychess]

Hi Paulwise3,


The Novag computer is a Savant, not a Savant II ("Savant I" to differentiate it from the "Savant II" used by Steve).

I agree with you that pondering is difficult to quantify.
However, if the move considered by The Emperor is the move played by Savant, 150% to 300% of the time (from 45+30 to 90+30 instead of 30 seconds) was used for his response: probably a move played at the next level (and 30 seconds for The Emperor is an average time, not a fixed time).

We can check the influence of pondering by replaying the game with an immediate response to The Emperor's move (pondering disabled is not possible): I think this is the only way to remove the advantage of pondering from one of the two computers and check its effect.
In the case of a "deep move" played by The Emperor in a game (piece's capture, mate in 7 ), it would be interesting to check this pondering effect.

Maybe Steve can use the Novag clock to calculate the Savant's time, and for L'Empereur another clock (Steve is a faster player, but 70 moves multiplied by 3 to 5 seconds is a lot of time). And we'll check L'Empereur's average time per move.


Cordialy.

[mychess]

Hi again,

We can evaluate how many times the pondering effect is correct : By using "Conseil" key, L'Empereur give you which move is the better (no, it is not a false response, L'Empereur doesn't cheat or lie ).

Cordialy.

Mychess

[Steve B]

Interestingly both the Savant and Savant II/Savant Royale manuals seem to suggest that the Savants do ponder
read page 7 of the manual at the top of the page:
http://alain.zanchetta.free.fr/docs/Nov ... ant_EN.pdf

however i cant recall the Savants ever replying instantly to an opponents move(which would be a telltale sign that it was thinking on the opponents time and correctly predicted the move)

perhaps the Savants do think while the opponent is on the move but this dosent result in a quicker response ,but instead it results in the computer still using the time allotted to think a bit deeper into the position?

Ill Have To Think About That Regards
Steve

[Steve B]

In Game 3....Both computers were out of book at move 3
SII wins a P early on but soon after blunders away a minor piece
In an effort to punish me for calling game 2 an early draw.. L'emp prolongs the game for 92 moves
refusing to bring its QR into the game until move 82 and bypassing several forced mates along the way

[Date "2024.3.21"]
[Round "3"]
[White "L'empereur"]
[Black "Savant II"]
[TimeControl 45 sec./avg]
[Result "1-0"]

.e4 e5 2.Bc4 c6 3.Nc3 b5 4.Be2 b4 5.Na4 Nf6 6.d4 Qa5 7.b3 Nxe4 8.Nf3 exd4 9.Qxd4 d5 10.Ng5 Nf6 11.Bf4 Bb7 12.Bxb8 Be7 13.Be5 O-O 14.Qf4 Nd7 15.Bc7 Bxg5 16.Qg3 Bd2+ 17.Kxd2 Nb6 18.Nxb6 axb6 19.Kd1 Rfe8 20.Be5 g6 21.Qf4 c5 22.Qf6 Kf8 23.Qg7+ Ke7 24.Bf6+ Ke6 25.Bg4+ Kd6 26.Qxf7 Bc8 27.Bxc8 Rexc8 28.Qxh7 Rg8 29.Qe7+ Kc6 30.Qe6+ Kc7 31.Be5+ Kb7 32.Qxd5+ Ka6 33.Bf6 Raf8 34.Re1 Ra8 35.f4 Rgf8 36.Bg5 Ka7 37.Re7+ Ka6 38.Re6 Rh8 39.Bf6 Rhg8 40.Re2 Rgf8 41.Qg5 Rg8 42.Qe5 Ra7 43.Qd5 Rf8 44.Re6 Rh7 45.h4 Ka7 46.Qd3 Qa3 47.Qxg6 Rd7+ 48.Ke1 Kb7 49.Qe4+ Ka6 50.Qc6 Rb7 51.Qxc5 Rff7 52.Rc6 Ka7 53.Qd4 Qa5 54.Be5 Rh7 55.g3 Rbd7 56.Rd6 Rde7 57.Rd5 Qa3 58.Kd1 Kb7 59.Bd6 Reg7 60.Bxb4 Qa7 61.Rg5 Rc7 62.Bd6 Rhd7 63.Rd5 Rc6 64.Be5 Rxd5 65.Qxd5 Qa8 66.Qd7+ Ka6 67.Qd3+ Ka7 68.f5 Rc5 69.Qd7+ Qb7 70.Qa4+ Ra5 71.Qf4 Qd5+ 72.Bd4 Qxf5 73.Qc7+ Ka8 74.Qc6+ Kb8 75.Qxb6+ Kc8 76.Qc6+ Kb8 77.Qe8+ Kb7 78.Qe3 Qg4+ 79.Kc1 Rd5 80.c3 Qh5 81.Kb2 Rd7 82.Re1 Qb5 83.g4 Kc8 84.h5 Kb7 85.h6 Kc8 86.Qe8+ Kb7 87.h7 Qxb3+ 88.axb3 Rxh7 89.Qe4+ Kc7 90.Qxh7+ Kd8 91.Bb6+ Kc8 92.Qc7# 1-0

L'Empereur wins the match 2.5-.5

On The Road To Paris Regards
Steve

[kamoj]

Nice!
Thank you for the L'empereur games. Very much appreciated.
I 'm also glad the the Savant's were part of this "revival".
/Lars