A french defence game

[Lyudmil Tsvetkov]

I've always thought database statistics were useless. So we agree there. The tree is just too vast and the samples so tiny, that the win/loss ratio means zero in nearly all situations.

Certainly the original line given looks very bad for black. My IDeA analysis in this variation would need substantially more effort to show black is in big trouble after White's move 6, but I may divert it from other things later this week to further investigate this.

Black's main alternatives are currently 6 ...Qa5 and ...Ne7 as played in the game.

Currently the scores favor White responding Qg4 (0.24) over Nf3 (0.00) in the ... Ne7 continuation.

Looking at how easily black can get into trouble even when played by a computer, the variation looks dangerous. The engines routinely fail to see the pawn pushes on the king side and strategic retreats like your Qf3! That said, I'm always surprised how many resources for defense can be found when talking about variations so early in the game. I'd be surprised if black was lost after White's move 6, but considering the issues you mentioned with the sheer depth, it is possible.

Mike

there are many defences, perfect play is difficult to get at.

I know that line almost by heart, as I have played thousands games like that against SF.

difficult to make sense what is going on in the sea of variations, I am basically doing a positional assessment:

- pair of bishops
- long pointed chain c3-d4-e5, with blocked center and backward black f7 shelter pawn
- better mobility

white a2 and c2/c3 isolated/doubled pawns are worth very small penalty, especially with the presence of bishop pair.

so, white has substantial evaluation lead, but what happens with perfect play no one knows.

I guess Nf3 is better than early Qg4 sally(with a black knight on f4, the queen feels a bit awkward), in some lines a straight f2-f4 push is even stronger, those 0.0, 0.24 scores are really funny, +70-80cps white advantage is of course much more realistic.

[zullil]

funny, SF does not spot Qf3 Rh8 h5 Ne3 fe3! Qg5 0-0-0 and white wins:

[d]r1b4r/p4pk1/1pn1p3/3pP1qP/2pP4/P1P1PQ2/2P4N/2KR3R b - - 0 4

other lines are more or less similar and lead to very convincing white scores.

asmFish finds Qf3 quickly, so Stockfish should too.

Code: Select all

info depth 27 seldepth 40 multipv 1 time 9755 nps 10144614 score cp -51 lowerbound nodes 98960713 tbhits 0 hashfull 15 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 10037 nps 10581027 score cp -43 lowerbound nodes 106201769 tbhits 0 hashfull 17 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 10805 nps 11691859 score cp -33 lowerbound nodes 126330542 tbhits 0 hashfull 19 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 12692 nps 13821887 score cp -17 lowerbound nodes 175427397 tbhits 0 hashfull 30 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 15597 nps 16964628 score cp 3 lowerbound nodes 264597303 tbhits 0 hashfull 47 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 22016 nps 22229699 score cp 31 lowerbound nodes 489409067 tbhits 0 hashfull 102 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 currmove h2g4 currmovenumber 2
info depth 27 currmove h5h7 currmovenumber 3
info depth 27 currmove e3f4 currmovenumber 4
info depth 27 currmove h5h6 currmovenumber 5
info depth 27 currmove e1d2 currmovenumber 6
info depth 27 currmove e1e2 currmovenumber 7
info depth 27 currmove e1c1 currmovenumber 8
info depth 27 currmove g5g6 currmovenumber 9
info depth 27 currmove h5g6 currmovenumber 10
info depth 27 currmove e3d2 currmovenumber 11
info depth 27 currmove e1d1 currmovenumber 12
info depth 27 currmove h1g1 currmovenumber 13
info depth 27 currmove h5h8 currmovenumber 14
info depth 27 currmove e3c1 currmovenumber 15
info depth 27 currmove f2f4 currmovenumber 16
info depth 27 currmove h2f3 currmovenumber 17
info depth 27 currmove h5g4 currmovenumber 18
info depth 27 currmove a1c1 currmovenumber 19
info depth 27 currmove a3a4 currmovenumber 20
info depth 27 currmove h2f1 currmovenumber 21
info depth 27 currmove h1f1 currmovenumber 22
info depth 27 currmove a1d1 currmovenumber 23
info depth 27 currmove a1a2 currmovenumber 24
info depth 27 currmove h5e2 currmovenumber 25
info depth 27 currmove e1g1 currmovenumber 26
info depth 27 currmove f2f3 currmovenumber 27
info depth 27 currmove e1f1 currmovenumber 28
info depth 27 currmove h5d1 currmovenumber 29
info depth 27 currmove a1b1 currmovenumber 30
info depth 27 currmove h5f7 currmovenumber 31
info depth 27 seldepth 41 multipv 1 time 24052 nps 23425287 score cp 26 nodes 563425009 tbhits 0 hashfull 114 pv h5f3 f8h8 h2g4 g7f8 g4f6 c6e7 h4h5 e7g8 e1c1 g8f6 e5f6 f8e8 e3f4 e8d7 f4e5 d7c6 g5g6 f5h6 a3a4 a7a5 f3f4 c6b7 c1d2 c8d7 g6g7 h8h7 d1a1 h6f5 f4g4 d7c6 a1a2 b7a7 h1b1 f5h6 g4f4 h6g8 f4g5 h7h6

and after more time

Code: Select all

info depth 42 seldepth 76 multipv 1 time 854389 nps 36479390 score cp 120 nodes 31167589739 tbhits 0 hashfull 999 pv h5f3 f8h8 h2g4 f5h4 f3f6 d8f6 g5f6 g7f8 e3h6 f8e8 h1h4 e8d7 e1e2 d7c7 a1h1 c8d7 g4e3 a7a5 h6f4 h8h4 h1h4 a8g8 h4h7 c6d8 f4h6 c7c6 h6g7 c6b5 h7h1 g8e8 f2f4 b5c6 e2f3 c6b5 e3g4 b5c6 h1a1 c6c7 a1g1 c7c6 g4e3 c6b5 g1a1 b5c6 a1b1 e8g8 f3e2 g8e8 b1h1 c6c7 h1h7 c7c6 e2d2

from that position you should squeeze at least +450cps.

so, your SF has nothing to be proud of, raving again somewhere in the fringe lines...

My asmFish will now hang its head in shame, as it can only reach +1.45.

Here's the latest "fringe line", after 17. Qf3:

Code: Select all

+1.45 17... Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Ke2 Kc7 24. Rah1 Bd7 25. Kd1 a5 26. f3 b5 27. Bd2 Rxh4 28. Rxh4 Rg8 29. Rh7 Be8 30. Be3 a4 31. Nh6 Rf8 32. Ke2 Nd8 33. Ng4 Rg8 34. Bh6 Kb6 35. Bd2 Nc6 36. Kd1 Nd8 37. Bc1 Nc6 38. Ke2 Nd8 39. Kf2 Rf8 40. Nh6 Kc6 41. Ke2 Kb6 42. Kd2 Kc7 43. Ng4 Nc6 44. Ke2 Nd8 45. Bh6 Rg8 46. Bd2 Kb6 47. Kf2 Kc7 48. Nh6 Rf8 49. Ke2 Bd7 50. Ng4 Bc6 51. Be3 (depth 58, 2:27:58)

[zullil]

funny, SF does not spot Qf3 Rh8 h5 Ne3 fe3! Qg5 0-0-0 and white wins:

[d]r1b4r/p4pk1/1pn1p3/3pP1qP/2pP4/P1P1PQ2/2P4N/2KR3R b - - 0 4

other lines are more or less similar and lead to very convincing white scores.

asmFish finds Qf3 quickly, so Stockfish should too.

Code: Select all

info depth 27 seldepth 40 multipv 1 time 9755 nps 10144614 score cp -51 lowerbound nodes 98960713 tbhits 0 hashfull 15 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 10037 nps 10581027 score cp -43 lowerbound nodes 106201769 tbhits 0 hashfull 17 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 10805 nps 11691859 score cp -33 lowerbound nodes 126330542 tbhits 0 hashfull 19 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 12692 nps 13821887 score cp -17 lowerbound nodes 175427397 tbhits 0 hashfull 30 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 15597 nps 16964628 score cp 3 lowerbound nodes 264597303 tbhits 0 hashfull 47 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 22016 nps 22229699 score cp 31 lowerbound nodes 489409067 tbhits 0 hashfull 102 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 currmove h2g4 currmovenumber 2
info depth 27 currmove h5h7 currmovenumber 3
info depth 27 currmove e3f4 currmovenumber 4
info depth 27 currmove h5h6 currmovenumber 5
info depth 27 currmove e1d2 currmovenumber 6
info depth 27 currmove e1e2 currmovenumber 7
info depth 27 currmove e1c1 currmovenumber 8
info depth 27 currmove g5g6 currmovenumber 9
info depth 27 currmove h5g6 currmovenumber 10
info depth 27 currmove e3d2 currmovenumber 11
info depth 27 currmove e1d1 currmovenumber 12
info depth 27 currmove h1g1 currmovenumber 13
info depth 27 currmove h5h8 currmovenumber 14
info depth 27 currmove e3c1 currmovenumber 15
info depth 27 currmove f2f4 currmovenumber 16
info depth 27 currmove h2f3 currmovenumber 17
info depth 27 currmove h5g4 currmovenumber 18
info depth 27 currmove a1c1 currmovenumber 19
info depth 27 currmove a3a4 currmovenumber 20
info depth 27 currmove h2f1 currmovenumber 21
info depth 27 currmove h1f1 currmovenumber 22
info depth 27 currmove a1d1 currmovenumber 23
info depth 27 currmove a1a2 currmovenumber 24
info depth 27 currmove h5e2 currmovenumber 25
info depth 27 currmove e1g1 currmovenumber 26
info depth 27 currmove f2f3 currmovenumber 27
info depth 27 currmove e1f1 currmovenumber 28
info depth 27 currmove h5d1 currmovenumber 29
info depth 27 currmove a1b1 currmovenumber 30
info depth 27 currmove h5f7 currmovenumber 31
info depth 27 seldepth 41 multipv 1 time 24052 nps 23425287 score cp 26 nodes 563425009 tbhits 0 hashfull 114 pv h5f3 f8h8 h2g4 g7f8 g4f6 c6e7 h4h5 e7g8 e1c1 g8f6 e5f6 f8e8 e3f4 e8d7 f4e5 d7c6 g5g6 f5h6 a3a4 a7a5 f3f4 c6b7 c1d2 c8d7 g6g7 h8h7 d1a1 h6f5 f4g4 d7c6 a1a2 b7a7 h1b1 f5h6 g4f4 h6g8 f4g5 h7h6

and after more time

Code: Select all

info depth 42 seldepth 76 multipv 1 time 854389 nps 36479390 score cp 120 nodes 31167589739 tbhits 0 hashfull 999 pv h5f3 f8h8 h2g4 f5h4 f3f6 d8f6 g5f6 g7f8 e3h6 f8e8 h1h4 e8d7 e1e2 d7c7 a1h1 c8d7 g4e3 a7a5 h6f4 h8h4 h1h4 a8g8 h4h7 c6d8 f4h6 c7c6 h6g7 c6b5 h7h1 g8e8 f2f4 b5c6 e2f3 c6b5 e3g4 b5c6 h1a1 c6c7 a1g1 c7c6 g4e3 c6b5 g1a1 b5c6 a1b1 e8g8 f3e2 g8e8 b1h1 c6c7 h1h7 c7c6 e2d2

from that position you should squeeze at least +450cps.

so, your SF has nothing to be proud of, raving again somewhere in the fringe lines...

My asmFish will now hang its head in shame, as it can only reach +1.45.

Here's the latest "fringe line", after 17. Qf3:

Code: Select all

+1.45 17... Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Ke2 Kc7 24. Rah1 Bd7 25. Kd1 a5 26. f3 b5 27. Bd2 Rxh4 28. Rxh4 Rg8 29. Rh7 Be8 30. Be3 a4 31. Nh6 Rf8 32. Ke2 Nd8 33. Ng4 Rg8 34. Bh6 Kb6 35. Bd2 Nc6 36. Kd1 Nd8 37. Bc1 Nc6 38. Ke2 Nd8 39. Kf2 Rf8 40. Nh6 Kc6 41. Ke2 Kb6 42. Kd2 Kc7 43. Ng4 Nc6 44. Ke2 Nd8 45. Bh6 Rg8 46. Bd2 Kb6 47. Kf2 Kc7 48. Nh6 Rf8 49. Ke2 Bd7 50. Ng4 Bc6 51. Be3 (depth 58, 2:27:58)

Even after more than 15 hours of searching, presumably of trillions of nodes, asmFish is unable to "squeeze at least +450 cps" from this position. So either Lyudmil or asmFish has somewhat misunderstood the position. Probably asmFish.

Code: Select all

+1.56 17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. O-O-O Kc7 24. Rdh1 Bd7 25. Be3 Rxh4 26. Rxh4 Rg8 27. f3 Nd8 28. Rh7 Be8 29. Kd1 Ba4 30. Nh6 Rf8 31. Kd2 Kc8 32. f4 Kc7 33. f5 Kd7 34. fxe6+ Kxe6 35. Kc1 Nc6 36. Kb2 Nd8 37. Bc1 Nc6 38. Bd2 Nd8 39. Be3 Nc6 40. Kc1 Nd8 41. Bf4 Nc6 42. Bg3 Nd8 43. Kb2 Nc6 44. Ng4 Nd8 45. Kc1 Kd7 46. Kd1 Rg8 47. Rg7 Rf8 48. Nh6 Ke6 49. Rh7 Kd7 50. Bf4 Ke6 51. Kd2 Bc6 52. Ke1 Ba4 53. Kd1 Nc6 54. Kd2 Nd8 55. Bg5 (depth 66, 15:07:10)

[Lyudmil Tsvetkov]

funny, SF does not spot Qf3 Rh8 h5 Ne3 fe3! Qg5 0-0-0 and white wins:

[d]r1b4r/p4pk1/1pn1p3/3pP1qP/2pP4/P1P1PQ2/2P4N/2KR3R b - - 0 4

other lines are more or less similar and lead to very convincing white scores.

asmFish finds Qf3 quickly, so Stockfish should too.

Code: Select all

info depth 27 seldepth 40 multipv 1 time 9755 nps 10144614 score cp -51 lowerbound nodes 98960713 tbhits 0 hashfull 15 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 10037 nps 10581027 score cp -43 lowerbound nodes 106201769 tbhits 0 hashfull 17 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 10805 nps 11691859 score cp -33 lowerbound nodes 126330542 tbhits 0 hashfull 19 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 12692 nps 13821887 score cp -17 lowerbound nodes 175427397 tbhits 0 hashfull 30 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 15597 nps 16964628 score cp 3 lowerbound nodes 264597303 tbhits 0 hashfull 47 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 seldepth 40 multipv 1 time 22016 nps 22229699 score cp 31 lowerbound nodes 489409067 tbhits 0 hashfull 102 pv h5f3
info depth 27 currmove h5f3 currmovenumber 1
info depth 27 currmove h2g4 currmovenumber 2
info depth 27 currmove h5h7 currmovenumber 3
info depth 27 currmove e3f4 currmovenumber 4
info depth 27 currmove h5h6 currmovenumber 5
info depth 27 currmove e1d2 currmovenumber 6
info depth 27 currmove e1e2 currmovenumber 7
info depth 27 currmove e1c1 currmovenumber 8
info depth 27 currmove g5g6 currmovenumber 9
info depth 27 currmove h5g6 currmovenumber 10
info depth 27 currmove e3d2 currmovenumber 11
info depth 27 currmove e1d1 currmovenumber 12
info depth 27 currmove h1g1 currmovenumber 13
info depth 27 currmove h5h8 currmovenumber 14
info depth 27 currmove e3c1 currmovenumber 15
info depth 27 currmove f2f4 currmovenumber 16
info depth 27 currmove h2f3 currmovenumber 17
info depth 27 currmove h5g4 currmovenumber 18
info depth 27 currmove a1c1 currmovenumber 19
info depth 27 currmove a3a4 currmovenumber 20
info depth 27 currmove h2f1 currmovenumber 21
info depth 27 currmove h1f1 currmovenumber 22
info depth 27 currmove a1d1 currmovenumber 23
info depth 27 currmove a1a2 currmovenumber 24
info depth 27 currmove h5e2 currmovenumber 25
info depth 27 currmove e1g1 currmovenumber 26
info depth 27 currmove f2f3 currmovenumber 27
info depth 27 currmove e1f1 currmovenumber 28
info depth 27 currmove h5d1 currmovenumber 29
info depth 27 currmove a1b1 currmovenumber 30
info depth 27 currmove h5f7 currmovenumber 31
info depth 27 seldepth 41 multipv 1 time 24052 nps 23425287 score cp 26 nodes 563425009 tbhits 0 hashfull 114 pv h5f3 f8h8 h2g4 g7f8 g4f6 c6e7 h4h5 e7g8 e1c1 g8f6 e5f6 f8e8 e3f4 e8d7 f4e5 d7c6 g5g6 f5h6 a3a4 a7a5 f3f4 c6b7 c1d2 c8d7 g6g7 h8h7 d1a1 h6f5 f4g4 d7c6 a1a2 b7a7 h1b1 f5h6 g4f4 h6g8 f4g5 h7h6

and after more time

Code: Select all

info depth 42 seldepth 76 multipv 1 time 854389 nps 36479390 score cp 120 nodes 31167589739 tbhits 0 hashfull 999 pv h5f3 f8h8 h2g4 f5h4 f3f6 d8f6 g5f6 g7f8 e3h6 f8e8 h1h4 e8d7 e1e2 d7c7 a1h1 c8d7 g4e3 a7a5 h6f4 h8h4 h1h4 a8g8 h4h7 c6d8 f4h6 c7c6 h6g7 c6b5 h7h1 g8e8 f2f4 b5c6 e2f3 c6b5 e3g4 b5c6 h1a1 c6c7 a1g1 c7c6 g4e3 c6b5 g1a1 b5c6 a1b1 e8g8 f3e2 g8e8 b1h1 c6c7 h1h7 c7c6 e2d2

from that position you should squeeze at least +450cps.

so, your SF has nothing to be proud of, raving again somewhere in the fringe lines...

My asmFish will now hang its head in shame, as it can only reach +1.45.

Here's the latest "fringe line", after 17. Qf3:

Code: Select all

+1.45 17... Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Ke2 Kc7 24. Rah1 Bd7 25. Kd1 a5 26. f3 b5 27. Bd2 Rxh4 28. Rxh4 Rg8 29. Rh7 Be8 30. Be3 a4 31. Nh6 Rf8 32. Ke2 Nd8 33. Ng4 Rg8 34. Bh6 Kb6 35. Bd2 Nc6 36. Kd1 Nd8 37. Bc1 Nc6 38. Ke2 Nd8 39. Kf2 Rf8 40. Nh6 Kc6 41. Ke2 Kb6 42. Kd2 Kc7 43. Ng4 Nc6 44. Ke2 Nd8 45. Bh6 Rg8 46. Bd2 Kb6 47. Kf2 Kc7 48. Nh6 Rf8 49. Ke2 Bd7 50. Ng4 Bc6 51. Be3 (depth 58, 2:27:58)

Even after more than 15 hours of searching, presumably of trillions of nodes, asmFish is unable to "squeeze at least +450 cps" from this position. So either Lyudmil or asmFish has somewhat misunderstood the position. Probably asmFish.

Code: Select all

+1.56 17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. O-O-O Kc7 24. Rdh1 Bd7 25. Be3 Rxh4 26. Rxh4 Rg8 27. f3 Nd8 28. Rh7 Be8 29. Kd1 Ba4 30. Nh6 Rf8 31. Kd2 Kc8 32. f4 Kc7 33. f5 Kd7 34. fxe6+ Kxe6 35. Kc1 Nc6 36. Kb2 Nd8 37. Bc1 Nc6 38. Bd2 Nd8 39. Be3 Nc6 40. Kc1 Nd8 41. Bf4 Nc6 42. Bg3 Nd8 43. Kb2 Nc6 44. Ng4 Nd8 45. Kc1 Kd7 46. Kd1 Rg8 47. Rg7 Rf8 48. Nh6 Ke6 49. Rh7 Kd7 50. Bf4 Ke6 51. Kd2 Bc6 52. Ke1 Ba4 53. Kd1 Nc6 54. Kd2 Nd8 55. Bg5 (depth 66, 15:07:10)

nice 11cps jump-up for just 15 hours.

let it run another 15, and score will reach +11 again.

each 156cps winning score sooner or later reaches 450cps.

the thing that I wonder is why Komodo varies between +100 and -100cps or so on each even/uneven ply?

SF does not do so, but its score also varies a lot.

[zullil]

each 156cps winning score sooner or later reaches 450cps.

the thing that I wonder is why Komodo varies between +100 and -100cps or so on each even/uneven ply?

SF does not do so, but its score also varies a lot.

Well, since the position is likely winning for White, the evaluation should eventually reach +128.xx (i.e., the score for a won 6-man position). I'm not going to wait for that! My guess is that there are enough defensive resources for Black that +4.50 is overly-optimistic here.

I'll switch to the latest Komodo and let it run for a bit, but I expect to see an evaluation and line quite similar to what asmFish gave.

[zullil]

[d]r1b4r/p4pk1/1pn1p3/3pP1qP/2pP4/P1P1PQ2/2P4N/2KR3R b - - 0 4



the thing that I wonder is why Komodo varies between +100 and -100cps or so on each even/uneven ply?

SF does not do so, but its score also varies a lot.

I did not see any such oscillation of scores using the latest Komodo. The evaluation was negative, then 0.00 for a while, and then positive once Qf3 was found. Here's depth 40:

Code: Select all

+1.32  17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Rh5 Bb7 24. Rg5 Rh7 25. Rg7 Rah8 26. Ke2 Nd8 27. Rxh7 Rxh7 28. Bf8 Bc6 29. Bd6 Rh8 30. Ne3 Kc8 31. Rg1 Be8 32. Rg7 Rh4 33. Kf3 Rh1 34. Rg8 Kd7 35. Ke2 Rh2 36. Ng4 Rh3 37. f3 Nc6 38. Kf2 a6 39. Kg2 Rh5 40. Ne3 Na7 41. a4 Rh4 (depth 40, 0:16:34)

Isn't this the same line given by asmFish, at least for a dozen plies or so?

[zullil]

This time with the correct diagram.

[d] r1bq1r2/p4pk1/1pn1p3/3pPnPQ/2pP3P/P1P1B3/2P2P1N/R3K2R w KQ - 1 17


the thing that I wonder is why Komodo varies between +100 and -100cps or so on each even/uneven ply?

SF does not do so, but its score also varies a lot.

I did not see any such oscillation of scores using the latest Komodo. The evaluation was negative, then 0.00 for a while, and then positive once Qf3 was found. Here's depth 40:

Code: Select all

+1.32  17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Rh5 Bb7 24. Rg5 Rh7 25. Rg7 Rah8 26. Ke2 Nd8 27. Rxh7 Rxh7 28. Bf8 Bc6 29. Bd6 Rh8 30. Ne3 Kc8 31. Rg1 Be8 32. Rg7 Rh4 33. Kf3 Rh1 34. Rg8 Kd7 35. Ke2 Rh2 36. Ng4 Rh3 37. f3 Nc6 38. Kf2 a6 39. Kg2 Rh5 40. Ne3 Na7 41. a4 Rh4 (depth 40, 0:16:34)

Isn't this the same line given by asmFish, at least for a dozen plies or so?

[Lyudmil Tsvetkov]

[d]r1b4r/p4pk1/1pn1p3/3pP1qP/2pP4/P1P1PQ2/2P4N/2KR3R b - - 0 4



the thing that I wonder is why Komodo varies between +100 and -100cps or so on each even/uneven ply?

SF does not do so, but its score also varies a lot.

I did not see any such oscillation of scores using the latest Komodo. The evaluation was negative, then 0.00 for a while, and then positive once Qf3 was found. Here's depth 40:

Code: Select all

+1.32  17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Rh5 Bb7 24. Rg5 Rh7 25. Rg7 Rah8 26. Ke2 Nd8 27. Rxh7 Rxh7 28. Bf8 Bc6 29. Bd6 Rh8 30. Ne3 Kc8 31. Rg1 Be8 32. Rg7 Rh4 33. Kf3 Rh1 34. Rg8 Kd7 35. Ke2 Rh2 36. Ng4 Rh3 37. f3 Nc6 38. Kf2 a6 39. Kg2 Rh5 40. Ne3 Na7 41. a4 Rh4 (depth 40, 0:16:34)

Isn't this the same line given by asmFish, at least for a dozen plies or so?

well, I used an older version with very limited searches.

probably, with smaller depth, due to the complexity of the position and the peculiar pruning techniques starting from different plies, would fail low and then fail high.

I don't see why queens should be changed and the h4 pawns given away.

interesting, what engines think of that one, after h5 instead of Ng4:

[d]r1bq3r/p4pk1/1pn1p3/3pPnPP/2pP4/P1P1BQ2/2P2P1N/R3K2R b KQ - 0 2

[zullil]

This time with the correct diagram.

[d] r1bq1r2/p4pk1/1pn1p3/3pPnPQ/2pP3P/P1P1B3/2P2P1N/R3K2R w KQ - 1 17


the thing that I wonder is why Komodo varies between +100 and -100cps or so on each even/uneven ply?

SF does not do so, but its score also varies a lot.

I did not see any such oscillation of scores using the latest Komodo. The evaluation was negative, then 0.00 for a while, and then positive once Qf3 was found. Here's depth 40:

Code: Select all

+1.32  17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Kd7 23. Rh5 Bb7 24. Rg5 Rh7 25. Rg7 Rah8 26. Ke2 Nd8 27. Rxh7 Rxh7 28. Bf8 Bc6 29. Bd6 Rh8 30. Ne3 Kc8 31. Rg1 Be8 32. Rg7 Rh4 33. Kf3 Rh1 34. Rg8 Kd7 35. Ke2 Rh2 36. Ng4 Rh3 37. f3 Nc6 38. Kf2 a6 39. Kg2 Rh5 40. Ne3 Na7 41. a4 Rh4 (depth 40, 0:16:34)

Isn't this the same line given by asmFish, at least for a dozen plies or so?

Now Komodo is just a bit more optimistic:

Code: Select all

+1.66 17. Qf3 Rh8 18. Ng4 Nxh4 19. Qf6+ Qxf6 20. gxf6+ Kf8 21. Bh6+ Ke8 22. Rxh4 Ba6 23. Rh5 Kd7 24. Rg5 Rh7 25. Rg7 Rah8 26. Kd2 Nd8 27. Rxh7 Rxh7 28. Bf8 Bb5 29. Rg1 Rh8 30. Be7 Nc6 31. Bd6 Nd8 32. Ne3 Kc8 33. Rg7 Be8 34. f4 Rh2+ 35. Kd1 Rh8 36. Ke1 Kd7 37. Kf2 Kc6 38. Be7 Nb7 39. Kg2 Kb5 40. Kg3 Kc6 41. Bb4 a5 42. Be7 a4 43. Kg2 Na5 44. f5 Kd7 45. Kf3 Nc6 46. Bd6 Na7 47. fxe6+ Kxe6 48. Kf2 (depth 50, 2:04:03)

[mbabigian]

Very interesting. The IDeA computers are working on something else at the moment, but they should finish in 4 days. I'll point them over to this variation at that time and let them spend a few days. Advantages in the early opening rarely hold up over the long-term. A static eval of the position may very well reach 70 or 80 cp for the initial board position. It will be interesting to see if extending the lines confirms that assessment or black squeaks out with a defense.

Thanks for your insights and commentary.
Mike.