Armageddon Opening set

[lkaufman]

I decided to create a set of ten Armageddon openings, so suitable for twenty game matches with Armageddon scoring (Black wins draws), with very demanding criteria. The opening lines run no more than 2 moves (i.e. 4 ply) long, White makes no suboptimal moves, all positions must have occurred at least 30 times in the current Hiarcs Opening Book, White must have won "about" half the games (maximum deviation was 11 games), and the Stockfish 15.1 eval after 40 or more plies is in the 0.80 to 1.20 range. Furthermore when the eval was outside the 0.90 to 1.10 range, the game stats must be in the opposite direction or at least very close to 50%. Finally, all Black moves must at least be "plausible", no obviously harmful moves, and the resulting positions must be ones that I felt would at least be deemed unclear by Grandmasters as to whether White should win or Black should draw. Assuming that I was successful in choosing positions that are truly near the win/draw line, Armageddon games to break ties could simply choose from the ten positions at random and assign color randomly. It is my intention that this set should work equally well for strong human players or for engines, and that it can be used not only for breaking ties but for determining quickly (with twenty games) whether engine X is stronger than engine Y (unless they are very closely matched).

Here are the ten openings with short comments:

1.e4 b6? Owen's defense, has the lowest eval, 0.84, but White won 51% of 577 games, the most popular of the ten.
1.d4 b5? Polish defense, has been played by strong players, but the eval at 0.96 is near the 1.00 threshold.
1.e4 e5 2.Nf3 d5?. "Elephant Gambit". Popular at amateur level; Black does get a little compensation for the pawn.
1.e4 c6 2.d4 Qc7?. Not a "good" move, but has some point. Eval and results right around the win/draw line.
1.e4 c6 2.d4 Na6?. Intending ...Nc7 next. Perhaps a tad better for White than the previous line.
1.e4 e5 2.Nf3 Qe7?. Intending ...g6 and Bg7. Eval above 1 but White wins below 50%.
1.d4 c5 2.d5 e5?. Looks quite playable for Black, but eval and results are both above threshold for White!
1.e4 e6 2.d4 Nc6?. Blocking c7 pawn here is dubious, but not silly. Eval just below threshold, results just above.
1.e4 g6 2.d4 Nf6?. Norwegian defense, played by Magnus Carlsen, but dubious. High eval but below 50% W wins.
1.e4 d6 2.d4 f5?. Balogh defense. White results and eval both a bit above threshold.

I imagine that many players will claim that some of these are defensible for Black and that others are losing, but I think that most grandmasters would admit to being uncertain about most of them. It is also far from clear whether stronger play (more time or using strong engines) favors White or Black in this Armageddon scenario.

[Peter Berger]

How can 1.e4 g6 2. d4 Nf6 be much worse than the Alekhine? Or this 3. g6 Caro line? 3. e5 is a concession releasing control of the white squares to some extent ( and chasing the knight to a decent square) . Contrary to your other lines I think this one may just be OK.

[Peter Berger]

As I was curious I gave the "Norwegian line" to Stockfish and let it run for a very long time. The eval kept going down and it actually transposed to a 3...g6 Caro line of sorts eventually, below 1.0 eval.

[Ajedrecista]

Hello Larry:

I decided to create a set of ten Armageddon openings, so suitable for twenty game matches with Armageddon scoring (Black wins draws), with very demanding criteria. The opening lines run no more than 2 moves (i.e. 4 ply) long, White makes no suboptimal moves, all positions must have occurred at least 30 times in the current Hiarcs Opening Book, White must have won "about" half the games (maximum deviation was 11 games), and the Stockfish 15.1 eval after 40 or more plies is in the 0.80 to 1.20 range. Furthermore when the eval was outside the 0.90 to 1.10 range, the game stats must be in the opposite direction or at least very close to 50%. Finally, all Black moves must at least be "plausible", no obviously harmful moves, and the resulting positions must be ones that I felt would at least be deemed unclear by Grandmasters as to whether White should win or Black should draw. Assuming that I was successful in choosing positions that are truly near the win/draw line, Armageddon games to break ties could simply choose from the ten positions at random and assign color randomly. It is my intention that this set should work equally well for strong human players or for engines, and that it can be used not only for breaking ties but for determining quickly (with twenty games) whether engine X is stronger than engine Y (unless they are very closely matched).

Here are the ten openings with short comments:

1.e4 b6? Owen's defense, has the lowest eval, 0.84, but White won 51% of 577 games, the most popular of the ten.
1.d4 b5? Polish defense, has been played by strong players, but the eval at 0.96 is near the 1.00 threshold.
1.e4 e5 2.Nf3 d5?. "Elephant Gambit". Popular at amateur level; Black does get a little compensation for the pawn.
1.e4 c6 2.d4 Qc7?. Not a "good" move, but has some point. Eval and results right around the win/draw line.
1.e4 c6 2.d4 Na6?. Intending ...Nc7 next. Perhaps a tad better for White than the previous line.
1.e4 e5 2.Nf3 Qe7?. Intending ...g6 and Bg7. Eval above 1 but White wins below 50%.
1.d4 c5 2.d5 e5?. Looks quite playable for Black, but eval and results are both above threshold for White!
1.e4 e6 2.d4 Nc6?. Blocking c7 pawn here is dubious, but not silly. Eval just below threshold, results just above.
1.e4 g6 2.d4 Nf6?. Norwegian defense, played by Magnus Carlsen, but dubious. High eval but below 50% W wins.
1.e4 d6 2.d4 f5?. Balogh defense. White results and eval both a bit above threshold.

I imagine that many players will claim that some of these are defensible for Black and that others are losing, but I think that most grandmasters would admit to being uncertain about most of them. It is also far from clear whether stronger play (more time or using strong engines) favors White or Black in this Armageddon scenario.

Please take a look to the thread of a Graham's tournament featuring gambits:

Stockfish 16CPU v Komodo Dragon 16CPU (LTC Gambits Match)

And a post of mine in the same thread suggesting other gambits:

Re: Stockfish 16CPU v Komodo Dragon 16CPU (LTC Gambits Match).

May I complete a little your descriptions in some of your picked openings with games of GMs:

------------------------

1.e4 e5 2.Nf3 Qe7: the Gunderam Defence. GM Radjabov (2670 Elo) played it in a must-win situation to finally draw against GM Adams (2731 Elo) at FIDE WC Knockout Tournament of 2004:

https://www.chessgames.com/perl/chessgame?gid=1292900

[pgn][Event "FIDE World Championship Knockout Tournament"]
[Site "Tripoli LIB"]
[Date "2004.07.04"]
[EventDate "?"]
[Round "6.4"]
[Result "1/2-1/2"]
[White "Michael Adams"]
[Black "Teimour Radjabov"]
[ECO "C40"]
[WhiteElo "2731"]
[BlackElo "2670"]
[PlyCount "88"]

1. e4 e5 2. Nf3 Qe7 3. Nc3 c6 4. d4 d6 5. Bc4 Bg4 6. dxe5 dxe5
7. h3 Bh5 8. g4 Bg6 9. Bg5 f6 10. Be3 Nd7 11. Nh4 O-O-O
12. Qe2 Nb6 13. Bb3 Qc7 14. Bd2 Bc5 15. O-O-O Ne7 16. Nf5 Nxf5
17. exf5 Bf7 18. Ne4 Bxb3 19. axb3 Be7 20. Ba5 Rd5 21. Bxb6
axb6 22. Rd3 Rxd3 23. Qxd3 Rd8 24. Qe2 b5 25. Rd1 Rd4 26. c3
Rd5 27. b4 Qb6 28. Kb1 Kb8 29. h4 Qd8 30. Kc2 Kc7 31. h5 Qa8
32. Kb1 Qg8 33. f3 Qd8 34. Rd2 Qd7 35. Kc2 h6 36. Rd1 Kb8
37. Ra1 b6 38. Qe1 c5 39. bxc5 bxc5 40. Qe2 c4 41. Rd1 Kc7
42. Rxd5 Qxd5 43. Qe3 Kb7 44. Nd2 Kb8 1/2-1/2[/pgn]

------------------------

1.e4 b6: the Owen Defence. Radjabov again against three GM (Nepomniachtchi, Carlsen and Giri), with mixed results (one win and two loses):

[pgn][Event "Airthings Masters"]
[Site "chess24.com INT"]
[Date "2020.12.30"]
[EventDate "2020.12.29"]
[Round "1.32"]
[Result "0-1"]
[White "Ian Nepomniachtchi"]
[Black "Teimour Radjabov"]
[ECO "B00"]
[WhiteElo "2784"]
[BlackElo "2765"]
[PlyCount "102"]

1. e4 b6 2. d4 Bb7 3. Nc3 e6 4. Nf3 d6 5. a4 a6 6. Bd3 Nd7
7. O-O h6 8. Re1 Ngf6 9. h3 Be7 10. Bf4 g5 11. Bh2 g4 12. hxg4
Nxg4 13. d5 Rg8 14. dxe6 fxe6 15. Bc4 Rg6 16. Ne2 Nxh2
17. Nxh2 Bg5 18. Nd4 Nf8 19. Qh5 Qf6 20. e5 Qf7 21. exd6 Be3
22. d7+ Nxd7 23. Ng4 Bxf2+ 24. Kh1 Bxe1 25. Rxe1 O-O-O
26. Bxe6 Rxe6 27. Qxf7 Rxe1+ 28. Kh2 Re4 29. Qg7 h5 30. Nh6
Kb8 31. c3 Ne5 32. a5 bxa5 33. Ne6 Nf3+ 34. gxf3 Rxe6 35. Nf5
Re2+ 36. Kg3 Rxb2 37. Nd4 h4+ 38. Kxh4 Rb6 39. Kg3 a4 40. Qe7
Rdd6 41. Nf5 Rd3 42. Qf8+ Ka7 43. Qc5 Rd5 44. Qf8 a3 45. Ne7
Rg5+ 46. Kf4 Ra5 47. Nc8+ Bxc8 48. Qxc8 Rc6 49. Qe8 Rc4+
50. Kg3 a2 51. Qe3+ Rcc5 0-1

[Event "FTX Crypto Cup"]
[Site "chess24.com INT"]
[Date "2021.05.29"]
[EventDate "2021.05.23"]
[Round "2.24"]
[Result "1-0"]
[White "Magnus Carlsen"]
[Black "Teimour Radjabov"]
[ECO "B00"]
[WhiteElo "2847"]
[BlackElo "2765"]
[PlyCount "77"]

1. e4 b6 2. d4 Bb7 3. Bd3 e6 4. Nf3 d6 5. O-O Nd7 6. a4 a6
7. Re1 Ngf6 8. Nbd2 Be7 9. Nf1 O-O 10. Ng3 e5 11. c3 Re8
12. Bd2 Bf8 13. b4 g6 14. Qc1 Bg7 15. a5 b5 16. Bh6 Bh8 17. d5
c5 18. h3 c4 19. Bc2 Nf8 20. Qd2 Bc8 21. Nh2 Bg7 22. Rf1 Bxh6
23. Qxh6 Kh8 24. Qd2 N6d7 25. f4 exf4 26. Rxf4 Ne5 27. Nf3
Nfd7 28. Rf1 Nxf3+ 29. R4xf3 Ne5 30. Rf6 Bf5 31. exf5 Qxf6
32. fxg6 Qh4 33. Nf5 Qh5 34. g7+ Kg8 35. Nxd6 Red8 36. Nf5 f6
37. d6 Ra7 38. Qd5+ Rf7 39. Bd1 1-0

[Event "Meltwater Tour Final 2021"]
[Site "Chess24 INT"]
[Date "2021.09.27"]
[EventDate "2021.09.23"]
[Round "3.5"]
[Result "1-0"]
[White "Anish Giri"]
[Black "Teimour Radjabov"]
[ECO "B00"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "78"]

1. e4 b6 2. d4 Bb7 3. Bd3 e6 4. Nf3 d6 5. Nbd2 Nd7 6. Nf1 Ngf6 7. Ng3 Be7
8. O-O O-O 9. Re1 e5 10. c3 Re8 11. h3 Bf8 12. Qc2 g6 13. Be3 Qe7 14. Rad1
Bg7 15. Qc1 a6 16. a4 Qf8 17. Bg5 Kh8 18. b4 Rac8 19. b5 axb5 20. axb5 Ra8
21. Qc2 Bh6 22. Bc4 Bg7 23. Bc1 h6 24. h4 exd4 25. cxd4 d5 26. exd5 Rxe1+
27. Rxe1 Bxd5 28. Bxd5 Nxd5 29. h5 Qd6 30. Ne4 Qe6 31. Nc5 Qd6 32. Nxd7
Qxd7 33. hxg6 Re8 34. Ne5 Bxe5 35. dxe5 fxg6 36. Qxg6 Re6 37. Qf5 Qe8 38.
f4 Qg8 39. Qh3 1-0[/pgn]

------------------------

1.d4 b5: the Polish Defence. Seen at fast TC against two 2700+ GMs:

https://old.chesstempo.com/gamedb/game/4665107
https://lichess.org/kMc1czvH

------------------------

1.e4 e5 2.Nf3 d5: the Elephant Gambit, seen at 2010 World Blitz Championship between Movsesian and MVL:

https://www.chessgames.com/perl/chessgame?gid=1599979

[pgn][Event "World Blitz Championship"]
[Site "Moscow RUS"]
[Date "2010.11.18"]
[EventDate "2010.11.16"]
[Round "33"]
[Result "1-0"]
[White "Sergei Mushegovic Movsesian"]
[Black "Maxime Vachier-Lagrave"]
[ECO "C40"]
[WhiteElo "2721"]
[BlackElo "2703"]
[PlyCount "129"]

1. e4 e5 2. Nf3 d5 3. exd5 e4 4. Qe2 Nf6 5. Nc3 Be7 6. Nxe4
O-O 7. d3 Nxd5 8. Qd1 Nc6 9. Be2 Bf5 10. O-O Qd7 11. Bd2 Rfe8
12. c3 Rad8 13. Qc2 Nf6 14. Rad1 Qe6 15. Ng3 Bg6 16. a3 Bc5
17. b4 Bb6 18. Ng5 Qe5 19. Rfe1 h6 20. Nh3 Qd6 21. Bf4 Qd7
22. a4 a6 23. Qa2 Nd5 24. Bd2 Re7 25. d4 Rde8 26. Bc4 Rxe1+
27. Rxe1 Rxe1+ 28. Bxe1 Nce7 29. Ne2 Qf5 30. a5 Ba7 31. Qd2 c6
32. Nhf4 Bh7 33. Bd3 Qd7 34. Bxh7+ Kxh7 35. Nxd5 cxd5 36. Nf4
Bb8 37. Nd3 Bd6 38. f3 Nf5 39. Qe2 Qb5 40. Kf1 Bxh2 41. g4
Ng3+ 42. Bxg3 Bxg3 43. Nc5 Qc6 44. Qd3+ g6 45. Qe2 Bd6 46. Kg2
Kg7 47. Qe3 Qc8 48. Qe2 h5 49. gxh5 gxh5 50. Qd3 h4 51. Qe3
h3+ 52. Kh1 Qf5 53. Nd3 Kh7 54. Ne5 f6 55. Qd3 Qxd3 56. Nxd3
Kg6 57. Nf2 h2 58. Kg2 Kf5 59. Nd3 Kg5 60. Nc5 Bxc5 61. dxc5
Kf4 62. b5 d4 63. cxd4 axb5 64. c6 bxc6 65. a6 1-0[/pgn]

------------------------

1.e4 c6 2.d4 Na6: played by Carlsen (2017) and Nakamura (2019) at fast TC:

https://www.chessgames.com/perl/chessgame?gid=1890619
https://www.chessgames.com/perl/chessgame?gid=1962999

[pgn][Event "Chess.com Speed Chess Championship 2017/18"]
[Site "chess.com INT"]
[Date "2017.10.04"]
[EventDate "2017.05.03"]
[Round "1.19"]
[Result "0-1"]
[White "Gadir Guseinov"]
[Black "Magnus Carlsen"]
[ECO "A40"]
[WhiteElo "2645"]
[BlackElo "2832"]
[PlyCount "102"]

1. e4 c6 2. d4 Na6 3. Nc3 Nc7 4. Nf3 d5 5. Bd3 Bg4 6. O-O e6
7. Be3 Bd6 8. h3 Bh5 9. e5 Be7 10. Be2 Nh6 11. Bxh6 gxh6
12. Ne1 Bg6 13. f4 Qd7 14. Nd3 O-O-O 15. b4 Bxd3 16. Bxd3 Bxb4
17. Na4 c5 18. Nxc5 Bxc5 19. dxc5 Qc6 20. Qh5 Qxc5+ 21. Kh1
Rd7 22. Qxh6 Kb8 23. Qg7 Qf8 24. Qf6 h5 25. Rab1 Rg8 26. Qh4
Qg7 27. Rf2 Qg3 28. Qxg3 Rxg3 29. Be2 h4 30. f5 a6 31. Rf4 Re3
32. Bf3 Rxe5 33. fxe6 Nxe6 34. Rxh4 Rc7 35. Rd1 Rxc2 36. Bxd5
Ka7 37. Rh7 Rc7 38. Bxe6 Rxe6 39. h4 Re2 40. a3 Rcc2 41. Rg1
Rf2 42. Kh2 Rf3 43. Ra1 b5 44. Rg7 Kb6 45. a4 b4 46. a5+ Kb5
47. Rg5+ Kc4 48. Rg4+ Kc3 49. Ra4 b3 50. Rgc4+ Kb2 51. Rg4 Rf5
0-1

[Event "PRO League Group Stage"]
[Site "chess.com INT"]
[Date "2019.03.13"]
[EventDate "2019.01.08"]
[Round "10"]
[Result "1/2-1/2"]
[White "Wang Yue"]
[Black "Hikaru Nakamura"]
[ECO "A40"]
[WhiteElo "2684"]
[BlackElo "2749"]
[PlyCount "211"]

1. d4 Na6 2. e4 c6 3. Nc3 Nc7 4. Nf3 d5 5. Bd3 dxe4 6. Nxe4
Nf6 7. Nxf6+ exf6 8. O-O Bd6 9. Re1+ Be6 10. c4 O-O 11. Qc2 g6
12. Bh6 Re8 13. h3 Bf8 14. Bxf8 Kxf8 15. b4 Kg7 16. a4 b5
17. axb5 cxb5 18. cxb5 Nd5 19. Qd2 Qd6 20. Be4 Rab8 21. Bxd5
Bxd5 22. Rxe8 Rxe8 23. Rxa7 Bxf3 24. gxf3 Qd5 25. Kg2 Qxb5
26. Ra5 Qd7 27. d5 Qd6 28. Rc5 Re5 29. h4 Rf5 30. Qc3 Rf4
31. Rc4 Rf5 32. Rd4 Rxd5 33. Rxd5 Qxd5 34. Qb2 Qb5 35. Qd4 h5
36. f4 Qc6+ 37. Kg3 Qh1 38. f3 Qe1+ 39. Kh3 Qe2 40. Kg3 Qf1
41. Qd2 Qb5 42. Qb2 Kg8 43. Kg2 Kg7 44. Kg3 Kg8 45. Kg2 Qb8
46. b5 Qb6 47. Qb4 Kg7 48. Qc4 Qe3 49. Qb4 Qb6 50. Qb2 Kf8
51. Qb4+ Ke8 52. Qe4+ Kf8 53. Qb4+ Kg8 54. Qc4 Kg7 55. Qc6 Qe3
56. b6 Qd2+ 57. Kf1 Qd1+ 58. Kf2 Qd2+ 59. Kf1 Qd3+ 60. Kf2
Qd4+ 61. Ke2 Qb2+ 62. Ke3 Qa3+ 63. Kd4 Qa6 64. Kc5 Qa3+
65. Kb5 Qb3+ 66. Ka5 Qa3+ 67. Kb5 Qd3+ 68. Ka5 Qa3+ 69. Kb5
Qd3+ 70. Kb4 Qd2+ 71. Ka4 Qxf4+ 72. Ka5 Qe5+ 73. Qb5 Qa1+
74. Qa4 Qc3+ 75. Qb4 Qe5+ 76. Ka6 Qa1+ 77. Kb5 Qf1+ 78. Ka5
Qa1+ 79. Qa4 Qc3+ 80. Ka6 Qd3+ 81. Ka7 Qe3 82. Ka6 Qd3+
83. Qb5 Qa3+ 84. Qa5 Qd3+ 85. Ka7 Qe3 86. Ka8 Qe8+ 87. Ka7 Qe3
88. Qb5 g5 89. hxg5 fxg5 90. Qb2+ Kg8 91. Ka6 Qe6 92. Ka5 Qd5+
93. Qb5 Qa2+ 94. Qa4 Qd5+ 95. Qb5 Qa2+ 96. Qa4 Qd2+ 97. Ka6
Qd6 98. Ka7 Qc5 99. Ka6 Qd6 100. Qa5 f6 101. Ka7 Qd4 102. Ka6
Qd3+ 103. Ka7 Qd4 104. Ka6 Qd3+ 105. Ka7 Qd4 106. Ka6 1/2-1/2[/pgn]

------------------------

1.d4 c5 2.d5 e5: Caruana tried it in 2018 without success against Aronian at fast TC:

https://old.chesstempo.com/gamedb/game/4374562
https://lichess.org/iw3tDNiz

------------------------

1.e4 g6 2.d4 Nf6: played by Carlsen, as you stated, and also by a lot of 2700+ GM against strong opposition, which surprised me. Special mention to Rapport (2698 Elo), who played it at least four times against Grischuk (2750 Elo) in August, 2017. Source: Lichess opening explorer.

------------------------

Regards from Spain.

Ajedrecista.

[lkaufman]

As I was curious I gave the "Norwegian line" to Stockfish and let it run for a very long time. The eval kept going down and it actually transposed to a 3...g6 Caro line of sorts eventually, below 1.0 eval.

The online chessbase analysis by SF 15.1 at 51 plies shows (after 3.e5 Nh5 4.Nf3 d6) 5.Bc4 +1.23. Even if it does drop below 1 slightly at some depth, it is clearly too close to the win/draw line for anyone to know if it is won or drawn in theory. With the knight on the edge, this is obviously a bad position for Black, just like the other nine.

[lkaufman]

Hello Larry:

I decided to create a set of ten Armageddon openings, so suitable for twenty game matches with Armageddon scoring (Black wins draws), with very demanding criteria. The opening lines run no more than 2 moves (i.e. 4 ply) long, White makes no suboptimal moves, all positions must have occurred at least 30 times in the current Hiarcs Opening Book, White must have won "about" half the games (maximum deviation was 11 games), and the Stockfish 15.1 eval after 40 or more plies is in the 0.80 to 1.20 range. Furthermore when the eval was outside the 0.90 to 1.10 range, the game stats must be in the opposite direction or at least very close to 50%. Finally, all Black moves must at least be "plausible", no obviously harmful moves, and the resulting positions must be ones that I felt would at least be deemed unclear by Grandmasters as to whether White should win or Black should draw. Assuming that I was successful in choosing positions that are truly near the win/draw line, Armageddon games to break ties could simply choose from the ten positions at random and assign color randomly. It is my intention that this set should work equally well for strong human players or for engines, and that it can be used not only for breaking ties but for determining quickly (with twenty games) whether engine X is stronger than engine Y (unless they are very closely matched).

Here are the ten openings with short comments:

1.e4 b6? Owen's defense, has the lowest eval, 0.84, but White won 51% of 577 games, the most popular of the ten.
1.d4 b5? Polish defense, has been played by strong players, but the eval at 0.96 is near the 1.00 threshold.
1.e4 e5 2.Nf3 d5?. "Elephant Gambit". Popular at amateur level; Black does get a little compensation for the pawn.
1.e4 c6 2.d4 Qc7?. Not a "good" move, but has some point. Eval and results right around the win/draw line.
1.e4 c6 2.d4 Na6?. Intending ...Nc7 next. Perhaps a tad better for White than the previous line.
1.e4 e5 2.Nf3 Qe7?. Intending ...g6 and Bg7. Eval above 1 but White wins below 50%.
1.d4 c5 2.d5 e5?. Looks quite playable for Black, but eval and results are both above threshold for White!
1.e4 e6 2.d4 Nc6?. Blocking c7 pawn here is dubious, but not silly. Eval just below threshold, results just above.
1.e4 g6 2.d4 Nf6?. Norwegian defense, played by Magnus Carlsen, but dubious. High eval but below 50% W wins.
1.e4 d6 2.d4 f5?. Balogh defense. White results and eval both a bit above threshold.

I imagine that many players will claim that some of these are defensible for Black and that others are losing, but I think that most grandmasters would admit to being uncertain about most of them. It is also far from clear whether stronger play (more time or using strong engines) favors White or Black in this Armageddon scenario.

Please take a look to the thread of a Graham's tournament featuring gambits:

Stockfish 16CPU v Komodo Dragon 16CPU (LTC Gambits Match)

And a post of mine in the same thread suggesting other gambits:

Re: Stockfish 16CPU v Komodo Dragon 16CPU (LTC Gambits Match).

May I complete a little your descriptions in some of your picked openings with games of GMs:

------------------------

1.e4 e5 2.Nf3 Qe7: the Gunderam Defence. GM Radjabov (2670 Elo) played it in a must-win situation to finally draw against GM Adams (2731 Elo) at FIDE WC Knockout Tournament of 2004:

https://www.chessgames.com/perl/chessgame?gid=1292900

[pgn][Event "FIDE World Championship Knockout Tournament"]
[Site "Tripoli LIB"]
[Date "2004.07.04"]
[EventDate "?"]
[Round "6.4"]
[Result "1/2-1/2"]
[White "Michael Adams"]
[Black "Teimour Radjabov"]
[ECO "C40"]
[WhiteElo "2731"]
[BlackElo "2670"]
[PlyCount "88"]

1. e4 e5 2. Nf3 Qe7 3. Nc3 c6 4. d4 d6 5. Bc4 Bg4 6. dxe5 dxe5
7. h3 Bh5 8. g4 Bg6 9. Bg5 f6 10. Be3 Nd7 11. Nh4 O-O-O
12. Qe2 Nb6 13. Bb3 Qc7 14. Bd2 Bc5 15. O-O-O Ne7 16. Nf5 Nxf5
17. exf5 Bf7 18. Ne4 Bxb3 19. axb3 Be7 20. Ba5 Rd5 21. Bxb6
axb6 22. Rd3 Rxd3 23. Qxd3 Rd8 24. Qe2 b5 25. Rd1 Rd4 26. c3
Rd5 27. b4 Qb6 28. Kb1 Kb8 29. h4 Qd8 30. Kc2 Kc7 31. h5 Qa8
32. Kb1 Qg8 33. f3 Qd8 34. Rd2 Qd7 35. Kc2 h6 36. Rd1 Kb8
37. Ra1 b6 38. Qe1 c5 39. bxc5 bxc5 40. Qe2 c4 41. Rd1 Kc7
42. Rxd5 Qxd5 43. Qe3 Kb7 44. Nd2 Kb8 1/2-1/2[/pgn]

------------------------

1.e4 b6: the Owen Defence. Radjabov again against three GM (Nepomniachtchi, Carlsen and Giri), with mixed results (one win and two loses):

[pgn][Event "Airthings Masters"]
[Site "chess24.com INT"]
[Date "2020.12.30"]
[EventDate "2020.12.29"]
[Round "1.32"]
[Result "0-1"]
[White "Ian Nepomniachtchi"]
[Black "Teimour Radjabov"]
[ECO "B00"]
[WhiteElo "2784"]
[BlackElo "2765"]
[PlyCount "102"]

1. e4 b6 2. d4 Bb7 3. Nc3 e6 4. Nf3 d6 5. a4 a6 6. Bd3 Nd7
7. O-O h6 8. Re1 Ngf6 9. h3 Be7 10. Bf4 g5 11. Bh2 g4 12. hxg4
Nxg4 13. d5 Rg8 14. dxe6 fxe6 15. Bc4 Rg6 16. Ne2 Nxh2
17. Nxh2 Bg5 18. Nd4 Nf8 19. Qh5 Qf6 20. e5 Qf7 21. exd6 Be3
22. d7+ Nxd7 23. Ng4 Bxf2+ 24. Kh1 Bxe1 25. Rxe1 O-O-O
26. Bxe6 Rxe6 27. Qxf7 Rxe1+ 28. Kh2 Re4 29. Qg7 h5 30. Nh6
Kb8 31. c3 Ne5 32. a5 bxa5 33. Ne6 Nf3+ 34. gxf3 Rxe6 35. Nf5
Re2+ 36. Kg3 Rxb2 37. Nd4 h4+ 38. Kxh4 Rb6 39. Kg3 a4 40. Qe7
Rdd6 41. Nf5 Rd3 42. Qf8+ Ka7 43. Qc5 Rd5 44. Qf8 a3 45. Ne7
Rg5+ 46. Kf4 Ra5 47. Nc8+ Bxc8 48. Qxc8 Rc6 49. Qe8 Rc4+
50. Kg3 a2 51. Qe3+ Rcc5 0-1

[Event "FTX Crypto Cup"]
[Site "chess24.com INT"]
[Date "2021.05.29"]
[EventDate "2021.05.23"]
[Round "2.24"]
[Result "1-0"]
[White "Magnus Carlsen"]
[Black "Teimour Radjabov"]
[ECO "B00"]
[WhiteElo "2847"]
[BlackElo "2765"]
[PlyCount "77"]

1. e4 b6 2. d4 Bb7 3. Bd3 e6 4. Nf3 d6 5. O-O Nd7 6. a4 a6
7. Re1 Ngf6 8. Nbd2 Be7 9. Nf1 O-O 10. Ng3 e5 11. c3 Re8
12. Bd2 Bf8 13. b4 g6 14. Qc1 Bg7 15. a5 b5 16. Bh6 Bh8 17. d5
c5 18. h3 c4 19. Bc2 Nf8 20. Qd2 Bc8 21. Nh2 Bg7 22. Rf1 Bxh6
23. Qxh6 Kh8 24. Qd2 N6d7 25. f4 exf4 26. Rxf4 Ne5 27. Nf3
Nfd7 28. Rf1 Nxf3+ 29. R4xf3 Ne5 30. Rf6 Bf5 31. exf5 Qxf6
32. fxg6 Qh4 33. Nf5 Qh5 34. g7+ Kg8 35. Nxd6 Red8 36. Nf5 f6
37. d6 Ra7 38. Qd5+ Rf7 39. Bd1 1-0

[Event "Meltwater Tour Final 2021"]
[Site "Chess24 INT"]
[Date "2021.09.27"]
[EventDate "2021.09.23"]
[Round "3.5"]
[Result "1-0"]
[White "Anish Giri"]
[Black "Teimour Radjabov"]
[ECO "B00"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "78"]

1. e4 b6 2. d4 Bb7 3. Bd3 e6 4. Nf3 d6 5. Nbd2 Nd7 6. Nf1 Ngf6 7. Ng3 Be7
8. O-O O-O 9. Re1 e5 10. c3 Re8 11. h3 Bf8 12. Qc2 g6 13. Be3 Qe7 14. Rad1
Bg7 15. Qc1 a6 16. a4 Qf8 17. Bg5 Kh8 18. b4 Rac8 19. b5 axb5 20. axb5 Ra8
21. Qc2 Bh6 22. Bc4 Bg7 23. Bc1 h6 24. h4 exd4 25. cxd4 d5 26. exd5 Rxe1+
27. Rxe1 Bxd5 28. Bxd5 Nxd5 29. h5 Qd6 30. Ne4 Qe6 31. Nc5 Qd6 32. Nxd7
Qxd7 33. hxg6 Re8 34. Ne5 Bxe5 35. dxe5 fxg6 36. Qxg6 Re6 37. Qf5 Qe8 38.
f4 Qg8 39. Qh3 1-0[/pgn]

------------------------

1.d4 b5: the Polish Defence. Seen at fast TC against two 2700+ GMs:

https://old.chesstempo.com/gamedb/game/4665107
https://lichess.org/kMc1czvH

------------------------

1.e4 e5 2.Nf3 d5: the Elephant Gambit, seen at 2010 World Blitz Championship between Movsesian and MVL:

https://www.chessgames.com/perl/chessgame?gid=1599979

[pgn][Event "World Blitz Championship"]
[Site "Moscow RUS"]
[Date "2010.11.18"]
[EventDate "2010.11.16"]
[Round "33"]
[Result "1-0"]
[White "Sergei Mushegovic Movsesian"]
[Black "Maxime Vachier-Lagrave"]
[ECO "C40"]
[WhiteElo "2721"]
[BlackElo "2703"]
[PlyCount "129"]

1. e4 e5 2. Nf3 d5 3. exd5 e4 4. Qe2 Nf6 5. Nc3 Be7 6. Nxe4
O-O 7. d3 Nxd5 8. Qd1 Nc6 9. Be2 Bf5 10. O-O Qd7 11. Bd2 Rfe8
12. c3 Rad8 13. Qc2 Nf6 14. Rad1 Qe6 15. Ng3 Bg6 16. a3 Bc5
17. b4 Bb6 18. Ng5 Qe5 19. Rfe1 h6 20. Nh3 Qd6 21. Bf4 Qd7
22. a4 a6 23. Qa2 Nd5 24. Bd2 Re7 25. d4 Rde8 26. Bc4 Rxe1+
27. Rxe1 Rxe1+ 28. Bxe1 Nce7 29. Ne2 Qf5 30. a5 Ba7 31. Qd2 c6
32. Nhf4 Bh7 33. Bd3 Qd7 34. Bxh7+ Kxh7 35. Nxd5 cxd5 36. Nf4
Bb8 37. Nd3 Bd6 38. f3 Nf5 39. Qe2 Qb5 40. Kf1 Bxh2 41. g4
Ng3+ 42. Bxg3 Bxg3 43. Nc5 Qc6 44. Qd3+ g6 45. Qe2 Bd6 46. Kg2
Kg7 47. Qe3 Qc8 48. Qe2 h5 49. gxh5 gxh5 50. Qd3 h4 51. Qe3
h3+ 52. Kh1 Qf5 53. Nd3 Kh7 54. Ne5 f6 55. Qd3 Qxd3 56. Nxd3
Kg6 57. Nf2 h2 58. Kg2 Kf5 59. Nd3 Kg5 60. Nc5 Bxc5 61. dxc5
Kf4 62. b5 d4 63. cxd4 axb5 64. c6 bxc6 65. a6 1-0[/pgn]

------------------------

1.e4 c6 2.d4 Na6: played by Carlsen (2017) and Nakamura (2019) at fast TC:

https://www.chessgames.com/perl/chessgame?gid=1890619
https://www.chessgames.com/perl/chessgame?gid=1962999

[pgn][Event "Chess.com Speed Chess Championship 2017/18"]
[Site "chess.com INT"]
[Date "2017.10.04"]
[EventDate "2017.05.03"]
[Round "1.19"]
[Result "0-1"]
[White "Gadir Guseinov"]
[Black "Magnus Carlsen"]
[ECO "A40"]
[WhiteElo "2645"]
[BlackElo "2832"]
[PlyCount "102"]

1. e4 c6 2. d4 Na6 3. Nc3 Nc7 4. Nf3 d5 5. Bd3 Bg4 6. O-O e6
7. Be3 Bd6 8. h3 Bh5 9. e5 Be7 10. Be2 Nh6 11. Bxh6 gxh6
12. Ne1 Bg6 13. f4 Qd7 14. Nd3 O-O-O 15. b4 Bxd3 16. Bxd3 Bxb4
17. Na4 c5 18. Nxc5 Bxc5 19. dxc5 Qc6 20. Qh5 Qxc5+ 21. Kh1
Rd7 22. Qxh6 Kb8 23. Qg7 Qf8 24. Qf6 h5 25. Rab1 Rg8 26. Qh4
Qg7 27. Rf2 Qg3 28. Qxg3 Rxg3 29. Be2 h4 30. f5 a6 31. Rf4 Re3
32. Bf3 Rxe5 33. fxe6 Nxe6 34. Rxh4 Rc7 35. Rd1 Rxc2 36. Bxd5
Ka7 37. Rh7 Rc7 38. Bxe6 Rxe6 39. h4 Re2 40. a3 Rcc2 41. Rg1
Rf2 42. Kh2 Rf3 43. Ra1 b5 44. Rg7 Kb6 45. a4 b4 46. a5+ Kb5
47. Rg5+ Kc4 48. Rg4+ Kc3 49. Ra4 b3 50. Rgc4+ Kb2 51. Rg4 Rf5
0-1

[Event "PRO League Group Stage"]
[Site "chess.com INT"]
[Date "2019.03.13"]
[EventDate "2019.01.08"]
[Round "10"]
[Result "1/2-1/2"]
[White "Wang Yue"]
[Black "Hikaru Nakamura"]
[ECO "A40"]
[WhiteElo "2684"]
[BlackElo "2749"]
[PlyCount "211"]

1. d4 Na6 2. e4 c6 3. Nc3 Nc7 4. Nf3 d5 5. Bd3 dxe4 6. Nxe4
Nf6 7. Nxf6+ exf6 8. O-O Bd6 9. Re1+ Be6 10. c4 O-O 11. Qc2 g6
12. Bh6 Re8 13. h3 Bf8 14. Bxf8 Kxf8 15. b4 Kg7 16. a4 b5
17. axb5 cxb5 18. cxb5 Nd5 19. Qd2 Qd6 20. Be4 Rab8 21. Bxd5
Bxd5 22. Rxe8 Rxe8 23. Rxa7 Bxf3 24. gxf3 Qd5 25. Kg2 Qxb5
26. Ra5 Qd7 27. d5 Qd6 28. Rc5 Re5 29. h4 Rf5 30. Qc3 Rf4
31. Rc4 Rf5 32. Rd4 Rxd5 33. Rxd5 Qxd5 34. Qb2 Qb5 35. Qd4 h5
36. f4 Qc6+ 37. Kg3 Qh1 38. f3 Qe1+ 39. Kh3 Qe2 40. Kg3 Qf1
41. Qd2 Qb5 42. Qb2 Kg8 43. Kg2 Kg7 44. Kg3 Kg8 45. Kg2 Qb8
46. b5 Qb6 47. Qb4 Kg7 48. Qc4 Qe3 49. Qb4 Qb6 50. Qb2 Kf8
51. Qb4+ Ke8 52. Qe4+ Kf8 53. Qb4+ Kg8 54. Qc4 Kg7 55. Qc6 Qe3
56. b6 Qd2+ 57. Kf1 Qd1+ 58. Kf2 Qd2+ 59. Kf1 Qd3+ 60. Kf2
Qd4+ 61. Ke2 Qb2+ 62. Ke3 Qa3+ 63. Kd4 Qa6 64. Kc5 Qa3+
65. Kb5 Qb3+ 66. Ka5 Qa3+ 67. Kb5 Qd3+ 68. Ka5 Qa3+ 69. Kb5
Qd3+ 70. Kb4 Qd2+ 71. Ka4 Qxf4+ 72. Ka5 Qe5+ 73. Qb5 Qa1+
74. Qa4 Qc3+ 75. Qb4 Qe5+ 76. Ka6 Qa1+ 77. Kb5 Qf1+ 78. Ka5
Qa1+ 79. Qa4 Qc3+ 80. Ka6 Qd3+ 81. Ka7 Qe3 82. Ka6 Qd3+
83. Qb5 Qa3+ 84. Qa5 Qd3+ 85. Ka7 Qe3 86. Ka8 Qe8+ 87. Ka7 Qe3
88. Qb5 g5 89. hxg5 fxg5 90. Qb2+ Kg8 91. Ka6 Qe6 92. Ka5 Qd5+
93. Qb5 Qa2+ 94. Qa4 Qd5+ 95. Qb5 Qa2+ 96. Qa4 Qd2+ 97. Ka6
Qd6 98. Ka7 Qc5 99. Ka6 Qd6 100. Qa5 f6 101. Ka7 Qd4 102. Ka6
Qd3+ 103. Ka7 Qd4 104. Ka6 Qd3+ 105. Ka7 Qd4 106. Ka6 1/2-1/2[/pgn]

------------------------

1.d4 c5 2.d5 e5: Caruana tried it in 2018 without success against Aronian at fast TC:

https://old.chesstempo.com/gamedb/game/4374562
https://lichess.org/iw3tDNiz

------------------------

1.e4 g6 2.d4 Nf6: played by Carlsen, as you stated, and also by a lot of 2700+ GM against strong opposition, which surprised me. Special mention to Rapport (2698 Elo), who played it at least four times against Grischuk (2750 Elo) in August, 2017. Source: Lichess opening explorer.

------------------------

Regards from Spain.

Ajedrecista.

Thanks for the extra info. But I don't see how the posts you mention regarding gambits have anything to do with this topic; almost all of those gambit lines offer serious compensation, and so are more relevant to normal chess than to Armageddon chess; they are almost all likely drawn with proper play. The Elephant Gambit is at least bad enough to be unclear as to whether it is drawn or lost; I rejected the Latvian Gambit (2...f5?) and the Englund Gambit (1d4 e5?) as almost certainly lost. Most other Black gambits such as Benko, Blumenfeld, Marshall, Budapest, Albin etc. are also too good for an Armageddon set, they are all likely good enough to draw with proper play. If you know of any other gambits that arise after just two moves each which you believe to be near the win/draw line, please mention them.

[Ajedrecista]

Hello Larry:

There are different kinds of 'equal' positions IMHO: some that are easy to play (or easy to find the good moves); and others that are only moves to save the draw for the weak side, thus difficult to play, and more prone to blunder and get a non draw result. I think that some of those openings fall on this ground, specially in human vs. human at fast TC.

Having said this, there are 72,078 different positions after four plies, too many for analyse them at high depths. I have been thinking in some openings (discarding gifts as: 1.e4 e5 2.Nf3 Qg5??; 1.f4 e6 2.g4?? Qh4#; and so on) and I come with some lines, limited to four plies. Since I do not know exactly where is the win/draw border regarding eval, I post some lines with SF 15.1 analysis using default NNUE net, single core (for being deterministic), Multi-PV=1, 512 MB of hash and depth 35:

Code: Select all

Stockfish 15.1 by the Stockfish developers (see AUTHORS file)
uci
id name Stockfish 15.1
[...]
uciok
setoption name Hash value 512
position fen XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
go depth 35

1.e4 Nf6 2.d4 Ne4 {Alekhine Defence, Mokele Mbembe (Bücker) Variation}.

Code: Select all

info depth 35 seldepth 50 multipv 1 score cp 131 nodes 44506012 nps 449269 hashfull 595 tbhits 0 time 99063 pv d2d4 f7f6 f2f3 e4g5 f3f4 g5f7 g1f3 e7e6 b1c3 d7d5 c1e3 f8b4 f1d3 c7c5 d4c5 b8d7 e5f6 d8f6 e1g1 e8g8 c3a4 f6e7 c2c3 b4c5 a4c5 d7c5 e3d4 b7b6 a2a4 c8a6 d3a6 c5a6 a4a5 a6c5
bestmove d2d4 ponder f7f6

1.e4 d5 2.exd5 e6 {Scandinavian Defence, 2. ... e6}.

Code: Select all

info depth 35 seldepth 43 multipv 1 score cp 114 nodes 37455694 nps 461015 hashfull 503 tbhits 0 time 81246 pv f1b5 c7c6 d5c6 b7c6 b5e2 g8f6 g1f3 c8b7 e1g1 c6c5 b1a3 b8c6 f1e1 f8e7 d2d3 d8c7 c2c3 h7h6 e2f1 a8d8 d1e2 e8g8 a3c4 f6d7 g2g3 d7b6 c1f4 c7c8 a1d1 f8e8 f1g2 e7f8 f3d2 e6e5 f4e3 b7a6 h2h4
bestmove f1b5 ponder c7c6

1.e4 e5 2.Nf3 Qf6 {Greco Defence}.

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 107 nodes 34244596 nps 464201 hashfull 471 tbhits 0 time 73771 pv b1c3 b8c6 d2d4 f8b4 c1d2 c6d4 f3d4 e5d4 c3b5 b4d2 d1d2 f6e5 e1c1 g8e7 f2f4 e5c5 f1e2 d7d6 d2d4 c5d4 d1d4 e8d8 e4e5 d6d5 c2c4 c7c6 b5d6 e7f5 d6f5 c8f5 e2f3 d8e7 g2g4 f5d7 h1d1 g7g6 c4d5 c6d5 d1d3 a8c8 c1d2 d7e6 f3d5 e6g4 d3a3 g4e6 a3a7 h8d8
bestmove b1c3 ponder b8c6

1.e4 d5 2.exd5 e5 {Scandinavian Defence, Böhnke Gambit}.

Code: Select all

info depth 35 seldepth 43 multipv 1 score cp 99 nodes 30896636 nps 451144 hashfull 414 tbhits 0 time 68485 pv d5e6 c8e6 g1f3 g8f6 d2d4 c7c5 f1b5 b8c6 e1g1 d8b6 b5c6 b6c6 f1e1 f8e7 c1e3 e8g8 b1d2 f6d5 c2c4 d5e3 e1e3 c5d4 f3d4 c6b6 d4e6 f7e6 d1b3 e7f6 b3b6 a7b6 e3b3 f8d8 d2e4 a8a2 e4f6 g7f6 a1b1
bestmove d5e6 ponder c8e6

1.e4 a6 2.d4 b5 {St. George Defence}.

Code: Select all

info depth 35 seldepth 39 multipv 1 score cp 97 nodes 40354133 nps 450813 hashfull 506 tbhits 0 time 89514 pv a2a4 b5b4 b1d2 e7e6 f1d3 d7d5 g1f3 g8f6 c2c4 b4c3 b2c3 c7c5 e1g1 b8c6 e4d5 f6d5 f3e5 c8b7 d2e4 c6e5 d4e5 d8c7 a1b1 b7c6 f1e1 f8e7 d1h5 h7h6 d3c2 a8d8 e4d6 e7d6 e5d6
bestmove a2a4 ponder b5b4

1.d4 f5 2.c4 g6 {Dutch Defence, 2. ... g6}.

Code: Select all

info depth 35 seldepth 46 multipv 1 score cp 95 nodes 49834663 nps 470546 hashfull 609 tbhits 0 time 105908 pv h2h4 f8g7 h4h5 c7c5 d4d5 d7d6 g1h3 g8h6 b1c3 g7c3 b2c3 h6f7 d1a4 c8d7 a4b3 d7c8 g2g3 b8d7 b3a4 d8b6 h3g5 f7g5 c1g5 b6b2 a1c1 g6h5 a4b3 b2b3 a2b3
bestmove h2h4 ponder f8g7

1.e4 Nf6 2.d4 Ng8 {Alekhine Defence, Brooklyn Defence (Retreat Variation)}.

Code: Select all

info depth 35 seldepth 43 multipv 1 score cp 92 nodes 44086791 nps 451958 hashfull 587 tbhits 0 time 97546 pv d2d4 d7d6 g1f3 d6e5 f3e5 b8d7 d1f3 d7e5 d4e5 c7c6 f3c3 c8e6 c1g5 h7h6 g5e3 e6d5 b1d2 e7e6 e1c1 g8e7 e3c5 e7f5 c5f8 h8f8 c3a3 b7b5 b2b3 d8c7 c2c4 b5c4 b3c4 a8b8 d2b3 d5e4 f2f3 f5e3 a3d6 c7d6 d1d6
bestmove d2d4 ponder d7d6

1.e4 e5 2.f4 Qf6 {King's Gambit, Norwald Variation}.

Code: Select all

info depth 35 seldepth 51 multipv 1 score cp 90 nodes 36036589 nps 458889 hashfull 476 tbhits 0 time 78530 pv b1c3 e5f4 g1f3 d7d6 c3d5 f6d8 d5f4 g8f6 d2d4 f6e4 f1d3 e4f6 e1g1 f8e7 f3g5 e8g8 d1f3 h7h6 g5e4 b8d7 e4g3 f6h7 f4h5 d7f6 c1h6 g7h6 f3f4 f6h5 f4h6 f7f5 g3h5 e7f6 c2c3 g8h8 a1e1 f8f7 h5f4 f7g7 h6h5 c8d7 f4g6 g7g6 h5g6 d8g8 d3f5 g8g6 f5g6 f6g5
bestmove b1c3 ponder e5f4

1.e4 e6 2.d4 Nf6 {French Defence, Mediterranean Defence}.

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 89 nodes 26757185 nps 466445 hashfull 364 tbhits 0 time 57364 pv e4e5 f6d5 c2c4 d5e7 b1c3 d7d6 e5d6 c7d6 g1f3 g7g6 h2h4 h7h6 c1f4 f8g7 c3b5 e8g8 d1d2 e7f5 g2g4 e6e5 d4e5 d6e5 e1c1 d8d2 f4d2 f5e7 g4g5 h6h5 b5c7 c8g4 f1e2 e5e4 f3g1 b8c6 c7a8 f8a8 e2g4 h5g4 g1e2 c6e5 e2f4 e5c4 d2c3 a8c8 c3g7 g8g7 c1b1 e7f5 d1d7 e4e3 f2e3
bestmove e4e5 ponder f6d5

1.e4 d5 2.exd5 c6 {Scandinavian Defence, Blackburne Gambit}.

Code: Select all

info depth 35 seldepth 51 multipv 1 score cp 82 nodes 36422964 nps 477778 hashfull 479 tbhits 0 time 76234 pv d5c6 b8c6 g1f3 e7e5 b1c3 g8f6 f1b5 f8d6 d2d4 e8g8 b5c6 e5d4 d1d4 b7c6 c1e3 c6c5 d4d2 d8b6 e1c1 d6e7 e3f4 f8e8 h1e1 c8b7 d2d3 e7f8 a2a3 b7a6 d3f5 g7g6 f5h3 a6b7 f4e5 a8d8 d1d8 e8d8
bestmove d5c6 ponder b8c6

You can further analyse some of these lines to higher depths if you think they can have potential.

The Dutch Defence line could lead to crazy variations with a pawn in the seventh rank at 10th move:

1.d4 f5 2.c4 g6 3.h4 Bg7 4.h5 c5 5.dxc5 Na6 6.e4 fxe4 7.hxg6 Qa5+ 8.Bd2 Qxc5 9.Rxh7 Rxh7 10.gxh7 Nf6 11.Nh3 Nxh7

Code: Select all

info depth 35 seldepth 45 multipv 1 score cp 76 nodes 38043019 nps 460747 hashfull 478 tbhits 0 time 82568 pv b1c3 c5f5 g2g4 f5f3 d1f3 e4f3 f1d3 h7f6 e1c1 d7d6 d1g1 f6g4 c3d5 e8f8 h3g5 g4e5 d3c2 a6c5 d2e3 a8b8 b2b3 b7b5 g5h7 f8f7 c4b5 b8b5 d5e7 c8d7 c2f5 d7e6 g1g3 c5d3 c1b1
bestmove b1c3 ponder c5f5

Regards from Spain.

Ajedrecista.

[lkaufman]

Hello Larry:

There are different kinds of 'equal' positions IMHO: some that are easy to play (or easy to find the good moves); and others that are only moves to save the draw for the weak side, thus difficult to play, and more prone to blunder and get a non draw result. I think that some of those openings fall on this ground, specially in human vs. human at fast TC.

Having said this, there are 72,078 different positions after four plies, too many for analyse them at high depths. I have been thinking in some openings (discarding gifts as: 1.e4 e5 2.Nf3 Qg5??; 1.f4 e6 2.g4?? Qh4#; and so on) and I come with some lines, limited to four plies. Since I do not know exactly where is the win/draw border regarding eval, I post some lines with SF 15.1 analysis using default NNUE net, single core (for being deterministic), Multi-PV=1, 512 MB of hash and depth 35:

Code: Select all

Stockfish 15.1 by the Stockfish developers (see AUTHORS file)
uci
id name Stockfish 15.1
[...]
uciok
setoption name Hash value 512
position fen XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
go depth 35

1.e4 Nf6 2.d4 Ne4 {Alekhine Defence, Mokele Mbembe (Bücker) Variation}.

Code: Select all

info depth 35 seldepth 50 multipv 1 score cp 131 nodes 44506012 nps 449269 hashfull 595 tbhits 0 time 99063 pv d2d4 f7f6 f2f3 e4g5 f3f4 g5f7 g1f3 e7e6 b1c3 d7d5 c1e3 f8b4 f1d3 c7c5 d4c5 b8d7 e5f6 d8f6 e1g1 e8g8 c3a4 f6e7 c2c3 b4c5 a4c5 d7c5 e3d4 b7b6 a2a4 c8a6 d3a6 c5a6 a4a5 a6c5
bestmove d2d4 ponder f7f6

Something is wrong here; I get a score of -.62 (meaning some comp. for White for the pawn but not enough, as you would expect). Your pv is illegal, starts with d4 when d4 has already been played. Probably you input the wrong position.

Regards from Spain.

Ajedrecista.

[Ajedrecista]

Hello Larry:

I think I analysed the correct FEN. You catched a typo: the line is 1.e4 Nf6 2.e5 Ne4. The same typo is in the other Alekhine Defence, where it is 1.e4 Nf6 2.e5 Ng8. The rest of lines seem correct, I mean, without typos.

I hope you can find some of the lines useful for your purpose.

Regards from Spain.

Ajedrecista.

[lkaufman]

Hello Larry:

There are different kinds of 'equal' positions IMHO: some that are easy to play (or easy to find the good moves); and others that are only moves to save the draw for the weak side, thus difficult to play, and more prone to blunder and get a non draw result. I think that some of those openings fall on this ground, specially in human vs. human at fast TC.

Having said this, there are 72,078 different positions after four plies, too many for analyse them at high depths. I have been thinking in some openings (discarding gifts as: 1.e4 e5 2.Nf3 Qg5??; 1.f4 e6 2.g4?? Qh4#; and so on) and I come with some lines, limited to four plies. Since I do not know exactly where is the win/draw border regarding eval, I post some lines with SF 15.1 analysis using default NNUE net, single core (for being deterministic), Multi-PV=1, 512 MB of hash and depth 35:

Code: Select all

Stockfish 15.1 by the Stockfish developers (see AUTHORS file)
uci
id name Stockfish 15.1
[...]
uciok
setoption name Hash value 512
position fen XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
go depth 35

1.e4 Nf6 2.d4 Ne4 {Alekhine Defence, Mokele Mbembe (Bücker) Variation}.

Code: Select all

info depth 35 seldepth 50 multipv 1 score cp 131 nodes 44506012 nps 449269 hashfull 595 tbhits 0 time 99063 pv d2d4 f7f6 f2f3 e4g5 f3f4 g5f7 g1f3 e7e6 b1c3 d7d5 c1e3 f8b4 f1d3 c7c5 d4c5 b8d7 e5f6 d8f6 e1g1 e8g8 c3a4 f6e7 c2c3 b4c5 a4c5 d7c5 e3d4 b7b6 a2a4 c8a6 d3a6 c5a6 a4a5 a6c5
bestmove d2d4 ponder f7f6

1.e4 d5 2.exd5 e6 {Scandinavian Defence, 2. ... e6}.

Code: Select all

info depth 35 seldepth 43 multipv 1 score cp 114 nodes 37455694 nps 461015 hashfull 503 tbhits 0 time 81246 pv f1b5 c7c6 d5c6 b7c6 b5e2 g8f6 g1f3 c8b7 e1g1 c6c5 b1a3 b8c6 f1e1 f8e7 d2d3 d8c7 c2c3 h7h6 e2f1 a8d8 d1e2 e8g8 a3c4 f6d7 g2g3 d7b6 c1f4 c7c8 a1d1 f8e8 f1g2 e7f8 f3d2 e6e5 f4e3 b7a6 h2h4
bestmove f1b5 ponder c7c6

1.e4 e5 2.Nf3 Qf6 {Greco Defence}.

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 107 nodes 34244596 nps 464201 hashfull 471 tbhits 0 time 73771 pv b1c3 b8c6 d2d4 f8b4 c1d2 c6d4 f3d4 e5d4 c3b5 b4d2 d1d2 f6e5 e1c1 g8e7 f2f4 e5c5 f1e2 d7d6 d2d4 c5d4 d1d4 e8d8 e4e5 d6d5 c2c4 c7c6 b5d6 e7f5 d6f5 c8f5 e2f3 d8e7 g2g4 f5d7 h1d1 g7g6 c4d5 c6d5 d1d3 a8c8 c1d2 d7e6 f3d5 e6g4 d3a3 g4e6 a3a7 h8d8
bestmove b1c3 ponder b8c6

1.e4 d5 2.exd5 e5 {Scandinavian Defence, Böhnke Gambit}.

Code: Select all

info depth 35 seldepth 43 multipv 1 score cp 99 nodes 30896636 nps 451144 hashfull 414 tbhits 0 time 68485 pv d5e6 c8e6 g1f3 g8f6 d2d4 c7c5 f1b5 b8c6 e1g1 d8b6 b5c6 b6c6 f1e1 f8e7 c1e3 e8g8 b1d2 f6d5 c2c4 d5e3 e1e3 c5d4 f3d4 c6b6 d4e6 f7e6 d1b3 e7f6 b3b6 a7b6 e3b3 f8d8 d2e4 a8a2 e4f6 g7f6 a1b1
bestmove d5e6 ponder c8e6

1.e4 a6 2.d4 b5 {St. George Defence}.

Code: Select all

info depth 35 seldepth 39 multipv 1 score cp 97 nodes 40354133 nps 450813 hashfull 506 tbhits 0 time 89514 pv a2a4 b5b4 b1d2 e7e6 f1d3 d7d5 g1f3 g8f6 c2c4 b4c3 b2c3 c7c5 e1g1 b8c6 e4d5 f6d5 f3e5 c8b7 d2e4 c6e5 d4e5 d8c7 a1b1 b7c6 f1e1 f8e7 d1h5 h7h6 d3c2 a8d8 e4d6 e7d6 e5d6
bestmove a2a4 ponder b5b4

1.d4 f5 2.c4 g6 {Dutch Defence, 2. ... g6}.

Code: Select all

info depth 35 seldepth 46 multipv 1 score cp 95 nodes 49834663 nps 470546 hashfull 609 tbhits 0 time 105908 pv h2h4 f8g7 h4h5 c7c5 d4d5 d7d6 g1h3 g8h6 b1c3 g7c3 b2c3 h6f7 d1a4 c8d7 a4b3 d7c8 g2g3 b8d7 b3a4 d8b6 h3g5 f7g5 c1g5 b6b2 a1c1 g6h5 a4b3 b2b3 a2b3
bestmove h2h4 ponder f8g7

1.e4 Nf6 2.d4 Ng8 {Alekhine Defence, Brooklyn Defence (Retreat Variation)}.

Code: Select all

info depth 35 seldepth 43 multipv 1 score cp 92 nodes 44086791 nps 451958 hashfull 587 tbhits 0 time 97546 pv d2d4 d7d6 g1f3 d6e5 f3e5 b8d7 d1f3 d7e5 d4e5 c7c6 f3c3 c8e6 c1g5 h7h6 g5e3 e6d5 b1d2 e7e6 e1c1 g8e7 e3c5 e7f5 c5f8 h8f8 c3a3 b7b5 b2b3 d8c7 c2c4 b5c4 b3c4 a8b8 d2b3 d5e4 f2f3 f5e3 a3d6 c7d6 d1d6
bestmove d2d4 ponder d7d6

1.e4 e5 2.f4 Qf6 {King's Gambit, Norwald Variation}.

Code: Select all

info depth 35 seldepth 51 multipv 1 score cp 90 nodes 36036589 nps 458889 hashfull 476 tbhits 0 time 78530 pv b1c3 e5f4 g1f3 d7d6 c3d5 f6d8 d5f4 g8f6 d2d4 f6e4 f1d3 e4f6 e1g1 f8e7 f3g5 e8g8 d1f3 h7h6 g5e4 b8d7 e4g3 f6h7 f4h5 d7f6 c1h6 g7h6 f3f4 f6h5 f4h6 f7f5 g3h5 e7f6 c2c3 g8h8 a1e1 f8f7 h5f4 f7g7 h6h5 c8d7 f4g6 g7g6 h5g6 d8g8 d3f5 g8g6 f5g6 f6g5
bestmove b1c3 ponder e5f4

1.e4 e6 2.d4 Nf6 {French Defence, Mediterranean Defence}.

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 89 nodes 26757185 nps 466445 hashfull 364 tbhits 0 time 57364 pv e4e5 f6d5 c2c4 d5e7 b1c3 d7d6 e5d6 c7d6 g1f3 g7g6 h2h4 h7h6 c1f4 f8g7 c3b5 e8g8 d1d2 e7f5 g2g4 e6e5 d4e5 d6e5 e1c1 d8d2 f4d2 f5e7 g4g5 h6h5 b5c7 c8g4 f1e2 e5e4 f3g1 b8c6 c7a8 f8a8 e2g4 h5g4 g1e2 c6e5 e2f4 e5c4 d2c3 a8c8 c3g7 g8g7 c1b1 e7f5 d1d7 e4e3 f2e3
bestmove e4e5 ponder f6d5

1.e4 d5 2.exd5 c6 {Scandinavian Defence, Blackburne Gambit}.

Code: Select all

info depth 35 seldepth 51 multipv 1 score cp 82 nodes 36422964 nps 477778 hashfull 479 tbhits 0 time 76234 pv d5c6 b8c6 g1f3 e7e5 b1c3 g8f6 f1b5 f8d6 d2d4 e8g8 b5c6 e5d4 d1d4 b7c6 c1e3 c6c5 d4d2 d8b6 e1c1 d6e7 e3f4 f8e8 h1e1 c8b7 d2d3 e7f8 a2a3 b7a6 d3f5 g7g6 f5h3 a6b7 f4e5 a8d8 d1d8 e8d8
bestmove d5c6 ponder b8c6

You can further analyse some of these lines to higher depths if you think they can have potential.

The Dutch Defence line could lead to crazy variations with a pawn in the seventh rank at 10th move:

1.d4 f5 2.c4 g6 3.h4 Bg7 4.h5 c5 5.dxc5 Na6 6.e4 fxe4 7.hxg6 Qa5+ 8.Bd2 Qxc5 9.Rxh7 Rxh7 10.gxh7 Nf6 11.Nh3 Nxh7

Code: Select all

info depth 35 seldepth 45 multipv 1 score cp 76 nodes 38043019 nps 460747 hashfull 478 tbhits 0 time 82568 pv b1c3 c5f5 g2g4 f5f3 d1f3 e4f3 f1d3 h7f6 e1c1 d7d6 d1g1 f6g4 c3d5 e8f8 h3g5 g4e5 d3c2 a6c5 d2e3 a8b8 b2b3 b7b5 g5h7 f8f7 c4b5 b8b5 d5e7 c8d7 c2f5 d7e6 g1g3 c5d3 c1b1
bestmove b1c3 ponder c5f5

Regards from Spain.

Ajedrecista.

I also found several of your above lines (but not all). The reason most were rejected is that they didn't have my required minimum of thirty games in the Hiarcs database so we wouldn't have to rely solely on engine evals. The 1.e4 d5 2.exd5 c6 gambit comes closest, with 23 games, but its eval is at the very low end of the allowed range, it's probably not that close to losing. Several others have just a single digit number of games. The St. George defense has enough games, but it is normally just going to transpose to my Polish defense line 1.d4 b5 after 2.e4 Bb7 with ...a6 soon needed. The only line above that seems to fulfill all requirements is your Dutch defense line. I must admit that I never thought of it, because I assumed it would just transpose to the Leningrad Dutch, but now I know that 2...g6? is an incorrect move order, 2...Nf6 first is necessary. The King's gambit line violates the principle that White must have played no clearly inferior moves; obviously the King's Gambit is objectively a mistake. The Bucker variation gives an eval that is too clearly won (above 1.2) for this purpose.
Obviously, the problem is that very few lines that are on the cusp of losing will occur in practical play thirty times, especially after just two moves where there are not likely to be complex tactics (well, your Dutch example seems an exception!). Probably there are others that neither of us thought of. I suppose that if we ever needed a substantially larger set of openings, I could just raise the two move limit to three moves.