Leisebein-Fels, 2004

[jdart]

This position is from a fairly recent correspondence game:

[d] rn1q1rk1/pp2bppp/2p1p3/6P1/1n1PN3/3B1Q1P/PPP5/R1B2RK1 w - -

Here White played Nf6+ with the followup .. gxf6 Bxh7+.

Both the initial sac and the followup are quite difficult. Shredder selects Nf6+ rapidly but expects 14. gxf6 instead of 14. Bxh7+. Rybka 3 did not like Nf6+ and preferred a3, at least at the depths I tried. (But note: I have not done a really deep search).

Full game follows:

[Event "CM.2004.0.00256"]
[Site "IECG"]
[Date "2004.09.15"]
[Round "?"]
[White "Leisebein, Peter"]
[Black "Fels, Bernhard"]
[Result "1-0"]
[ECO "D00"]

1. d4 Nf6 2. f3 d5 3. e4 dxe4 4. Nc3 exf3 5. Nxf3 Bg4 6. h3 Bxf3 7. Qxf3 c6 8.
g4 e6 9. g5 Nd5 10. Bd3 Be7 11. O-O O-O 12. Ne4 Nb4 13. Nf6+ gxf6 14. Bxh7+ Kg7
15. Qh5 fxg5 16. c3 Nd5 17. h4 Nf6 18. Rxf6 Bxf6 19. hxg5 Bxd4+ 20. cxd4 Qxd4+
21. Kg2 f6 22. Qh6+ Kf7 23. Bg6+ Ke7 24. Qg7+ Kd6 25. Qxf8+ 1-0

[ernst]

At first glance it seems that 16...Nd5 threw the draw away.

Edit: No i have to recall this; it is not that simple. Interresting game!

[ernst]

12...Nb4 is the error, leaving the black king without enough protection.

Analysis by Rybka 3:

13.Nf6+ gxf6
=/+ (-0.34) Depth: 11 00:00:00 0kN
13.Nf6+ gxf6
=/+ (-0.34) Depth: 12 00:00:00 0kN
13.Nf6+ gxf6
-/+ (-0.81) Depth: 13 00:00:11 2823kN
13.Nf6+ gxf6 14.gxf6[] Bxf6 15.Qxf6[] Qxf6 16.Rxf6[] Nd7 17.Rf3 f5 18.Bf4 Kf7 19.c4 Rad8 20.Bb1 Rg8+ 21.Kf2 Nb6 22.a3
-/+ (-0.75) Depth: 14 00:00:18 4533kN
13.Nf6+ gxf6 14.gxf6[] Bxf6 15.Qxf6[] Qxf6 16.Rxf6[] Nd7 17.Rf3 f5 18.Bf4 Kf7 19.c4 Rad8 20.Bb1 Rg8+ 21.Kf2 Nb6 22.a3
-/+ (-0.75) Depth: 15 00:00:26 6621kN
13.Nf6+ gxf6 14.gxf6[] Bxf6 15.Qxf6[] Qxf6 16.Rxf6[] Nd7 17.Rf3 f5 18.Bf4 Kf7 19.c4 Nxd3 20.Rxd3 Rad8 21.Rad1 Nf6 22.Kf2 Rg8 23.Ra3
-/+ (-0.72) Depth: 16 00:00:45 11885kN
13.Nf6+
=/+ (-0.52 !) Depth: 17 00:03:25 48929kN
13.Nf6+
=/+ (-0.32 !) Depth: 17 00:04:51 67090kN
13.Nf6+
= (0.08 !) Depth: 17 00:11:27 158mN
13.Nf6+
+/- (0.88 !) Depth: 17 00:18:33 255mN
13.Nf6+ gxf6 14.Bxh7+[] Kg7 15.Qh5[] fxg5 16.c3[] Nd7 17.h4[] Nf6 18.Rxf6[] Bxf6 19.hxg5[] Bxd4+ 20.Kf1[] f6 21.g6[] Be3[] 22.Bxe3 Qd3+[] 23.Kf2 Qf5+ 24.Qxf5 exf5 25.cxb4[] Rad8 26.Rh1 Rfe8 27.Bh6+ Kh8 28.Rh5
+/- (1.13) Depth: 17 00:25:32 349mN

(, 01.01.2009)

[Dr.Wael Deeb]

12...Nb4 is the error, leaving the black king without enough protection.
Analysis by Rybka 3:

13.Nf6+ gxf6
=/+ (-0.34) Depth: 11 00:00:00 0kN
13.Nf6+ gxf6
=/+ (-0.34) Depth: 12 00:00:00 0kN
13.Nf6+ gxf6
-/+ (-0.81) Depth: 13 00:00:11 2823kN
13.Nf6+ gxf6 14.gxf6[] Bxf6 15.Qxf6[] Qxf6 16.Rxf6[] Nd7 17.Rf3 f5 18.Bf4 Kf7 19.c4 Rad8 20.Bb1 Rg8+ 21.Kf2 Nb6 22.a3
-/+ (-0.75) Depth: 14 00:00:18 4533kN
13.Nf6+ gxf6 14.gxf6[] Bxf6 15.Qxf6[] Qxf6 16.Rxf6[] Nd7 17.Rf3 f5 18.Bf4 Kf7 19.c4 Rad8 20.Bb1 Rg8+ 21.Kf2 Nb6 22.a3
-/+ (-0.75) Depth: 15 00:00:26 6621kN
13.Nf6+ gxf6 14.gxf6[] Bxf6 15.Qxf6[] Qxf6 16.Rxf6[] Nd7 17.Rf3 f5 18.Bf4 Kf7 19.c4 Nxd3 20.Rxd3 Rad8 21.Rad1 Nf6 22.Kf2 Rg8 23.Ra3
-/+ (-0.72) Depth: 16 00:00:45 11885kN
13.Nf6+
=/+ (-0.52 !) Depth: 17 00:03:25 48929kN
13.Nf6+
=/+ (-0.32 !) Depth: 17 00:04:51 67090kN
13.Nf6+
= (0.08 !) Depth: 17 00:11:27 158mN
13.Nf6+
+/- (0.88 !) Depth: 17 00:18:33 255mN
13.Nf6+ gxf6 14.Bxh7+[] Kg7 15.Qh5[] fxg5 16.c3[] Nd7 17.h4[] Nf6 18.Rxf6[] Bxf6 19.hxg5[] Bxd4+ 20.Kf1[] f6 21.g6[] Be3[] 22.Bxe3 Qd3+[] 23.Kf2 Qf5+ 24.Qxf5 exf5 25.cxb4[] Rad8 26.Rh1 Rfe8 27.Bh6+ Kh8 28.Rh5
+/- (1.13) Depth: 17 00:25:32 349mN

(, 01.01.2009)

Exactly,giving the white a big chance for tactical shots over the f6 square
Never the less,it's not an easy move when playing over the board....

[ernst]

Exactly,giving the white a big chance for tactical shots over the f6 square
Never the less,it's not an easy move when playing over the board....

Or even for engines, as both DeepFritz11 and Naum4 stick with 13.a3 after 30 minutes on my 3GHz quad.

[Marc Lacrosse]

Never the less,it's not an easy move when playing over the board....

Not played "over the board"

This is an IECG game (=correspondence one)

Pieter Leisebein is a brilliant correspondence master, one of the most famous exponents of daring unorthodox openings.

Marc

[Dr.Wael Deeb]

Never the less,it's not an easy move when playing over the board....

Not played "over the board"
This is an IECG game (=correspondence one)

Pieter Leisebein is a brilliant correspondence master, one of the most famous exponents of daring unorthodox openings.

Marc

I know Marc,I meant to say especialy if it's played over the board....

[Anil]

This position is from a fairly recent correspondence game:

[d] rn1q1rk1/pp2bppp/2p1p3/6P1/1n1PN3/3B1Q1P/PPP5/R1B2RK1 w - -

Here White played Nf6+ with the followup .. gxf6 Bxh7+.

Both the initial sac and the followup are quite difficult. Shredder selects Nf6+ rapidly but expects 14. gxf6 instead of 14. Bxh7+. Rybka 3 did not like Nf6+ and preferred a3, at least at the depths I tried. (But note: I have not done a really deep search).

Full game follows:

[Event "CM.2004.0.00256"]
[Site "IECG"]
[Date "2004.09.15"]
[Round "?"]
[White "Leisebein, Peter"]
[Black "Fels, Bernhard"]
[Result "1-0"]
[ECO "D00"]

1. d4 Nf6 2. f3 d5 3. e4 dxe4 4. Nc3 exf3 5. Nxf3 Bg4 6. h3 Bxf3 7. Qxf3 c6 8.
g4 e6 9. g5 Nd5 10. Bd3 Be7 11. O-O O-O 12. Ne4 Nb4 13. Nf6+ gxf6 14. Bxh7+ Kg7
15. Qh5 fxg5 16. c3 Nd5 17. h4 Nf6 18. Rxf6 Bxf6 19. hxg5 Bxd4+ 20. cxd4 Qxd4+
21. Kg2 f6 22. Qh6+ Kf7 23. Bg6+ Ke7 24. Qg7+ Kd6 25. Qxf8+ 1-0

A good game showing us that engines are yet to catch up with understanding sacrifices (2-piece sacrifice in this game). Is there any chess engine which gets these 2 moves (13. Nf6+ followed by 14. Bxh7+)?

As few have already pointed out the mistakes in begining section of the game, so such a position might not arise when playing against good chess engines. Still, one more point to add to the list for chess engine developers:
1. Fortress detection
2. Fortress creation
3. Understanding traps and making piece sacrifices for winning a game

[jdart]

It looks like Rybka will play it, given enough time.

But the general problem remains: many, many chess positions in the search tree involve loss of material for inadequate or no compensation; and these can and should be pruned. But some sacs, even of multiple pieces, are sound. It is very hard to choose only these without deep search.

This particular problem is especially hard because realizing White's advantage from the sacs requires very accurate followup play including several "quiet" moves that are not captures or checks.

[ernst]

A good game showing us that engines are yet to catch up with understanding sacrifices (2-piece sacrifice in this game). Is there any chess engine which gets these 2 moves (13. Nf6+ followed by 14. Bxh7+)?

3rd post from above...