Old chess computers nostalgia

[Rubinus]

[Event "test"] [Site "?"] [Date "2024.08.11"] [Round "1"] [White "Novag Constellation 3.6, level 7"] [Black "Mephisto MMI, level 6"] [Result "1-0"] [ECO "C78"] [PlyCount "71"] 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O {last theo.} b5 6. Bb3 Bb7 7. Re1 Bc5 8. c3 {For already playing under his own power without a book, he's doing surprisingly well.} d6 9. d4 Bb6 10. Bg5 h6 11. Bh4 g5 {last theo. One of the longest variants MMI has in the book.} 12. Bg3 Qe7 $6 {0,80} (12... O-O 13. a4 exd4 14. cxd4 Re8 15. d5 Na5 16. Bc2 Nc4 $11) 13. d5 $6 {They don't see that pawns on d4 can't be taken.} (13. Nbd2 {and is a losing mistake} exd4 $4 14. e5 $3 $18) 13... Na5 14. a4 (14. Bc2) 14... O-O {I was expecting him to take a bishop, but from the info he was considering 0-0-0 for quite a while.} ( 14... Nxb3) 15. axb5 (15. Ba2) 15... axb5 {0,54} 16. Nbd2 $2 {The Kittinger program doesn't seem to have much respect for the bishops.} (16. Bc2) 16... Kh7 $2 (16... Nxb3) 17. Bc2 Ba6 $2 {And we get to the phase where in a closed position the machines have trouble finding the objectives of the game, and so the weird moves phase begins. It can take a long time in some of those old machine fights.} (17... Nc4) 18. h4 $2 (18. Nd4 $18 {That's Stockfish's idea. I kind of humanly considered b4 to control c5 and not be left with a delayed pawn on b2 after Nc4.}) 18... Rg8 $2 {He's tempted by the bound Bg3.} (18... Nh5 19. Nf1 g4 20. N3h2 Rg8 21. Kh1 $16) 19. hxg5 Nd7 $2 {They see that you can't immediately take for Nd4, but this retreat is bad. Optically it looks good at the same time, a tempting c5 square, but no time.} (19... Nh5 20. Nf1 hxg5 21. Nd4 $1 $18) 20. Qe2 $2 (20. Nf1 hxg5 21. Nd4 $18) 20... Nc4 $2 {0,35} (20... b4 21. Qd1 bxc3 22. bxc3 Kh8 {to prevent Nd4 from converting to f5} 23. Ra2 Bb7 24. Nf1 hxg5 $14) 21. Nxc4 $2 (21. Bh4 $18) 21... bxc4 {On the surface it looks good for Black because Bg3 is uncovered due to the bind and White has a late b2 pawn, but the kingside position plays a role.} 22. Ba4 $4 ({Funny variant of Stockfish} 22. Bf4 hxg5 23. Nd4) 22... hxg5 $2 (22... Nc5) 23. Nd2 $2 (23. Bxd7 Qxd7 24. Nxe5 {White will figure out the motive eventually, but we'll wait.}) 23... Rh8 $2 {Black sees nothing. When I played with him 40 years ago, he often had trouble spotting very obvious threats.} (23... Kg7) 24. Bxd7 (24. Bc6 Ra7 25. Bxd7) 24... Qxd7 25. Qh5+ Kg8 {0,88 - pro srovnání, Stockfish cca 4,3} 26. Qxg5+ Kh7 ({There is no escape.} 26... Kf8 $2 27. Nxc4) 27. Nf3 c6 $2 (27... Rag8 28. Nxe5 dxe5 (28... Rxg5 29. Nxd7 Rxg3 30. Nxb6 Rd3 31. Rxa6 cxb6 32. Rxb6 $18) 29. Qh5+ Kg7 30. Bxe5+) 28. Qh5+ {He can't see it yet on the first try.} Kg8 29. Qg5+ Kh7 30. Nxe5 $1 dxe5 {2,83} 31. Bxe5 f6 32. Bxf6 Bxf2+ {And the rest of it is delaying the inevitable, just opening up the lane for a faster checkmate.} (32... Qf7 33. Qh4+ Kg6 34. Bxh8) 33. Kxf2 Qa7+ 34. Kf3 Qe3+ 35. Rxe3 Bb5 36. Rh1# 1-0
If the second game is interesting, I'll probably let more of them play. Constellation was quite surprised how he managed without a book. Even though there are a lot of mistakes due to the limited calculation, it was fun to watch.
Novag Constellation 3.6 - spacious mind - He started the 1985 Prague Open with a 4.5/11. Better than he deserved, but anti-computer strategy was not well known at the time.

[Rubinus]

[Event "test"] [Site "?"] [Date "2024.08.13"] [Round "2"] [White "Mephisto MMI, level 6"] [Black "Novag Constellation 3.6, level 7"] [Result "0-1"] [ECO "E51"] [PlyCount "74"] 1. d4 Nf6 2. c4 e6 3. Nc3 Bb4 {last theo} 4. Bd2 O-O 5. e3 d5 6. Nf3 c5 7. a3 Bxc3 8. Bxc3 Ne4 9. Rc1 {last theo} Nd7 10. Bd3 {0,11} (10. cxd5 exd5 11. dxc5) 10... Ndf6 11. O-O b6 {Even without the book Novag builds the figures quite reasonably. But he probably has less respect for the bishops than it deserves.} (11... Nxc3) 12. b4 $5 (12. dxc5 Nxc5 13. Be2 dxc4 14. Bxc4 $14) 12... cxd4 $2 (12... dxc4 13. Bxc4 Nxc3 14. Rxc3 cxb4 15. axb4 a6 $11) 13. Bxd4 Bd7 $2 {The position was already difficult, but this unnatural move among stronger players would have deserved to lose.} (13... dxc4) 14. cxd5 $6 {0,42} (14. c5 $18) 14... exd5 15. Qb3 $6 (15. Ne5) 15... Rc8 16. Be2 Re8 17. Rfd1 $6 {The old machine is unable to execute a longer term plan, it just tries to optimize rating gains over a few moves. And so this kind of nanny-like shifting of pieces sometimes occurs.} (17. h3) 17... Rxc1 (17... Ng4) 18. Rxc1 Bg4 19. Qa4 {0,45} Qe7 20. Qb5 $6 (20. h3) 20... Bd7 $2 {It is obvious that Black sees his position as weak and would accept a repetition of moves.} (20... Rd8) 21. Qa6 Bg4 22. Qd3 (22. Bb5 Bd7 23. Bd3 $18) 22... Bh5 23. Qb5 Rd8 24. h3 $6 {0,33} ( 24. Ne5) 24... Bxf3 25. Bxf3 Nd2 26. Be2 Nfe4 {White has almost given up his advantage, but Black's exit is just a harmless little cloud in the sky.} 27. Qa4 $2 {This is a mess, though, he's taken the Queen far from the scene.} Qh4 $11 {White's fundamental problem is that he must accept the loss of a pawn and find compensation. But this is beyond his power. And if he doesn't reveal the weakness of the king because of the possibility of inserting intermediate moves, it will be a quick and actually quite surprising ending. A small cloud becomes a tornado.} 28. g3 $4 {He loses this game immediately. It's hard to say, perhaps he didn't see it going to checkmate, after all there are enough half-moves to count.} (28. Kh2 Qxf2 29. Qd7 $1 $11) 28... Qxh3 $19 29. Qxa7 ( 29. Be5 Nxf2 30. Kxf2 Qh2+ 31. Ke1 Ne4 32. Kd1 Qg1+ 33. Kc2 Rc8+ $19) 29... Ng5 {And the rest of the lot is just delaying the inevitable.} 30. f4 {-1,00 The rating alone suggests that White sees nothing... Stockfish shows -#8...} Qxg3+ 31. Kh1 Nge4 32. Qxf7+ {:)} Kxf7 33. Rc7+ Kf8 34. Rf7+ Kxf7 35. Bh5+ Ke7 36. Bc5+ bxc5 37. bxc5 Nf2# 0-1
Novag is stronger, but by how much? This looks pretty clear so far, once the king is played for, Mephisto doesn't stand a chance because he doesn't see the threats, or too late. On the other hand, Novag is toying with those bishops, if Mephisto can keep his composure, he probably has a chance to overwhelm his opponent. Unfortunately for the SSDF, that old Mephisto MMI are missing. I'm curious, so I'm going to extend the game for two more games. Unlike the Lexibook fight, it's a little more watchable....

[Rubinus]

[Event "?"] [Site "?"] [Date "2024.08.13"] [Round "3"] [White "Mephisto MMI, level 6"] [Black "Novag Constellation 3.6, level 7"] [Result "0-1"] [ECO "D03"] [PlyCount "94"] 1. Nf3 Nf6 2. d4 d5 3. Bg5 {last theo} g6 4. e3 Ne4 {last theo} 5. Bb5+ $2 {A very childish idea with a waste of time.} (5. Bh4) 5... c6 6. Be2 Bg7 $6 {One can understand not wanting to give up a centralized slider, but taking gave the biggest advantage. The opponent doesn't get it either, so he just delays a move.} (6... Nxg5) 7. O-O $6 (7. Bf4) 7... Nxg5 8. Nxg5 Nd7 {I thought I played e5 immediately} 9. Qd2 $6 ({The best is supposed to be Nf3, I just thought} 9. f4 e5 {Oh, it doesn't hinder anything and only weakens the position.}) 9... O-O (9... e5) 10. c4 {0,14} e5 11. Nf3 e4 12. Ne1 Nf6 {The recap is clear. The opening is won by Black and stands as if it started with White. But there's not that much going on yet.} 13. Qb4 $6 (13. cxd5) 13... Qc7 $6 {The beginning of the turnaround. From now on, Mephisto's game will feel a little more confident.} (13... a5) 14. Nc3 Bf5 (14... a5) 15. Qb3 {-0,16} Qd7 $6 (15... dxc4) 16. Nc2 $2 (16. cxd5 cxd5 17. g3 $17) 16... Rac8 $2 (16... dxc4 17. Qxc4 b5 18. Qc5 Rfc8 $19) 17. Rac1 $2 Rc7 $2 {Neither wants to start, both holding the tension because they can't calculate the consequences correctly.} 18. Rfd1 $2 Rcc8 $2 {A human player would set a plan to improve his position and try to implement it. This cannot be expected with computers with a depth of calculation of only a few halfmoves. They just don't see how to improve their position and the piece-shuffling phase ensues. So from the eye, White's play looked better to me, so I thought he eventually leveled the game at least. } 19. h3 $6 Be6 $6 20. Nb4 {-0,16} Qd8 21. Rd2 {Raising the rooks would make sense, but you have to watch the retreat of the knight.} Bh6 {It seems that he considered f3 to be the biggest threat.} 22. a4 Qd6 $2 (22... a6) 23. Nc2 Qc7 { At this stage it was fun to watch the roulette ball fall on even or odd and not have a bet. With these opponents, it can go either way. Fortunately, things are already starting to happen.} 24. cxd5 cxd5 25. Nb4 $6 {0,05 The queen behind the knight will be uncovered, so loss of tempo.} (25. Na3 Rb8 26. Rdc2 a6 27. Ncb5 Qb6 28. a5 Qxa5 29. Nc7 Rfd8 30. Rc5 b5 31. Nxe6 fxe6 32. Rc6 Rb6 $11) 25... Qb6 26. Bd1 (26. Qa3) 26... Rfd8 (26... Rb8) 27. Bc2 $2 {And this is where Mephisto loses the thread again. And it's gonna get worse!} (27. Nba2 $11) 27... Rc4 $6 {I liked the move, but Stockfish shows that it could have been a lot better.} (27... Rd6 28. Bd1 Bf8 $19) 28. Nbxd5 $4 {And the party's over again. Most likely a miscalculation in the search for peace. Maybe, like the sample Minimax from the book Chess on PC, he has limited it to only a few half-moves and so hasn't calculated that it won't be two pawns per quality, but a whole rook.} Bxd5 29. Nxd5 (29. Qxb6 axb6 {It wouldn't have changed the outcome, he would have played longer.}) 29... Qxb3 30. Nxf6+ {He obviously understands now what he's done... Not as brutal, but it happens to me too. } Kg7 {That's just it, hopefully he didn't get to the difference between Kh8 and Kg7 in his calculation...} 31. Nh5+ gxh5 32. Bxb3 Rxc1+ 33. Kh2 h4 {But this deserves praise! White has a rook less, and the one he has left has nowhere to play. And there's going to be trouble with the king. An unintended consequence of creating at least a double pawn for the opponent - but also a free column.} 34. d5 {-3,16} Kf6 35. Rd4 Ke5 {I was expecting the Bf8-d6 maneuver} 36. Rd2 f5 37. Re2 Kd6 38. f3 f4 $1 39. fxe4 fxe3 40. g4 Rf8 41. Kg2 Rff1 42. g5 Rg1+ 43. Kh2 Rh1+ 44. Kg2 Rcg1+ 45. Kf3 Rg3+ 46. Kf4 Rf1+ 47. Rf2 Rxf2# {I wanted to switch it off after 33.-h4, but I was curious about the implementation. I'm glad for the quick ending, it also could have been a nagging ending until somewhere in the 90th move before the win could be realized. There were computers that were only pushed to victory by the triple repetition rule and 50 moves. And those that couldn't recognize triple repetition, like the already demonstrated Lexibook, sometimes tied with triple repetition with an extra rook...} 0-1
One of the biggest mistakes I've seen MMI play. Usually - with humans - he lost by eating what he didn't have. But a miscalculation like that? Quite disappointing. If MMI doesn't win the next game, they will move on to another opponent. When I progress to the MMII tests, Novag will run into the "stronger brother".
I'll continue on Friday and think about the next opponent. For example, I have SciSys Kasparov Turbo 16K in my collection, but that will probably be weaker than MMI. Only Fidelity was able to face H+G (Mephisto) at that time, the others were about a year or two behind. I don't have anything that weak from Fidelity, yet. My great uncle had a Chess Challenger, a version from maybe '77, and it was no match for the MMI. 6 years is almost an eternity between computers.

[Vinvin]

Unfortunately for the SSDF, that old Mephisto MMI are missing. I'm curious, so I'm going to extend the game for two more games. Unlike the Lexibook fight, it's a little more watchable....

2 rating lists with a lot of old dedicated machines :
https://www.schach-computer.info/wiki/i ... -Elo-Liste
and
https://www.schachcomputer.info/html/ak ... liste.html

[Rubinus]

[Event "test"] [Site "?"] [Date "2024.08.16"] [Round "4"] [White "Novag Constellation 3.6, level 7"] [Black "Mephisto MMI, level 6"] [Result "0-1"] [ECO "D38"] [PlyCount "106"] 1. c4 {last theo} e6 2. Nf3 Nf6 3. Nc3 Bb4 4. d4 d5 5. cxd5 {A clear disadvantage of the book according to the embedded variants. Even if the computer returns to a book position under its own power, it is no longer able to use the knowledge and must play under its own power anyway.} ({If MMI was playing against itself, it would be playing} 5. Bg5 h6 6. Bxf6 Bxc3+ {- the most common position that appears in his automatic games.}) 5... exd5 {last theo} 6. Bf4 Nc6 {-0,42} 7. e3 O-O 8. Bb5 (8. Bd3) 8... Ne4 9. Qc1 {Strange, he obviously didn't like the weave after Bg4.} (9. Rc1) 9... Bg4 10. Bxc6 bxc6 {-0,34} 11. Ne5 Bd7 12. O-O Nxc3 13. bxc3 Bd6 14. Rb1 $6 (14. c4 $16) 14... Bf5 15. Nxc6 $2 {A mistake that will cause defeat. It's a position evaluation problem with old engines. 5 for a rook, 3-3.5 for a light piece, if they got pawns to do it, they were easily tempted to overly optimistic sacrifices of quality. Even Lang's programs, whether Genius or Mephisto machines, suffered from this.} (15. Ra1 $11) 15... Qd7 {-0,78} 16. Ne5 Bxe5 17. Bxe5 f6 {I wouldn't play that, but it doesn't spoil anything.} 18. Bf4 Bxb1 19. Qxb1 Qc6 $6 {Stockfish doesn't like it, but I agreed with the lot. I might play it that way too. So Black secures the b column and starts to stifle White by pushing the a pawn advance.} (19... a5 $19) 20. Qb4 (20. Qc2) 20... Rab8 21. Qa3 Rb7 22. Bg3 {He doesn't know what to play. That's the disadvantage of a computer, it doesn't know who it's playing with. You'd have to adjust your plan according to your opponent. Here it would be clear, the only way to beat or at least somehow derail MMI is to attack the king by creating weaknesses around him. In fact, in that aforementioned 1984 match, I saved two games with him by perpetual check without overlooking quality.} Rfb8 {I would have inserted Kf7 to avoid being bothered by the queen, but it doesn't seem to matter.} 23. Bf4 { Sometimes a losing position can be held by doing nothing. But that's not the case here.} (23. h4) 23... a6 $6 (23... Rb1 {and checkmate cannot be taken} 24. Qxa7 Rxf1+ 25. Kxf1 Qc4+ {#-4}) 24. Qe7 {The Queen alone can do nothing there. In addition, Black could even banish her by inserting Re8 and Kf7.} (24. h4) 24... Qc4 $6 (24... g5 25. Bxg5 fxg5 26. Qxg5+ {Stockfish would certainly avoid perpetual check or position repetition, I wouldn't be sure about MMI with his calculation, a move in game is probably safer from his point of view.} ) 25. Qe6+ Kh8 26. Qf5 (26. g4 {Well, what did I write a moment ago?}) 26... Qxa2 {-2,06 5 Steps to Pawn Glory a6...} 27. Qd3 a5 28. e4 Kg8 $6 (28... Rb1) 29. exd5 Qxd5 30. Re1 (30. c4) 30... Qc6 $6 {-2,00 And again, a move that Stockfish says is not the best, but I quite like it - it supports the next pawn advance and can possibly exchange or drive away the white rook via e8. Less pieces, less worries, say many players here.} (30... a4) 31. d5 Qd7 (31... Re8 32. Ra1 Qb5) 32. d6 {If it weren't for the free pawn, White could still hope for some strength, but of course that won't happen. And MMI hasn't overlooked anything trivial this time.} cxd6 33. Bxd6 (33. c4) 33... Rd8 34. Qc4+ Kh8 35. Bf4 a4 {-2,37 Wednesday, a well-known cafe player would announce - the transformation ends with a free weekend...} 36. Qe2 a3 37. Qa6 Qd1 38. Kf1 Qb3 39. Qe2 Qa4 40. Bc1 {An act of desperation, but MMI won't overlook that last line} a2 {-7,12} 41. Bb2 Rdb8 42. Ba1 Rb1 43. Qd3 Qe8 44. Qe3 Qb5+ 45. Kg1 Re8 46. h3 Rxe3 47. Kh2 Rexe1 48. Kg3 Rxa1 {He does not see the checkmate and so he consumes. But it doesn't matter, chess is not played on the score of the number of moves.} 49. c4 Qg5+ 50. Kf3 Qf5+ 51. Kg3 Re4 52. f3 Qe5+ 53. f4 Qxf4# 0-1
So now we see how MMI can win. Novag beats itself with an optimistic sacrifice of material. All in all, I quite commend Mephisto MMI for that realization of superiority, though of course there were gaps. He played quite logically and if he didn't overlook anything, there was nothing to worry about.
At least two more games will be played, one today and one probably on Monday. I'm busy next week, then it will be played only on Thursday and Sunday.

[Rubinus]

[Event "test"] [Site "?"] [Date "2024.08.19"] [Round "5"] [White "Novag Constellation 3.6, level 7"] [Black "Mephisto MMI, level 6"] [Result "1/2-1/2"] [ECO "B40"] [PlyCount "85"] 1. e4 c5 {25% for MMI, otherwise only e5} 2. Nf3 e6 3. d4 cxd4 4. Nxd4 {last theo} Nf6 5. Bd3 Nc6 6. Nxc6 bxc6 {last theo} 7. Bf4 {Interesting, Stockfish would play it too, at least for a short time.} d5 {0,25} 8. e5 $6 (8. Nd2) 8... Nd7 9. O-O Qb6 (9... g5 10. Bg3 h5 11. h3 g4 $15 {and when} 12. hxg4 Qg5) 10. Qh5 $6 {a waste of time, he's gonna go somewhere else anyway.} (10. Nd2 Qxb2 11. c4) 10... g6 {-0,48} 11. Qe2 Qxb2 12. Nd2 {Interesting position. What's white's pawn? Only piece development, but that's a temporary advantage. But it needs to be turned into a permanent advantage by active play. Can a machine with a few moves calculate that?} Bg7 13. Nf3 O-O 14. Rfb1 (14. h4 $11) 14... Qa3 15. Bd2 {He slowly closes the net around the black queen. Black, with his calculations, clearly can't see it.} c5 $2 {-0,71} (15... a5) (15... Qe7) 16. Bb5 {He slowly closes the net around the black queen. Black, with his calculations, clearly can't see it.} a6 $8 17. Rb3 axb5 18. Rxa3 Rxa3 19. Qxb5 Nxe5 $2 {Watching the game, I estimated the position to be won for White, but according to Stockfish, the mistake is here. Correctly, the queen should have dismantled the unplayed pieces, but the machine would have to be able to see quite a few moves further.} (19... Ra7 $11) 20. Nxe5 $18 Bxe5 {0,11 - Stockfish 2,15} 21. Qxc5 Ra8 {Black needs to build some sort of fortress to hold back the advance of the free "a" pawn. And he also has weaknesses around the king.} 22. Bc3 $2 {Strategic mistake. Given the weaknesses of h6 and f6, a black bishop would be useful for White, as it eventually offers a plan of attack for the king as h4-h5.} (22. Bh6 Rd8 23. Rb1 Bf6 (23... Rxa2 $4 24. Qe7) 24. h4 e5 (24... Bxh4 25. Qd4) 25. Rb6 $18) 22... Bxc3 $16 23. Qxc3 Bd7 24. a3 (24. h4) 24... Ba4 25. Qd2 Rfb8 $6 (25... Rfc8) 26. c3 (26. g4 Rc8 27. h4 { Roughly along the lines of what I wrote above.}) 26... Rb3 {-0,83 The MMI rating is completely off...} 27. h3 $6 {Well, if you don't play hard enough, the advantage gradually dissolves. And if it's in compensation for another disadvantage, it can easily turn a won game into a lost one. But we're not there yet.} (27. Qd4) 27... Rc8 28. c4 $2 {He does not see his chances in weakening the fields around the king correctly. But now it is clear that the advantage has turned to Black's side, although holding is still possible. Play to win with a comparable opponent, perhaps even for both sides, however, too.} (28. Qd4 Rc4 29. Qf6 Rcxc3 30. Qd8+ Kg7 31. Qd6 Rc4) 28... dxc4 {Oddly enough, right. My guess for the game was that I would take the rook and hold the fort.} (28... Rxc4 $2 29. Rc1 Rd3 30. Qxd3 Rxc1+ 31. Kh2 Rc4 {The fortress is there, but Black has no active chances, and he would still have to watch to make sure that he doesn't finish off those black squares if White somehow gets the king on h6, for example.}) 29. Qa5 Be8 {Suddenly a black free pawn is at least as dangerous as a white one. But the queen can hold it by attacking pieces.} ( 29... Bc6) 30. Rc1 c3 31. Qa6 Bd7 {-1,08} 32. Qa7 Bb5 33. a4 Bd3 34. Qd7 Ba6 35. Qd6 Be2 36. Qg3 Ra3 {-1,02} 37. Qe5 {Which figures would you prefer to play as? Black must guard the queen's double attacks, while White must hold the c-pawn with all his might.} Ba6 $2 {It was clear that some sort of mistake was bound to come if both veterans were only counting a few moves ahead.} ( 37... Bc4 38. Rxc3 Rxa4 39. Qd4 e5 40. Qxe5 Rca8 $11 {Seeing Stockfish's variant, I understand that Black didn't want to change the free pawns, because his was two moves away!}) 38. Qa5 Rc6 39. Qd8+ Kg7 40. Qd4+ {Right up to here. But he can't find a way to win against these unplayed pieces.} Kg8 41. Qd8+ $2 {The question mark doesn't really belong there yet, because he could still repeat it once, and if it was a game with a time bonus per turn, that could come in handy.} Kg7 42. Qd4+ Kg8 43. Qd8+ ({One can enjoy the Stockfish variant, where he discards the black free pawn and keeps his own. It would still be tough, but somehow winnable for white it certainly is.} 43. Qb4 $1 Ra2 44. Qb8+ Kg7 45. Qb3 Rb2 46. Qa3 e5 47. Rxc3 Rb1+ 48. Kh2 Rbb6 49. Qe7 Re6 50. Qc5) 1/2-1/2
Saturday, as planned, was played, but I cancelled both attempts. First the game was played exactly the same as game 4 above, so Novag would have lost once more, and in the second game the Spanish with b5 from game 1 was repeated there too until somewhere around move 19. So he would have lost to MMI, as well. These don't have self-learning, the opening books are quite small, so there aren't really many variations of unique games they can play against each other. 1.c4 e6 is exhausted because Novag plays with his own power and MMI has a single line and the Spanish one probably does too, there's nothing else MMI plays with Black, unless Novag also plays the exchange 4.Bxc6, I don't know, I played a single game with him when I bought him sometime before last year. I don't care for doubles, so if there's any threat of a repeat, I'm cancelling the game. There's one game left, probably on Thursday, when MMI will play White, I hope he plays 1.e4 - about 25% - and then I don't think I'll try the two against each other, since the variations keep repeating.

From that conclusion, would you find a win or would you also rather draw?
White on move.

8
7
6
5
4
3
2
1
	a	b	c	d	e	f	g	h

6k1/5p1p/b1r1p1p1/8/P2Q4/r1p4P/5PP1/2R3K1 w - - 0 43

[Jouni]

Really difficult. Engines show +3 for Qb4 or Qd8, but don't show any winning line.

[Rebel]

Unfortunately for the SSDF, that old Mephisto MMI are missing. I'm curious, so I'm going to extend the game for two more games. Unlike the Lexibook fight, it's a little more watchable....

2 rating lists with a lot of old dedicated machines :
https://www.schach-computer.info/wiki/i ... -Elo-Liste
and
https://www.schachcomputer.info/html/ak ... liste.html

Rebel Portoroz 100 MHz

Never heard of that one.

[Ajedrecista]

Hello Ed:

Rebel Portoroz 100 MHz

Never heard of that one.

It might be the Rebel version that played at the World Microcomputer Chess Championship of 1989 at the city of Portoroz, then in Yugoslavia.

Regards from Spain.

Ajedrecista.

[Rebel]

Aha, 1989. I took a break for about a year, they (H&G) likely played with the MMV or Polgar.