Sinclair Spectrum 48K Emulation

[spacious_mind]

I have not played with a Spectrum emulator for a while so this morning I felt like messing around with the Fuse Emulator for Spectrum.

The Sinclair 48K was a Z80 home computer that came out in 1982 (36 years ago!) with 48K RAM and the CPU ran at 3.5 Mhz.

So anyway on the Fuse Emulator I set up a game between Superchess 3.5 written by Chris Whittington in 1985 (33 years ago) against Mephisto Monte Carlo IV Limited Edition from 1990, a chess program by Ed Schroeder.

Mephisto Monte Carlo IV Limited Edition - 1990



http://www.spacious-mind.com/html/monte ... iv_le.html

Fuse Emulator - Spectrum 48K - Superchess 3.5 - 1985



Chris Whittington famous for Chess System Tal, created his first programs on the Spectrum 16K in 1983 with Superchess 1.0. It was only later around 1987 that he started working on Commodore Amiga/Atari ST and later still with DOS and PC.



Originally Superchess 3.5 on a Spectrum 48K Z80 3.5 MHz had a play strength of approximately 1500 ELO.

Well, I knew that Superchess 3.5 wouldn't have much of a chance playing at this speed against Monte Carlo 4 LE, so therefore instead I set up Superchess 3.5 to play 960 seconds per move = 16 minutes per move and therefore Superchess 3.5 had 32 times more thinking time than MC4 LE! who played at 30 seconds per move. However speeding up the emulator to 3200% the thinking time was equalized where both played at around 30 seconds per move in real time. In fact the average thinking time that Superchess 3.5 used had it played in real time was 13 minutes and 36 seconds per move.

MC4 LE is rated at 2016 ELO in Active Chess at Schachcomputer.Info.

https://www.schachcomputer.info/html/ak ... liste.html


105 Mephisto Monte Carlo IV LE 2016 218 (+ 92,= 46,- 80), 52.8 %

Tasc R30 V2.2 1 (+ 0,= 0,- 1), 0.0 %
Mephisto Magellan 2 (+ 2,= 0,- 0), 100.0 %
Phoenix Resurrection Ruffian 1 (+ 0,= 0,- 1), 0.0 %
Mephisto Risc 2 / CM 14 MHz Gideon 3.1 1 (+ 0,= 0,- 1), 0.0 %
Mephisto Mondial 8 (+ 7,= 0,- 1), 87.5 %
Mephisto Academy 10 (+ 4,= 4,- 2), 60.0 %

At Schacomputer.Info you can see that MC4 LE played 218 Active chess games in order to obtain its active chess rating of 2016 ELO. The above is just a small example of the opponents it played against.

OK back to Superchess 3.5, if you take a speed doubling factor of 100 ELO per doubling you would have the following:

Superchess 3.5 Speed Doubling

30S = 1500 ELO
60S = 1600 ELO
120S = 1700 ELO
240S = 1800 ELO
480S = 1900 ELO
960S = 2000 ELO

Therefore with the 960 seconds per move setting, I figured that Superchess 3.5 might be able to compete with Mephisto Monte Carlo IV Limited Edition playing at 30 seconds per move.

Here is the game:

[Event "Test Match"]
[Site "Pelham, AL"]
[Date "2018.03.03"]
[Round "1"]
[White "SP48K Superchess 3.5, 960S."]
[Black "Mephisto Monte Carlo IV LE, 30S."]
[Result "1-0"]
[ECO "A29"]
[WhiteElo "2000"]
[BlackElo "2016"]
[Annotator "spacious_mind"]
[PlyCount "167"]
[EventDate "2018.03.03"]
[EventCountry "USA"]

1. d4 d5 2. e3 e6 {SP48K Superchess 3.5 out of book} 3. Bb5+ {Mephisto Monte Carlo 4 LE out of book} Bd7 4. Nc3 Bb4 5. Bd3 c5 6. dxc5 Bxc3+ 7. bxc3 Qa5 8. Bd2 Qxc5 9. Rb1 Bc6 10. Nf3 Qa3 11. Nd4 b6 12. Qc1 Qxc1+ 13. Rxc1 Nf6 14. Nb5 Kd7 15. f3 e5 16. Rb1 a6 17. Na3 Kc7 18. Be2 b5 19. O-O h5 20. h4 Nbd7 21. Kh2 Rae8 22. Kh3 Re6 23. g4 hxg4+ 24. fxg4 Ne4 25. Be1 f6 26. h5 Rhe8 27. Rb4 Ndc5 28. Nb1 a5 29. Rb2 Na4 30. Rxb5 Ng5+ 31. Kh2 Bxb5 32. Bxb5 Rb8 33. c4 Nb6 34. Bxa5 Kb7 35. cxd5 Nxd5 36. Bd2 Ne4 37. Re1 Nxd2 38. Nxd2 Nc3 39. Bd3 e4 40. Bc4 Rd6 41. Nb1 Rc8 42. Bb3 Nxa2 43. Kg3 Nb4 44. Re2 Nd5 45. c4 Nb4 46. Nd2 Rd3 47.
Ba4 Kb6 48. Kf2 Ra3 49. Bd7 Nd3+ 50. Kg3 Rd8 51. Bf5 Nc1 52. Rg2 Rxe3+ 53. Kf4 Rdd3 54. Bxe4 Re2 55. Nf1 Rxg2 56. Bxg2 Rd4+ 57. Kf5 Rxc4 58. Ne3 Rd4 59. Kg6 Rd7 60. Bf1 Nd3 61. Bxd3 Rxd3 62. Nf5 Rd7 63. Nxg7 Rd4 64. Kf5 Kc6 65. Ne6 Rd2 66. Nf4 Rd4 67. h6 Rd7 68. Kg6 Rd6 69. h7 f5+ 70. Kg5 Rd8 71. gxf5 Rh8 72. Kh6 Kc7 73. Kg7 Rxh7+ 74. Kxh7 Kd6 75. Ng6 Kd5 76. f6 Ke6 77. Kg7 Kd5 78. f7 Kc4 79. f8=Q Kd5 80. Qb4 Ke6 81. Qc4+ Kd7 82. Qc5 Kd8 83. Ne5 Ke8 84. Qf8# 1-0

Critical Position

[fen]8/6p1/1k3pK1/5N1P/6P1/3r4/8/8 w - - 0 62[/fen]

Overall for most of the game Mephisto Monte Carlo 4 LE seemed to be the better program, but missed a few good opportunities to win.

In the above position however Superchess 3.5 has the edge and Mephisto has only one move that draws which is 62. ... Rd1! everything else loses. MC4 LE played 62. ... Rd7? and so loses a game in the endgame which perhaps it should have won.

Lichess Evaluation

https://lichess.org/x4d5FSQh

SP48K Superchess 3.5, 960S. (2000)
6 Inaccuracies
6 Mistakes
1 Blunders
34 Average centipawn loss

Mephisto Monte Carlo IV LE, 30S. (2016)
10 Inaccuracies
6 Mistakes
3 Blunders
49 Average centipawn loss

Final Position

[fen]4kQ2/6K1/8/4N3/8/8/8/8 w - - 0 84[/fen]

Well a speed doubling of 5 times 100 ELO seems to be about right giving Superchess 3.5 a rating of @ 2000 ELO playing on approximate hardware of a Z80 in a range somewhere between 94 to 112 MHz in speed.

How much more it could improve is impossible to know as 999 seconds per move is the max setting for Superchess 3.5. I played at 960 seconds per move.

To conclude, Superchess 3.5 on good hardware is another good program for players who play at around 2000 ELO. If you are a 1500 ELO player, then you would also probably have fun playing Superchess 3.5 on an original Sinclair Spectrum 48K computer or emulator.

Best regards

[microhenri]

Hi spacious_mind
I still like the idea to run old chess software on new hardware. Although it is emulated it is possible to speed it up and compare it with newer chess programs or engines. This gives an idea of how good (or bad) the algorithm was at that time.

One question:
In this case you compare the old super Chess 3.5 that runs on a 4MHz Z80A processor of a ZX-Spectrum, with a chess program that runs on an 8MHz 65C02 processor. With your theory the Z80A processor should run at about 100MHz. Is the algorithm of the Monte Carlo so much better (by a factor 12?). I do believe that chess engine algorithms did improve over the years but did it improve this much in only 5 years?
The super Chess program did win, so this could be an indication that you gave the super Chess too much time?

Is your assumption of 100 ELO points per doubling correct?

I was wondering when you give super Chess less time (480s or 240s or less) when does it start to lose.

microHenri

[spacious_mind]

Hi spacious_mind
I still like the idea to run old chess software on new hardware. Although it is emulated it is possible to speed it up and compare it with newer chess programs or engines. This gives an idea of how good (or bad) the algorithm was at that time.

One question:
In this case you compare the old super Chess 3.5 that runs on a 4MHz Z80A processor of a ZX-Spectrum, with a chess program that runs on an 8MHz 65C02 processor. With your theory the Z80A processor should run at about 100MHz. Is the algorithm of the Monte Carlo so much better (by a factor 12?). I do believe that chess engine algorithms did improve over the years but did it improve this much in only 5 years?
The super Chess program did win, so this could be an indication that you gave the super Chess too much time?

Is your assumption of 100 ELO points per doubling correct?

I was wondering when you give super Chess less time (480s or 240s or less) when does it start to lose.

microHenri

Hi Henri,

If I remember the 6502 is approximately 3.5 times faster than the z80. Therefore a Spectrum 48k in speed was approximately the equivalent of a 6502 Commodore 64 which had 1 Mhz.

Therefore if you take MC4 LE you have 6502 8 MHz times 3.5Mhz = z80 28 MHz. With me playing the emulator at 960 seconds per move = approximately 112 MHz = approximately a doubling factor of 2:

Factor 0 = 28 MHz
Factor 1 = 56 MHz
Factor 2 = 112 MHz

MC4 LE is also a much better chess program with much better opening book written 5 years later. Also a 6502 I believe has access to larger tables therefore its processing is probably more efficient than a z80.

The Program size for Superchess 3.5 is 37KB, which leaves 11 KB for sound, graphics and program evaluations. Therefore I also think that MC4 LE with its dedicated 8KB RAM is more efficient in this area as well.

So I think the comparison between Spectrum 48K 112 MHz and MC4 LE is about right.

ps. also FYI the z80 used on Spectrum had a speed of 3.5 MHz (not 4 MHz that seems to be a mathematical round up that was used)

Regards

[spacious_mind]

Here is a test game this time between Spectrum 48K Colossus 4 against Mephisto Monte Carlo IV LE.

Colossus 4 played this time at the setting of 59 minutes and 59 seconds per move. = 3599 seconds per move. Which would be approximately a Z80 with 420 MHz speed!

Spectrum 48K Colossus 4 - 1986 - Martin Bryant



If you were to play a match between Colossus 4 against Superchess 3.5 on a regular Sinclair Spectrum 48K then overall Colossus 4 would win fairly easily. It was rated at around 1700 ELO.

However my experience with Colossus 4 also shows that this program's improvements through speed doubling are not as impressive as Chris Whittington's programs, therefore setting it up at its maximum per move setting of 3599 seconds or the equivalent of running it at Z80 420 MHz would be an interesting challenge for MC4 LE.

You see not all chess programs were born equal.

Here is the game:

[Event "Test Match"]
[Site "Pelham, AL"]
[Date "2018.03.03"]
[Round "1"]
[White "Mephisto Monte Carlo IV LE, 30S."]
[Black "SP48K Colossus 4, 60M."]
[Result "0-1"]
[ECO "A40"]
[WhiteElo "2016"]
[BlackElo "2000"]
[Annotator "spacious_mind"]
[PlyCount "116"]
[EventDate "2018.03.03"]
[EventCountry "USA"]

1. d4 e6 2. c4 Bb4+ {Mephisto Monte Carlo IV LE out of book} 3. Bd2 {SP48K Colossus 4 out of book} Nc6 4. Bxb4 Nxb4 5. e4 Ne7 6. a3 Na6 7. Nf3 O-O 8. Nc3 f5 9. Bd3 fxe4 10. Nxe4 d5 11. Neg5 h6 12. Nh3 Nf5 13. O-O Bd7 14. Qb3 Bc6 15. Nf4 Qd6 16. g3 Rf6 17. c5 Qf8 18. Rfe1 Re8 19. Rac1 g5 20. Nh5 Rf7 21. Bxa6 bxa6 22. Qc2 Bb5 23. Ne5 Rh7 24. g4 Nh4 25. a4 Bd7 26. f4 Kh8 27. fxg5 Nf3+ 28. Nxf3 Qxf3 29. Nf6 Ree7 30. c6 Be8 31. Qd2 hxg5 32. Qxg5 Reg7 33. Qe5 Bf7 34. Rf1 Qb3 35. Nxh7 Kxh7 36. Rf2 Qxa4 37. g5 Qa5 38. Qf6 Qb4 39. Rg2 Bg6 40. Rc3 Qd6 41. Re2 Bf5 42. Rb3 Qe7 43. Qh6+ Kg8 44. Rb8+ Kf7


Critical Position 1

[fen]1R6/p1p1qkr1/p1P1p2Q/3p1bP1/3P4/8/1P2R2P/6K1 w - - 0 45[/fen]

OMG MC4 LE so far has totally outplayed Colossus 4. In the above position I was a few moves away from resigning for Colossus 4 and putting it out of its misery. In this position MC4 LE had a forced mate with the move 44. Rg2! which of course MC4 LE did not see. MC4 LE played 45. Qh5+ which was not as good but still devastating.

45. Qh5+ Rg6 46. h4 Qf6 47. Re5 Bd3 48. Qh7+ Qg7 49. Qxg7+ Rxg7 50. Rb7 Kg6 51. Rxe6+ Kh5 52. Rh6+ Kg4


Critical Position 2

[fen]8/pRp3r1/p1P4R/3p2P1/3P2kP/3b4/1P6/6K1 w - - 0 53[/fen]

MC4 LE missed dozens of moves to win this game sooner. In this position now after chasing black's King to g4, I began to worry that MC4 LE would mess it all up and lose this game after all. Black's Bishop on d3 and the potential move of black's Rook to e7 would threaten checkmate if MC4 LE is not careful.

In this position the correct and winning move is 53. Rb3! which neutralizes all Black's threats.

But of course MC4 LE did not see any of this and played 53. Rxa7?.

53. Rxa7 Re7


Critical Position 3

[fen]8/R1p1r3/p1P4R/3p2P1/3P2kP/3b4/1P6/6K1 w - - 0 54[/fen]

Now MC4 LE is in trouble. The move 54. Rf6! can still force a draw. Everything else loses. MC4 LE played 54. Kf2? and loses! 54. Kf2? forces a checkmate in 5. This is a good example that shows that with it's 6502 8 MHz processor MC4 LE is not capable of searching deeply enough to see threats such as checkmate in 5!

54. Kf2 Re2+ 55. Kf1 Kg3 56. Rf6 Re8+ 57. Kg1 Re1+ 58. Rf1 Rxf1# 0-1


Final Position

[fen]8/R1p5/p1P5/3p2P1/3P3P/3b2k1/1P6/5rK1 w - - 0 59[/fen]

Well this game as in the previous game shows that MC4 LE has better control of the opening and middle game, and this is where it needs to win its games against a speeded up Spectrum program. The speeded up Spectrum programs however in endgames are superior because the extra search depth allows them to see winning opportunities easier.

This game also hopefully shows that when speeded up Chris Whittington's programs have a lot more potential for improvement than Martin Bryant and his Colossus programs.

So MC4 LE loses its first two games as a result of its own self inflicted errors.

Lichess Evaluation

https://lichess.org/kp8GHhfy

Mephisto Monte Carlo IV LE, 30S. (2016)
1 Inaccuracies
7 Mistakes
4 Blunders
82 Average centipawn loss

SP48K Colossus 4, 60M. (2000)
3 Inaccuracies
10 Mistakes
3 Blunders
67 Average centipawn loss

Best regards

[spacious_mind]

Here is another Test match this time between Spectrum 48K Clock Chess 89 written by Chris Whittington and Ed Schroeder's Mephisto Monte Carlo IV Limited Edition.

Sinclair Spectrum 48K - Clock Chess 89 - Chris Whittington - 1989



Clock Chess 89 is essentially from appearance the same as Superchess 3.5 from 1985. It does however have other levels which it's predecessors did not have. For example:

Average seconds per move.
Infinite
Matching times
Easy
Clock Chess (time per game)

The average seconds per move still maxed out at 999 seconds per move.

Here is the game:

[Event "Test Match"]
[Site "Pelham, AL"]
[Date "2018.03.04"]
[Round "1"]
[White "SP48K Clock Chess 89, 960S."]
[Black "Mephisto Monte Carlo IV LE, 30S."]
[Result "0-1"]
[ECO "A28"]
[WhiteElo "2000"]
[BlackElo "2016"]
[Annotator "spacious_mind"]
[PlyCount "110"]
[EventDate "2018.03.04"]
[EventCountry "USA"]

1. c4 e5 2. Nf3 Nc6 {SP48K Clock Chess 89 out of book} 3. Nc3 {Mephisto Monte Carlo IV LE out of book} Nf6 4. d4 e4 5. d5 exf3 6. dxc6 fxg2 7. cxd7+ Qxd7 8. Qxd7+ Bxd7 9. Bxg2 O-O-O 10. Be3 Be6 11. c5 Be7 12. c6 b6 13. O-O Ng4 14. Bf4 g5 15. Bc1 f5 16. e4 Rhg8 17. exf5 Bxf5 18. Nd5 Bd6 19. h3 Ne5 20. Nf6 Rg7 21. Re1 g4 22. hxg4 Nxg4 23. Nh5 Rgg8 24. Ng3 Rdf8 25. Be3 Nxf2 26. Bxf2 Bxg3 27. Bxg3 Rxg3 28. Kh2 Rfg8 29. Re8+ Rxe8 30. Kxg3 Rg8+ 31. Kh2 Kd8 32. Rd1+ Ke7 33. Bh3 Bxh3 34. Kxh3 Rd8 35. Re1+ Kf7 36. Rf1+ Kg6 37. Rg1+ Kh6 38. b4 a5 39. a3
axb4 40. axb4 b5 41. Re1 Rd6 42. Re5 Rxc6 43. Rxb5 Kg6 44. Kg4 Rc4+ 45. Kf3 h5 46. Rb8 Rc3+ 47. Ke4 Rc1 48. Rg8+ Kh7 49. Rg5 Kh6


Critical Position

[fen]8/2p5/7k/6Rp/1P2K3/8/8/2r5 w - - 0 50[/fen]

The game followed the same pattern as the previous two test games. MC4 LE had opening and middle game advantages and by the time this position was reached the advantages slipped away and the game seemed to be drawn.

But this time in this position Clock Chess 89 makes a critical error with the move 50. Rc5? which allows the exchange of rooks and loses the game because of Black's extra pawn advantage.

50. Rc5 Rxc5 51. bxc5 Kg5 52. Kf3 h4 53. Kg2 Kh6 54. c6 Kh5 55. Kf2 Kg5 0-1


Critical Position

[fen]8/2p5/2P5/6k1/7p/8/5K2/8 w - - 0 56[/fen]

Ok I resigned here for Clock Chess, perhaps I should have a played on a little longer but it seems clear to me that MC4 LE wins this game.

Lichess Evaluation

https://lichess.org/oMaDZD37

SP48K Clock Chess 89, 960S. (2000)
3 Inaccuracies
2 Mistakes
1 Blunders
35 Average centipawn loss

Mephisto Monte Carlo IV LE, 30S. (2016)
2 Inaccuracies
3 Mistakes
0 Blunders
16 Average centipawn loss

Other than the additional levels it is hard to tell if there are any chess programming improvements made on Clock Chess 89 when compared to its predecessor Superchess 3.5 from 1985.

Clock Chess 89 came out for the Spectrum +3 which had 128 KB ram so perhaps 48K is too small for it?

Anyway MC4 LE gets its first win!

Best regards

[spacious_mind]

Ok here is the final test for today. In the previous games I used Spectrum 48K which in some ways seems to be slightly handicapped by its 48K leaving very little RAM for move processing. Therefore this time I thought I would use Sinclair Spectrum +3 which has the same z80 3.5 Mhz but 128KB of RAM which should allow for more processing RAM space.

Sinclair Spectrum +3 = Clock Chess 89 - Chris Whittington - 1989



The level setting was the same with 960 seconds per move and 3200% emulator speed increase. Here is the game:

[Event "Test Match"]
[Site "Pelham, AL"]
[Date "2018.03.04"]
[Round "1"]
[White "SP+3 Clock Chess 89, 960S."]
[Black "Mephisto Monte Carlo IV LE, 30S."]
[Result "1/2-1/2"]
[ECO "A00"]
[WhiteElo "2000"]
[BlackElo "2016"]
[Annotator "spacious_mind"]
[PlyCount "96"]
[EventDate "2018.03.04"]
[EventCountry "USA"]

1. Nf3 d5 2. b4 {Mephisto Monte Carlo IV LE out of book} e6 {SP+3 Clock Chess 89 out of book} 3. c3 a5 4. bxa5 Rxa5 5. d4 Bd7 6. Nbd2 Ba4 7. Nb3 Nf6 8. Bg5 Be7 9. Qb1 Rb5 10. e3 Rb6 11. Bd3 Nbd7 12. O-O h6 13. Bf4 c5 14. c4 cxd4 15. exd4 dxc4 16. Bxc4 Rb4 17. Rc1 Nb6 18. Bd3 Bc6 19. Bd2 Ra4 20. Ne5 Qa8 21. Nc5 Bxc5 22. dxc5 Nbd7 23. Nxc6 bxc6 24. Be3 O-O 25. Rc3 Nd5 26. Rc1 Rb8 27. Qc2 Nb4 28. Qd2 Rd8 29. Rd1 Nxd3 30. Qxd3 Rxa2 31. Rxa2 Qxa2 32. Qf1 e5 33. Rd6 Qa4 34. g3 f5 35. Qd3 f4


Critical Position

[fen]3r2k1/3n2p1/2pR3p/2P1p3/q4p2/3QB1P1/5P1P/6K1 w - - 0 36[/fen]

This was the first game where a Spectrum program actually showed some advantages in the middle game. In the above position Spectrum +3 Clock Chess 89 misses a good opportunity by not playing 36. gxf4! exf4 37. Bd4 with a really good advantage.

Instead Spectrum +3 Clock Chess 89 played 36. Rxd7 which only draws as MC4 LE can force this through repetition of checks!

36. Rxd7 Qa1+ 37. Kg2 Rxd7 38. Qxd7 fxe3 39. Qe6+ Kh7 40. fxe3 Qc3 41. Qf5+ Kg8 42. Qc8+ Kh7 43. Qxc6 Qxe3 44. Qd5 Qe2+ 45. Kh3 Qh5+ 46. Kg2 Qe2+ 47. Kh3 Qh5+ 48. Kg2 Qe2+ {Draw by 3x repetition} 1/2-1/2

Final Position

[fen]8/6pk/7p/2PQp3/8/6P1/4q1KP/8 w - - 0 49[/fen]

Lichess Evalaution

https://lichess.org/U9rEKacR

SP+3 Clock Chess 89, 960S. (2000)
3 Inaccuracies
4 Mistakes
0 Blunders
29 Average centipawn loss

Mephisto Monte Carlo IV LE, 30S. (2016)
3 Inaccuracies
5 Mistakes
0 Blunders
30 Average centipawn loss

Although this game ended in a draw, my feeling is that if I were to play a tournament then I would use the Spectrum +3 128 KB RAM to give the Spectrum programs a little more processing power.

Conclusion

There are several Tournaments that you could set up to see how the performance improves. In every case I would use Spectrum +3 128 KB:

Tournament 1

Play the programs at approximately the same speed setting which would mean the Spectrum programs play at 240 seconds per move and Emulator speed factor of 800% to simulate 30 seconds real move time. You then add some dedicated computers 30 seconds per move to compare the performance equally. I would probably pick the following for such a match:

Dedicated Computers
Mephisto Monte Carlo 4 LE (Ed Shroeder)
Fidelity Avantgarde V2 (Dan & Kathe Spracklen)
Mephisto Roma 68000 (Richard Lang)
Novag Super Expert or Forte C (David Kittinger)

Spectrum Programs
Cyrus is Chess (Richard Lang?)
Colossus Chess (Martin Bryant)
Super Chess 16K - 1983 (Chris Whittington)
Super Chess II 2.1 - 1984 (Chris Whittington)
Super Chess III 3.0 - 1984 (Chris Whittington)
Super Chess III 3.1 - 1985 (Chris Whittington)
Super Chess III 3.5 - 1985 (Chris Whittington)
Clock Chess 89 - 1989 (Chris Whittington)

Double rounds = 132 games

With Cyrus is Chess you would have to use infinite mode and manually stop each move as longest level is level 8 which is 180 seconds per move.

So for this tournament I would expect that the dedicated computers all come out on top!

Tournament 2

Same programs but this time use 960 seconds per move for the Spectrum +3 Emulator with 3200% speed increase for real time and compare final tournament results against Test 1

Tournament 3

Same programs keep 960 seconds and 800% speed increase for Spectrum +3 Emulator and dedicated computers play at 2 minutes per move!

This setting would give dedicated decent thinking time and keeps the hardware comparison equal hence improve dedicated computer game quality.

Tournament 1 and 2 results could be accurately matched against Schachcomputer.Info active rating list. Tournament 3 is a little trickier as the equivalent of 2 minutes per move is played instead of 3, therefore not exactly comparable against the Tournament rating list.

Best regards

[microhenri]

Hi Nick,
There are more interesting chess programs for the ZX spectrum

What do you think about:
- Chess - The Turk
- Chess (Sinclair Research Ltd)(Psion)
- Chess Master
- The chessmaster 2000
- Chess 3D

I have even more.
Regards, microHenri

[spacious_mind]

Hi Nick,
There are more interesting chess programs for the ZX spectrum

What do you think about:
- Chess - The Turk
- Chess (Sinclair Research Ltd)(Psion)
- Chess Master
- The chessmaster 2000
- Chess 3D

I have even more.
Regards, microHenri

Hi Henri,

Yes there are many chess games for Spectrum. Most of them are quite weak though. From your above list CM2000 is the only one I would consider including. The others are not that strong.

Also I forgot about Cyrus II which I would include as well.

BTW Brian Whitby did something for Spectrum called Chess Diary, but it does not exist anywhere now.

Perhaps Brian could tell us what Chess Diary was all about.

Regards

[microhenri]

Then probably you know this link, but maybe interesting for others:

http://www.worldofspectrum.org/infoseek ... loadpics=0[/url]

[spacious_mind]

Yes I know the site.

It has a couple missing but otherwise it is mostly complete. Here for example Andromeda Chess:



Best regards

[microhenri]

Here for example Andromeda Chess:

Found it here
https://www.emuparadise.me/ZX_Spectrum_ ... eda)/74314

I am playing it on my cellphone

Regards, microHenri

[BenRedic]

It has a couple missing but otherwise it is mostly complete. Here for example Andromeda Chess:



Best regards

In playable form here:

https://archive.org/details/zx_Andromed ... _Andromeda

Looks like a clone of Mikrogen Masterchess, also known as ZX Spectrum Chess. Only board colors are different. Interface is the same, and it makes the same moves.

http://www.spacious-mind.com/html/spect ... chess.html

[Bryan Whitby]

Hi Nick,
There are more interesting chess programs for the ZX spectrum

What do you think about:
- Chess - The Turk
- Chess (Sinclair Research Ltd)(Psion)
- Chess Master
- The chessmaster 2000
- Chess 3D

I have even more.
Regards, microHenri

Hi Henri,

Yes there are many chess games for Spectrum. Most of them are quite weak though. From your above list CM2000 is the only one I would consider including. The others are not that strong.

Also I forgot about Cyrus II which I would include as well.

BTW Brian Whitby did something for Spectrum called Chess Diary, but it does not exist anywhere now.

Perhaps Brian could tell us what Chess Diary was all about.

Regards

Hi Nick

My Spectrum chess program was actually called 'The Chess Data System' which was a graphical chess database, similar to ChessBase that enabled me to keep a computerised list of all my postal chess games. Back then I played in British, European and World postal tournaments.

In those days moves where sent either by postcard or scoresheet using a window envelope. In the British games after producing The Chess Data System, when a new game started I would ask my opponent if he had a Sinclair Spectrum and if they had then I would ask them if they would like to play our game on my Chess Data System via a cassette tape instead of the usual paper methods.

This is how we played our game. Playing white I would load a fresh version of my 'CDS' onto a blank cassette tape, make my first move and save the game. I would then post the cassette to my opponent who would load the tape onto his Spectrum and my move would appear on his TV screen or monitor. He would then make his move, save the game and post the cassette back to me and and so it went on posting the cassette back and too to one another.

Every one I played with my 'CDS' really enjoyed the novelty of playing postal games like this as they had never played like this before, plus they also got a free copy of my chess program.

Sir Clive Regards
Bryan

[spacious_mind]

Every one I played with my 'CDS' really enjoyed the novelty of playing postal games like this as they had never played like this before, plus they also got a free copy of my chess program.

Sir Clive Regards
Bryan

Hi Bryan,

So what was the name your chess program. Do you still have it?

Regards
Nick

[Bryan Whitby]

Every one I played with my 'CDS' really enjoyed the novelty of playing postal games like this as they had never played like this before, plus they also got a free copy of my chess program.

Sir Clive Regards
Bryan

Hi Bryan,

So what was the name your chess program. Do you still have it?

Regards
Nick

Hi Nick

The chess program I was referring to was my Chess Data System that I wrote about, not an actual chess engine. Sorry for the confusion. Unfortunately I haven't kept a copy nor my other Spectrum program I advertise and sold which was called The Encyclopaedia of the FA Cup.

Regards
Bryan