Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Test Positions / Alekine 6.Nxf7: maybe humans can't see all the mates & wins
- - By MrKris (***) Date 2020-01-15 18:14 Edited 2020-01-15 18:41
A recent G/7'+2" played here:

Stockfish (6/8HI) 1-0 Lc0 (gtx1650S) 1. e4 Nf6 2. e5 Nd5 {book} 3. d4 {+0.43/28 29s} d6
{-1.07/8 6.2s} 4. Nf3 {+0.73/26 3.5s} dxe5 {-0.94/8 8.1s} 5. Nxe5 {+0.81/24
4.0s} Nd7 {-0.74/8 8.5s} 6. Nxf7 {+0.97/26 4.7s}

6.Nxf7 at ChessTempo.com does worse than 6.Nf3
'all games'
93 6.Nxf7 wh+34.4% ==25.8% bk+39.8%
93 6.Nf3  wh+53.8% ==33.3% bk+12.8% , various others mostly high white %
2400+
27 6.Nf3  wh+48.7% ==44.4% bk+14.8%
17 6.Nxf7 wh+17.6% ==29.4% bk+52.9% , 6 6.Bc4 wh+33% ==67%, 3 various moves


[Event "i7-4770K 6HT  Sy.6 2Ghash"]
[Site "GTX 1650 Super ; 5 rare"]
[Date "2020.01.15"]
[Round "G/7'+2''"]
[White "Stockfish-6ths_20011312c0"]
[Black "Leelenstein-191226_12.1"]
[Result "1-0"]
[ECO "B04"]
[Opening "Alekhine's defense"]
[PlyCount "48"]
[Termination "adjudication"]
[TimeControl "420+2"]
[Variation "Modern, Larsen Variation"]

1. e4 {book} Nf6 {book} 2. e5 {book} Nd5 {book} 3. d4 {+0.43/28 29s} d6
{-1.07/8 6.2s} 4. Nf3 {+0.73/26 3.5s} dxe5 {-0.94/8 8.1s} 5. Nxe5 {+0.81/24
4.0s} Nd7 {-0.74/8 8.5s} 6. Nxf7 {+0.97/26 4.7s} Kxf7 {+0.14/9 9.3s} 7.
Qh5+ {+1.09/26 4.0s} Ke6 {-0.02/11 3.0s} 8. c4 {+1.23/29 16s} N5f6
{-0.90/11 13s} 9. d5+ {+1.38/27 3.7s} Kd6 {-1.10/11 6.2s} 10. Qf7 {+1.32/30
12s} Nb8 {-0.59/11 22s} (10. .. Ne5 {ChessTempo.com 'all games' played in
27 of 36} 11. Bf4 {Sf -all 8 threads- and 27 of 27 human} c5 {Sf and 25 of 27} 12. Nc3 {Sf
and 21 of 25} a6 {Sf 21 of 21} 13. b4 {Sf and only 9 of 21 human} Qb6 {Sf
and 8 of 9} 14. Rc1 {Sf and 5 of 9} g6 {1 of 5 human but 0-1} (14. .. g5 {2
of 5 human} 15. Bg3 {Sf and 2 of 2} h5 {Sf 7' 41 plies -4.31 +-, and 1 of 2
human} (15. .. Bf5 16. Bd3 {Sf and 1 of 2, 1' 35 plies +9.84; and in 2012
Manca 2424 1-0 Sulskis in 16}) 16. bxc5+ Qxc5 17. Qg6 Qd4 18. Qc2 {Sf 3:34
40 plies +4.73, with:} (18. c5+ Kc7 19. Qxg5 Nd7 {in 1992, Szilagyi unr.
0-1 Krantz 2305 in 29}) 18. .. h4 19. c5+ Kd7 20. Rd1 Qxc5 21. Bxe5 {etc.})
15. Be2 Qc7 16. bxc5+ {Sf 8ths 3' 39 plies +4.66} (16. Na4 {in 1997, Prie
2470 0-1 Nataf 2410, 28 moves}) 16. .. Qxc5 (16. .. Kxc5 17. Na4+ Kd4 18.
Rd1+ Ke4 19. Bxe5 Kxe5 (19. .. Qxe5 20. Nc5+ Kf4 (20. .. Kf5 21. g4+ {Sf
8ths 00:01 38 plies +M12}) 21. Nd3+ {Sf 8ths 0:44 +M15}) 20. O-O {Sf 8ths
1:24 +14.11}) 17. Bxe5+ Kxe5 18. O-O Bh6 19. Na4 Rf8 20. Qxf8 Bxf8 21. Nxc5
Kd6 22. Nb3 b6 23. Rfe1 Bh6 24. Rcd1 {Sf 8ths 1' 35 plies +5.11}) 11. c5+
{+3.73/27 4.3s} Kd7 {-1.69/12 7.5s} 12. Bb5+ {+3.92/28 7.8s} c6 {-2.85/13
6.5s} 13. dxc6+ {+4.23/27 5.8s} bxc6 {-3.11/11 6.7s} 14. O-O {+3.53/31 16s
----Diagram----} Qa5 {-3.26/11 13s} (14. .. Kc7 15. Bf4+ {Sf 1' 35 plies
+8.88 with:} Kb7 16. Qb3 a6 17. Rd1 Qxd1+ 18. Qxd1 Bd7 19. Be2 e5 20. Qb3+
Kc8 21. Bxe5 Bxc5 22. Nd2 Re8 23. Qc3 Bb4 24. Qxb4 Rxe5 25. Bf3 Re6 26. Nc4
{etc.}) 15. Rd1+ {+3.91/28 4.4s} Kc7 {-3.90/11 5.7s} 16. Bf4+ {+4.15/30
10s} Kb7 {-4.64/9 0.10s} 17. Be2 {+4.35/28 4.5s} Qb4 {-4.47/8 33s} 18. Bd2
{+4.72/29 16s} Qxc5 {-4.34/9 7.0s} 19. Be3 {+4.90/29 10s} Qa5 {-4.25/8 18s}
20. Qb3+ {+4.76/27 4.7s} Kc7 {-4.67/9 5.8s} 21. Na3 {+4.98/29 4.6s} e5
{-4.84/9 10s} 22. Nb5+ {+7.27/30 5.0s} cxb5 {-7.63/9 20s} 23. Rac1+
{+11.07/29 5.9s} Nc6 {-11.59/8 12s} 24. Bxb5 {+16.53/28 6.6s} Bd7 {-13.38/7
9.7s, White wins by adjudication} 1-0


Actual game continuation after "--Diagram--" above: 11. c5+ {+3.73/27 4.3s, CT only 3 games} Kd7 {-1.69/12 7.5s, same 3} 12. Bb5+ {+3.92/28 7.8s, same 3} c6 {-2.85/13 6.5s, same 3} 13. dxc6+ {+4.23/27 5.8s, same 3} bxc6 {-3.11/11 6.7s, same 3} 14. O-O {+3.53/31 16s, same 3}
rnbq1b1r/p2kpQpp/2p2n2/1BP5/8/8/PP3PPP/RNB2RK1 b - - 1 14

1980 Hammar 2385 0-1 Cramling 2330, 41: 14...Qa5 15.Bc4 --Sf played 15.Rd1+ +3.91/28-- Kc7
1979 Ragelis 2315 ½-½ Kaumas 2290, 19: 14...Kc7 15.Bf4+  --Sf +8.88, see above-- 15.Bf4+ Kb7 16.Ba6+? (Sf 16.Qb3) Kxa6 17.Qc4+ Kb7 18.Qb4+ Ka7 19.Qc4+.
2005 Mutschnik 2408 1-0 Pitschka 2212, 27: 14...Qa5 15.Bc4 --Sf played 15.Rd1+ +3.91/28-- Kd8
Parent - - By Peter Grayson (****) Date 2020-01-15 23:20

> Alekine 6.Nxf7: maybe humans can't see all the mates & wins


Perhaps they do and that is why 5..Nd7 has only been played in three games in the last 10 years by players 2300+
6..c6 and 5..g6 are the preferred Black moves.

The fact the NN engine chose a tactically weak move confirms NN engines are still susceptable to the tactical sucker punch. A quick check with Stockfish 130120 64 default settings confirms it would only consider c6 or g6, settling on c6 in the game context.

There is nothing to commend Black playing 5..Nd7??

Peter
Parent - - By MrKris (***) Date 2020-01-16 05:10

> 6..c6 and 5..g6 are the preferred Black moves.
> There is nothing to commend Black playing 5..Nd7??


Yes, ChessTempo.com both 2700+ 5...c6 19 games, 5...g6 2, no other moves.

> The fact the NN engine chose a tactically weak move confirms NN engines are still susceptible to the tactical sucker punch.


Definitely, even if the NN loses fewer in a match.


I missed (thinking 6.Nf3 still popular) that both 2300+, CT, after 5...Nd7? in the last 10 years all where with 6.Nxf7.

However these where 8 games, white +2, black +1, 5 draws, white 56.25%:
Parent - By Peter Grayson (****) Date 2020-01-17 14:59 Edited 2020-01-17 15:02
Thanks for the extra games information. Reading down the list I could only find games 2,6 and 7 in my reference database. I also have 5..e6 and 5..Bf5 have been played.

http://www.mediafire.com/file/6cj5cqqnmgafvf0/Alekhines_5_e6_and_Bf5.zip/file

Peter
Parent - By user923005 (****) Date 2020-01-17 22:42
There are more mistakes after these, but these are two important false steps.  Already down a pawn, Black played Nb8, which loses the game here (Ne5 is far better):
r1bq1b1r/pppnpQpp/3k1n2/3P4/2P5/8/PP3PPP/RNB1KB1R b KQ - acd 36; acs 64; bm Ne5; am Nb8; cce -203; ce -112;
pm Ne5 {41} Nb8 {15} Nb6 {6}; pv Ne5 Bf4 g5 c5+ Kxc5 Bxe5 Qxd5 Qxd5+ Kxd5 Bxc7 Bd7 Nc3+ Ke6 Bc4+ Kf5 Bb3 Rc8 Bg3 Kg6 O-O Bf5 Rfe1 Rd8 f3 h5 Bf2 a6 Re5 Rd2 Rae1 Nd7 R5e2 Rxe2 Nxe2 e6 Ng3 Bc5 Bxe6 Bxe6 Rxe6+ Kf7 Re1 Bxf2+ Kxf2 Nf6 Rc1 Rd8 Rc7+ Ke6 Rxb7 Rd2+ Kg1 Nd7 Nxh5; white_wins 645; black_wins 238; draws 151;

Not as bad, but black played  Nd7 here and c6 is better:
rnbqkb1r/ppp1pppp/8/3nN3/3P4/8/PPP2PPP/RNBQKB1R b KQkq - acd 36; acs 900; bm c6; am Nd7; cce -113; ce -40;
pm c6 {793} g6 {414} Nd7 {182} e6 {26} Bf5 {6} c5 {1} Nb6 {1} Nc6 {1} Nf6 {1}; pv c6 Be2 Nd7 Nf3 g6 c4 Nc7 Nc3 Bg7 Bf4 O-O O-O b5 cxb5 cxb5 Nxb5 Nxb5 Bxb5 Bb7 Re1 Rc8 Qe2 Nf6 Be5 e6 Bd3 Nd5 Bxg7 Kxg7 Be4 Nf4 Qd2 Bxe4 Rxe4 Nd5 Ne5 Qc7 Re2 Rfe8 h3 Red8 a3 f6 Nd3; white_wins 9734; black_wins 4198; draws 8532;
Attachment: ga.epd - The analyzed moves of the game (for each position, c3 is the move actually played) (16k)
Up Topic The Rybka Lounge / Test Positions / Alekine 6.Nxf7: maybe humans can't see all the mates & wins

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill