Engine exchange sacrifice

[AdminX]

I came across an interesting read on Chessbase.com about Komodo. In it the writer shows a positional exchange sacrifice Komodo plays in a endgame versus Hannibal.

[d]r7/3br1kp/2p2pp1/1pRp4/pP1P1P1P/4P3/P4PB1/2R3K1 w - - 0 1

https://en.chessbase.com/post/komodo-11-review

By chance I came across the following position in a game Houdini(Tactical) played versus Critter

[d]r1b2rk1/pp1nqpbp/4pnp1/2Pp4/2P2B2/5NP1/PPQNPPBP/R4RK1 b - - 0 11

Here Critter chose 11. ... e5 instead of regaining it's pawn. Thus allowing Houdini to chose the positional exchange sacrifice 12. Nxe5 Nxe5 13. cxd5

[d]r1b2rk1/pp2qpbp/5np1/2PPn3/5B2/6P1/PPQNPPBP/R4RK1 b - - 0 13

From here the pawns proved to be to much for the pieces.

[pgn]

[Event "40 Moves in 2 min; 20 Moves in 1 min; Re"]
[Site "Engine Match"]
[Date "2017.10.13"]
[Round "6"]
[White "Houdini 6.02 Pro x64 Pext [3 T"]
[Black "Critter 1.6a 64-bit \[3\]"]
[Result "1-0"]
[ECO "A07"]
[PlyCount "117"]
[EventDate "2017.??.??"]

1. Nf3 {book 0s} d5 {book 0s} 2. g3 {book 0s} g6 {book 0s} 3. c4 {book 0s} c6 {
book 0s} 4. Qa4 {book 0s} e6 {book 0s} 5. d4 {book 1s} Nd7 {[%eval 27,18]
[%emt 0:00:03]} 6. Qc2 {[%eval 65,16] [%emt 0:00:03]} Ngf6 {[%eval 29,18]
[%emt 0:00:00] (Ne7)} 7. Bg2 {[%eval 58,16] [%emt 0:00:03]} Bg7 {[%eval 27,19]
[%emt 0:00:00] (b5)} 8. Bf4 {[%eval 61,17] [%emt 0:00:03]} O-O {[%eval 33,19]
[%emt 0:00:00]} 9. O-O {[%eval 67,15] [%emt 0:00:03] (Nbd2)} Qe7 {[%eval 25,19]
[%emt 0:00:22] (b5)} 10. Nbd2 {[%eval 60,17] [%emt 0:00:06] (Nc3)} c5 {[%eval
5,17] [%emt 0:00:03] (h6)} 11. dxc5 {[%eval 15,17] [%emt 0:00:06]} e5 {[%eval
10,18] [%emt 0:00:02] (Nxc5)} 12. Nxe5 {[%eval 65,16] [%emt 0:00:02] (Bg5)}
Nxe5 {[%eval -28,16] [%emt 0:00:03]} 13. cxd5 {[%eval 74,19] [%emt 0:00:00]}
Nh5 {[%eval -25,18] [%emt 0:00:04] (Bf5)} 14. Bxe5 {[%eval 80,18] [%emt 0:00:
05] (d6)} Qxe5 {[%eval -18,18] [%emt 0:00:02]} 15. Nc4 {[%eval 91,18] [%emt 0:
00:02]} Qc7 {[%eval -22,19] [%emt 0:00:00]} 16. Nd6 {[%eval 99,18] [%emt 0:00:
02] (Ne3)} Be5 {[%eval -22,17] [%emt 0:00:03] (Rb8)} 17. Nxc8 {[%eval 129,18]
[%emt 0:00:02]} Raxc8 {[%eval -12,17] [%emt 0:00:00] (Rfxc8)} 18. Rac1 {
[%eval 64,19] [%emt 0:00:07]} Qa5 {[%eval -3,19] [%emt 0:00:00] (b6)} 19. a3 {
[%eval 125,18] [%emt 0:00:03] (Rfd1)} Nf6 {[%eval -6,17] [%emt 0:00:02]} 20.
Rfd1 {[%eval 117,22] [%emt 0:00:00] (e3)} Nd7 {[%eval 0,18] [%emt 0:00:02]} 21.
b4 {[%eval 86,22] [%emt 0:00:06] (c6)} Qxa3 {[%eval 11,17] [%emt 0:00:02]} 22.
Rb1 {[%eval 70,21] [%emt 0:00:05]} Rfd8 {[%eval 6,19] [%emt 0:00:00] (Nxc5)}
23. Bh3 {[%eval 128,16] [%emt 0:00:02] (d6)} Qa6 {[%eval 20,17] [%emt 0:00:03]
(Kh8)} 24. e4 {[%eval 137,16] [%emt 0:00:02] (f4)} Bg7 {[%eval 30,16] [%emt 0:
00:02]} 25. b5 {[%eval 127,19] [%emt 0:00:03]} Qd6 {[%eval 42,19] [%emt 0:00:
00]} 26. c6 {[%eval 129,19] [%emt 0:00:02]} Rc7 {[%eval 41,19] [%emt 0:00:00]
(f5)} 27. f4 {[%eval 169,16] [%emt 0:00:02]} f6 {[%eval 41,19] [%emt 0:00:01]}
28. Rbc1 {[%eval 133,21] [%emt 0:00:10] (Be6+)} b6 {[%eval 40,19] [%emt 0:00:
04]} 29. Bxd7 {[%eval 125,21] [%emt 0:00:00] (Be6+)} Rdxd7 {[%eval 40,18]
[%emt 0:00:02]} 30. Qd2 {[%eval 97,20] [%emt 0:00:10] (Kg2)} Rd8 {[%eval 39,21]
[%emt 0:00:05]} 31. h4 {[%eval 102,20] [%emt 0:00:02] (Kg2)} Re8 {[%eval 29,18]
[%emt 0:00:05] (h6)} 32. Rc4 {[%eval 98,18] [%emt 0:00:07] (Qd3)} Rd8 {[%eval
33,19] [%emt 0:00:05] (h6)} 33. Kg2 {[%eval 119,16] [%emt 0:00:01]} Bf8 {
[%eval 33,19] [%emt 0:00:02] (Re8)} 34. Qc3 {[%eval 123,17] [%emt 0:00:01]
(Qd4)} Re7 {[%eval 50,19] [%emt 0:00:13] (Bg7)} 35. Rd3 {[%eval 121,18] [%emt
0:00:03] (Qb4)} Kf7 {[%eval 60,17] [%emt 0:00:02] (Rc8)} 36. Rd2 {[%eval 167,
15] [%emt 0:00:00] (h5)} Rc8 {[%eval 52,18] [%emt 0:00:01]} 37. g4 {[%eval 190,
19] [%emt 0:00:00] (h5)} Ree8 {[%eval 71,14] [%emt 0:00:02] (a5)} 38. Qf3 {
[%eval 336,15] [%emt 0:00:00] (Qg3)} Kg8 {[%eval 133,14] [%emt 0:00:02]} 39. g5
{[%eval 316,20] [%emt 0:00:00]} Be7 {[%eval 179,14] [%emt 0:00:02]} 40. gxf6 {
[%eval 444,20] [%emt 0:00:00]} Bxf6 {[%eval 211,14] [%emt 0:00:01] (Qxf6)} 41.
Qg3 {[%eval 454,18] [%emt 0:00:02] (e5)} Qd8 {[%eval 185,15] [%emt 0:00:04]
(g5)} 42. d6 {[%eval 582,19] [%emt 0:00:06] (h5)} Bg7 {[%eval 293,14] [%emt 0:
00:02] (a6)} 43. Qg4 {[%eval 884,18] [%emt 0:00:02] (d7)} Re7 {[%eval 456,14]
[%emt 0:00:02] (h5)} 44. e5 {[%eval 1229,19] [%emt 0:00:07] (Qxc8)} Rf7 {
[%eval 555,15] [%emt 0:00:10]} 45. Rcc2 {[%eval 1645,21] [%emt 0:00:04] (e6)}
Rf5 {[%eval 643,13] [%emt 0:00:01] (g5)} 46. c7 {[%eval 1924,19] [%emt 0:00:04]
} Qd7 {[%eval 1316,16] [%emt 0:00:01]} 47. e6 {[%eval 1977,18] [%emt 0:00:00]}
Qxe6 {[%eval 1398,16] [%emt 0:00:03]} 48. d7 {[%eval 2135,17] [%emt 0:00:00]}
Bf8 {[%eval 1396,16] [%emt 0:00:03] (Rcf8)} 49. dxc8=Q {[%eval 2255,15] [%emt
0:00:00]} Qxc8 {[%eval 1491,17] [%emt 0:00:02]} 50. Rd8 {[%eval 32746,31]
[%emt 0:00:00]} Qe6 {[%eval 2304,17] [%emt 0:00:05] (Rc5)} 51. c8=Q {[%eval
32750,25] [%emt 0:00:00] (Rxf8+)} Qe4+ {[%eval 32752,18] [%emt 0:00:02] (Qf7)}
52. Qf3 {[%eval 32752,23] [%emt 0:00:00]} Qxf3+ {[%eval 32754,26] [%emt 0:00:
01]} 53. Kxf3 {[%eval 32754,48] [%emt 0:00:00]} Rf7 {[%eval 32756,29] [%emt 0:
00:02] (Kg7)} 54. Rxf8+ {[%eval 32756,20] [%emt 0:00:00] (Rc7)} Rxf8 {[%eval
32758,39] [%emt 0:00:01]} 55. Qe6+ {[%eval 32758,85] [%emt 0:00:00]} Kg7 {
[%eval 32760,64] [%emt 0:00:00]} 56. Rc7+ {[%eval 32760,99] [%emt 0:00:00]} Kh6
{[%eval 32762,64] [%emt 0:00:00]} 57. Qe5 {[%eval 32762,99] [%emt 0:00:00]} Rf7
{[%eval 32764,64] [%emt 0:00:00] (Rf5)} 58. Rxf7 {[%eval 32764,12] [%emt 0:00:
00]} a5 {[%eval 32766,64] [%emt 0:00:00] (g5)} 59. Qg5# {[%eval 32766,11]
[%emt 0:00:00]} 1-0

[/pgn]

[Nordlandia]

Is it just me or i got this feeling that exchange sacrifices has become more normalized.

Chess masters in the past did seldomly sacrifice the exchange. Nowdays it's treated as almost normal.

[Nordlandia]

Here is an example of a human exchange sacrifice.

14. Ra2!?

[d]r2qr1k1/p1pn1ppp/1p1p1n2/8/2bPP3/P1P1PQN1/R5PP/1BB1K2R b K - 0 14

A brilliant concept, turning the game around. It may not lead to victory, but it gives White an excellent fighting chance. By sacrificing the exchange, White kills many birds with one stone: he can castle, his light-squared bishop dominate on the diagonal a2-g8, and his pawn centre becomes more fluid. Only his bishop on c1 is a little bit sleepy.

[pgn][Event "11th Open"]
[Site "Chicago USA"]
[Date "2002.05.27"]
[Round "7"]
[White "Gregory Kaidanov"]
[Black "Alexander Onischuk"]
[Result "1-0"]
[ECO "E28"]
[WhiteElo "2616"]
[BlackElo "2641"]
[SetUp "1"]
[FEN "r2qr1k1/p1pn1ppp/1p1p1n2/8/2bPP3/P1P1PQN1/6PP/RBB1K2R w KQ - 0 14"]
[PlyCount "55"]
[EventDate "2002.??.??"]

14. Ra2 Bxa2 15. Bxa2 Nf8 16. O-O Ng6 17. Bd2 Qd7 18. Nf5 Qd8 19. e5 dxe5 20.
e4 exd4 21. Nxg7 Ne5 22. Qh3 dxc3 23. Bxc3 Kxg7 24. Qg3+ Ng6 25. Bxf6+ Qxf6 26.
Rxf6 Kxf6 27. h4 Re5 28. Qf3+ Kg7 29. h5 f6 30. hxg6 hxg6 31. Qd3 Rae8 32. Qc4
R5e7 33. Qc6 Rxe4 34. Qxc7+ R8e7 35. Qb8 Kh6 36. Qf8+ Kg5 37. Bf7 Rc7 38. Qg7
Rc1+ 39. Kf2 Rc2+ 40. Kf3 Rf4+ 41. Ke3 1-0

[/pgn]

[MikeB]

I came across an interesting read on Chessbase.com about Komodo. In it the writer shows a positional exchange sacrifice Komodo plays in a endgame versus Hannibal.

[d]r7/3br1kp/2p2pp1/1pRp4/pP1P1P1P/4P3/P4PB1/2R3K1 w - - 0 1

https://en.chessbase.com/post/komodo-11-review

...[snip]

[/pgn]

K11.2 fins Rxc6 about 1 second faster than SF-McB 3.0 on my machine ( released date tomorrow 10/14/2017)

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 32	+1.20 	1.13G	0:45.36	Rxc6 Bxc6 Rxc6 Ree8 Bxd5 Rac8 Kg2 Red8 a3 Kh6 Be4 Re8 Bf3 Red8 Rxf6 Rc3 Rb6 Rxa3 Rxb5 Rb3 Ra5 Rxb4 Bc6 Rdb8 d5 Rb2 Rxa4 Rd2 Kf3 Rbb2 Ke4 Rd1 Ke5 Rxf2 
 32	+1.16?	1.09G	0:43.75	Rxc6 Bxc6? 
 31	+1.23 	868.6M	0:35.22	Rxc6 Bxc6 Rxc6 Ree8 Bxd5 Rac8 Kg2 Rxc6 Bxc6 Re6 Bxb5 Rb6 Bxa4 Rxb4 Bb3 Kf8 Kf1 Ke7 Ke2 Rb5 d5 Kd6 Kd3 Rb4 f3 h6 e4 Kc5 Ke3 Kd6 Ke2 Kc5 Kd3 h5 Ke3 Kd6 Ke2 Kc5 Kd3 
 31	+1.21!	619.9M	0:24.95	Rxc6! 
 30	+1.07 	452.6M	0:17.91	Rxc6 Bxc6 Rxc6 Ree8 Bxd5 Rac8 Rb6 Rc2 Rxb5 Rd8 Kg2 Kf8 Kf3 Ke7 h5 gxh5 f5 Rc3 Kg2 Rd7 Be6 Ra7 Rb8 Rc2 b5 Rb2 e4 Rc7 e5 fxe5 dxe5 Rcc2 Rb7+ Ke8 f6 Rxf2+ Kh3 Rh2+ Kg3 Rhd2 Rb8+ Rd8 Rxd8+ Kxd8 
 30	+1.02!	236.8M	0:09.14	Rxc6! 
 30	+0.91!	230.7M	0:08.91	Rxc6! 
 29	+0.80 	225.0M	0:08.69	Rxc6 Bxc6 Rxc6 Ree8 Bxd5 Rec8 Kg2 Rxc6 Bxc6 Ra6 Bxb5 Rb6 Bxa4 Rxb4 Bb3 Kf8 Kf3 Ke7 Ke4 Rb5 d5 Rb4+ Kd3 Kd6 e4 Kc5 f3 h6 Kd2 Rb8 Ke3 Re8 Kd3 h5 a4 Rf8 Bc4 g5 fxg5 fxg5 hxg5 Rxf3+ Ke2 Ra3 
 29	+1.02!	223.2M	0:08.63	Rxc6! 
 29	+0.91!	217.7M	0:08.42	Rxc6! 
 28	+0.80 	200.9M	0:07.78	Rxc6 Bxc6 Rxc6 Ree8 Bxd5 Rec8 Kg2 Rxc6 Bxc6 Ra6 Bxb5 Rb6 Bxa4 Rxb4 Bb3 Kf8 Kf3 Ke7 Ke4 Kd6 Bg8 h6 Bb3 Rb5 Kf3 Rh5 Kg4 Rb5 Kh3 Rb8 Kg3 Rc8 f5 g5 hxg5 hxg5 f4 Rc3 fxg5 fxg5 Kf2 
 28	+0.67?	163.3M	0:06.35	Rxc6 Bxc6? 
 28	+1.10!	150.7M	0:05.88	Rxc6! 
 28	+0.95!	137.6M	0:05.37	Rxc6! 
 28	+0.85!	130.3M	0:05.10	Rxc6! 
 27	+0.74 	118.5M	0:04.64	Rxc6 Bxc6 Rxc6 Ree8 Bxd5 Rec8 Kf1 Rxc6 Bxc6 Ra6 Bxb5 Rb6 Bxa4 Rxb4 Bb3 Rb5 d5 Kf8 Ke2 Ke7 Kd3 Kd6 f3 Kc5 Kc3 Rb4 a4 Rb7 e4 Ra7 e5 fxe5 fxe5 Rf7 d6 Rxf3+ Kb2 Kc6 Bd1 Rf2+ Kc3 
 27	+0.86!	113.0M	0:04.44	Rxc6! 
 27	+0.70!	101.7M	0:04.00	Rxc6! 
 27	+0.59!	99.3M  	0:03.90	Rxc6! 
 27	+0.51!	94.0M  	0:03.69	Rxc6! 
 26	+0.44 	85.9M  	0:03.37	a3 Ree8 Kh2 Rac8 Bf3 h6 Kg3 Kf8 h5 g5 f5 Kf7 Be2 Kg7 Kf3 Kf8 Bf1 Kf7 Kg2 Ke7 Bd3 Kf7 Kh2 Kg7 Be2 Kf7 Bf1 
 26	+0.48?	81.0M  	0:03.17	a3 Ree8? 
 25	+0.55 	68.2M  	0:02.66	a3 Ree8 Kf1 Rac8 Ke2 Kf7 Kd2 Ke7 f3 Kf7 e4 Rcd8 Ke3 Re7 Kf2 dxe4 fxe4 Be6 d5 cxd5 exd5 Bf5 Bf1 Be4 Bxb5 Rxd5 Rxd5 
 24	+0.65 	48.2M  	0:01.87	a3 Ree8 Kh2 h6 Bh3 f5 Kg3 Rac8 Rh1 Rh8 Rcc1 Kf6 Bf1 g5 fxg5+ hxg5 hxg5+ Kxg5 Bd3 Rcg8 Kf3 Kf6 Ra1 Ke6 Rad1 Kf6 Rc1 Ke7 Kf4 Rxh1 Rxh1 
 23	+0.62 	45.0M  	0:01.75	a3 Ree8 Kh2 h6 Bh3 f5 Kg3 Rac8 Rh1 Rh8 Rcc1 Kf6 Bf1 g5 fxg5+ hxg5 hxg5+ Kxg5 Bd3 Rcg8 Kf3 Kf6 Ra1 Ke6 Rad1 Kf6 Rc1 

[Graham Banks]

Is it just me or i got this feeling that exchange sacrifices has become more normalized

Quite frequent in engine v engine games these days, although usually a pawn or two are added bonuses.

[MikeB]

...
[snip]

By chance I came across the following position in a game Houdini(Tactical) played versus Critter

[d]r1b2rk1/pp1nqpbp/4pnp1/2Pp4/2P2B2/5NP1/PPQNPPBP/R4RK1 b - - 0 11

Here Critter chose 11. ... e5 instead of regaining it's pawn. Thus allowing Houdini to chose the positional exchange sacrifice 12. Nxe5 Nxe5 13. cxd5

[d]r1b2rk1/pp2qpbp/5np1/2PPn3/5B2/6P1/PPQNPPBP/R4RK1 b - - 0 13

From here the pawns proved to be to much for the pieces.
...
[snip]

McBrain likes Nxe5 early and sticks with it, but the score never gets convincing even after 45 minutes and 58 billion nodes

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 42	+0.47!	58.1G	46:38.61	Nxe5! 
 42	+0.35!	39.0G	31:28.91	Nxe5! 
 42	+0.27!	36.2G	29:12.92	Nxe5! 
 42	+0.22!	29.1G	23:31.78	Nxe5! 
 41	+0.17 	28.5G	23:04.97	Nxe5 Nxe5 cxd5 Rd8 e4 Bd7 Rfd1 a5 Bxe5 Qxe5 Nc4 Qh5 Nb6 Rab8 h4 h6 Rd2 Bg4 Qb3 Nd7 f3 Nxb6 fxg4 Qxg4 cxb6 Be5 Rd3 Rbc8 Rf1 a4 Qd1 Qxd1 Rfxd1 Rc2 b3 a3 R1d2 Rxd2 Rxd2 Bxg3 Kf1 Rd6 Ke2 Rxb6 Kd3 f5 exf5 gxf5 h5 Kf7 Kc4 Be5 Kc5 Rf6 
 41	+0.27!	24.5G	19:52.97	Nxe5! 
 41	+0.22!	24.3G	19:41.08	Nxe5! 
 40	+0.50!	12.6G	10:15.52	Nxe5! 
 40	+0.35!	11.0G	8:59.00	Nxe5! 
 40	+0.24!	10.5G	8:34.20	Nxe5! 
 40	+0.17!	10.3G	8:28.04	Nxe5! 
 40	+0.13!	10.2G	8:18.96	Nxe5! 
 39	+0.09 	9.68G	7:56.57	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Qe5 Qc3 Re8 b4 Qxe3+ Qxe3 Rxe3 d6 Rxe2 b5 Rd2 c6 bxc6 Rfd1 Rxd1+ Rxd1 Bd7 bxc6 Nf6 cxd7 Rd8 Kf2 Rxd7 Bc6 Ng4+ Ke2 Rd8 Rd4 f5 d7 Nxh2 Ra4 Ng4 Rxa7 Ne5 Bb5 Kf7 a4 Ke7 
 39	+0.25!	7.05G	5:47.34	Nxe5! 
 39	+0.18!	6.86G	5:38.16	Nxe5! 
 39	+0.14!	6.69G	5:29.01	Nxe5! 
 38	+0.10 	6.51G	5:20.09	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Qe5 Qc3 Re8 b4 Qxe3+ Qxe3 Rxe3 d6 Rxe2 b5 Rd2 Rfd1 Rxd1+ Rxd1 Nf6 c6 bxc6 bxc6 Rb8 Bf1 Bd7 cxd7 Nxd7 Bc4 Kg7 Kf2 Ne5 Bb3 Ng4+ Kg1 Rd8 h3 Nf6 Rd4 Nd7 Kg2 Nc5 Bc2 
 38	+0.12?	5.82G	4:44.58	Nxe5 Nxe5? 
 37	+0.20 	5.28G	4:17.71	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Qe5 Qc3 Re8 b4 Qxe3+ Qxe3 Rxe3 d6 Rxe2 b5 Rd2 Rfd1 Rxd1+ Rxd1 Nf6 c6 bxc6 bxc6 Rb8 Bf1 Bd7 cxd7 Nxd7 Kf2 Kf8 a4 Rb6 Bb5 Ke8 Re1+ Kd8 Re7 Rxd6 Rxf7 Rf6+ Rxf6 Nxf6 
 37	+0.09!	5.04G	4:06.02	Nxe5! 
 37	  0.00?	4.96G	4:02.33	Nxe5 Nxe5? 
 37	+0.10?	4.48G	3:38.89	Nxe5 Nxe5? 
 37	+0.23!	4.23G	3:26.97	Nxe5! 
 36	+0.18 	4.18G	3:24.24	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Qe5 Qc3 Re8 Qxe5 Rxe5 b4 Rxe3 d6 Rxe2 Rcd1 Rxg2+ Kxg2 Bd7 Rde1 f5 Re7 Nf6 Rb1 Bc6+ Kh3 Nd5 Re5 Nc3 Rb3 Ne4 b5 Bd7 c6 bxc6 Re7 Be6 d7 Kf8 Rxe6 Ng5+ Kh4 Nxe6 
 36	+0.27?	3.76G	3:04.23	Nxe5 Nxe5? 
 36	+0.35?	3.70G	3:01.26	Nxe5 Nxe5? 
 35	+0.42 	3.54G	2:53.28	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Bd7 b4 Qe5 Qc3 Qxc3 Rxc3 a5 a3 axb4 axb4 Ra6 c6 bxc6 dxc6 Bg4 c7 Rb6 Bf3 Be6 Rd1 Rxb4 Rd8 Bc8 Rcd3 Be6 Rxf8+ Kxf8 Rd8+ Ke7 c8=Q Bxc8 Rxc8 Nf6 Kf2 
 35	+0.45!	3.42G	2:47.43	Nxe5! 
 35	+0.36!	3.15G	2:34.39	Nxe5! 
 35	+0.30!	2.92G	2:23.26	Nxe5! 
 34	+0.25 	2.58G	2:06.92	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Bd7 b4 Rae8 d6 Qb8 a4 a6 Qc4 Bc6 Bf3 Rxe3 Bxh5 gxh5 b5 Bd7 c6 bxc6 Qc5 Re6 bxc6 Bxc6 Qxc6 Qxd6 Qxd6 Rxd6 Rc7 f6 Rf5 Re8 Kf2 Rd2 Rxh5 Rexe2+ Kf3 Rf2+ Kg4 Rxh2 Ra5 Rdf2 R 
 34	+0.22!	2.42G	1:58.91	Nxe5! 
 34	+0.10?	2.39G	1:57.30	Nxe5 Nxe5? 
 34	+0.24!	2.33G	1:54.46	Nxe5! 
 34	+0.17?	2.20G	1:48.08	Nxe5 Nxe5? 
 33	+0.25 	2.01G	1:38.85	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Bd7 b4 a6 a4 Rae8 Qb3 Qe5 b5 axb5 axb5 Nf6 c6 Bf5 Qc3 Kg7 c7 Bc8 Qxe5 Rxe5 d6 Rxe3 Bf3 Rfe8 Rfd1 b6 Rc6 Nd7 Rb1 R3e5 
 33	+0.20!	1.87G	1:31.95	Nxe5! 
 33	+0.03?	1.82G	1:29.58	Nxe5 Nxe5? 
 33	+0.26!	1.56G	1:17.14	Nxe5! 
 33	+0.19!	1.49G	1:13.71	Nxe5! 
 33	+0.15!	1.45G	1:11.62	Nxe5! 
 32	+0.11 	1.37G	1:07.64	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Bd7 b4 Qe5 Qc3 Qxc3 Rxc3 a5 a3 Bb5 bxa5 Rxa5 g4 Bxe2 gxh5 Bxf1 Kxf1 Rc8 hxg6 hxg6 d6 Kf8 Ke2 Raxc5 Rxc5 Rxc5 Kd3 
 32	+0.17!	1.34G	1:06.03	Nxe5! 
 32	+0.08?	1.22G	1:00.53	Nxe5 Nxe5? 
 32	+0.19?	1.18G	0:58.29	Nxe5 Nxe5? 
 32	+0.27?	1.11G	0:55.12	Nxe5 Nxe5? 
 31	+0.34 	1.09G	0:53.78	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Bd4 Rac1 Bxe3 fxe3 Qe5 Qc3 Re8 Qxe5 Rxe5 b4 Rxe3 d6 Rb8 Bd5 Re5 Bb3 Bf5 b5 Rxe2 c6 Nf6 Rcd1 Kg7 d7 bxc6 Rxf5 gxf5 
 31	+0.17?	909.7M	0:45.17	Nxe5 Nxe5? 
 31	+0.25?	879.3M	0:43.68	Nxe5 Nxe5? 
 30	+0.32 	777.3M	0:38.70	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Qe5 b4 Nf6 b5 Ng4 Nxg4 Bxg4 e3 Qh5 Rab1 Bf3 Bxf3 Qxf3 Qc4 h5 d6 a6 bxa6 Rxa6 e4 Ra3 Rxb7 Rd3 Rb3 Rd4 Qc2 Qxe4 Qxe4 Rxe4 
 30	+0.35!	575.9M	0:28.68	Nxe5! 
 30	+0.24!	501.6M	0:25.01	Nxe5! 
 30	+0.17!	474.1M	0:23.65	Nxe5! 
 30	+0.12!	437.5M	0:21.83	Nxe5! 
 29	+0.08 	411.5M	0:20.52	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Nd6 Be5 Nxc8 Raxc8 Rac1 b6 c6 f5 e3 Nf6 Rfd1 Bd6 a3 Rfe8 b4 Qe7 Qd2 Rc7 Qd4 Qf7 Qd3 Qe7 Rc2 Ng4 Rb2 Nf6 Rdb1 Ne4 Rb3 
 28	+0.07 	317.3M	0:15.83	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Nd6 Be5 Nxc8 Raxc8 Rac1 b6 c6 f5 e3 Nf6 Rfd1 Bd6 Qd3 Qe7 a3 Rfe8 b4 Rc7 Rc2 Ng4 Ra2 Nf6 Rb2 Qg7 Rc2 Qe7 Rb1 Ng4 
 27	+0.07 	293.5M	0:14.66	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Nd6 Be5 Nxc8 Raxc8 Rac1 b6 c6 f5 e3 Nf6 Rfd1 Bd6 Qd3 Qe7 a3 Rfe8 b4 Rc7 Rc2 Ng4 Ra2 Nf6 Rb2 Qg7 Rc2 Qe7 
 27	+0.12!	277.0M	0:13.83	Nxe5! 
 26	+0.08 	270.9M	0:13.53	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Nd6 Be5 Nxc8 Raxc8 Rac1 b6 c6 f5 e3 Nf6 Rfd1 Bd6 Qd3 Qe7 b4 Rfe8 a4 Rc7 Rc2 Ng4 h3 Ne5 Qd4 Qg5 a5 bxa5 bxa5 
 26	+0.19!	257.3M	0:12.84	Nxe5! 
 26	+0.12?	235.4M	0:11.75	Nxe5 Nxe5? 
 25	+0.20 	184.1M	0:09.19	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Ne3 Qe5 b4 Nf6 Bf3 Bh3 Rfd1 Rae8 Rac1 Ne4 Nc4 Qf5 d6 Ng5 Bxb7 Ne4 Bxe4 Qxe4 Qxe4 Rxe4 
 25	+0.17?	178.7M	0:08.92	Nxe5 Nxe5? 
 24	+0.24 	159.0M	0:07.93	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Nd6 Be5 Nxc8 Raxc8 Rac1 b6 c6 f5 e3 Nf6 Rfd1 Bd6 a3 Qe7 Qd3 Ng4 b4 Rc7 Rc2 Qf6 Qb3 Re8 
 24	+0.35!	157.4M	0:07.85	Nxe5! 
 23	+0.29!	119.1M	0:05.94	Nxe5! 
 22	+0.23 	73.3M  	0:03.66	Nxe5 Nxe5 cxd5 Nh5 Bxe5 Qxe5 Nc4 Qc7 Nd6 Rb8 Rab1 b6 b4 bxc5 bxc5 Ba6 Bf3 Nf6 Rxb8 Rxb8 Qa4 Bc8 Rc1 Bf8 Qf4 Bxd6 cxd6 Qd8 e4 Bb7 Rb1 Nd7 
 22	+0.32!	67.7M  	0:03.38	Nxe5! 
 22	+0.19!	66.3M  	0:03.31	Nxe5! 
 21	+0.27 	17.8M  	0:00.89	Bg5 d4 b4 h6 Bh4 g5 Nxg5 hxg5 Bxg5 a5 a3 Qe6 Rad1 Qg4 Bh4 Re8 e3 Nf8 exd4 exd4 Nf3 axb4 axb4 Bf5 Qb2 

[jdart]

It is not uncommon in human games. The Sicilian Dragon is just full of Rxc3 sacs. There is also this one (12. .. b6). The compensation is very long-term and probably outside even an engine's horizon, but Topalov won the game. I am pretty sure it was posted before:

[D] r6r/pp1kpp1p/4bnp1/2p5/2P5/2P3P1/P3PPBP/1RB1K2R b K - 0 12

[pgn]
[Event "Sarajevo"]
[Site "Sarajevo BIH"]
[Date "2000.05.24"]
[EventDate "2000.05.17"]
[Round "7"]
[Result "0-1"]
[White "Etienne Bacrot"]
[Black "Veselin Topalov"]
[ECO "A41"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "66"]

1.d4 d6 2.Nf3 g6 3.c4 Bg7 4.Nc3 c5 5.dxc5 Bxc3+ 6.bxc3 dxc5
7.Qxd8+ Kxd8 8.Ne5 Be6 9.g3 Nd7 10.Nxd7 Kxd7 11.Bg2 Nf6 12.Rb1
b6 13.Bxa8 Rxa8 14.f3 Ne8 15.Bf4 Bxc4 16.Rb2 Kc6 17.Rd2 Nd6
18.Bxd6 exd6 19.Kf2 d5 20.e4 dxe4 21.fxe4 Re8 22.Re1 Be6
23.Kf3 Kb5 24.Kf4 Kc4 25.Re3 a5 26.Kg5 b5 27.Kh6 b4 28.cxb4
axb4 29.Kxh7 Ra8 30.Kg7 Kb5 31.h4 c4 32.g4 c3 33.Rg2 Rxa2 0-1[/pgn]

[AdminX]

Here is another one from current TCEC 10

[d]6k1/2qr1p1p/1p2rbp1/p1pp1n2/P7/1Q1PP1PP/1P1B1PB1/3RR1K1 b - - 7 27

Here Nirvana played 27. ... Nxg3

[Vinvin]

Here is another one from current TCEC 10

[d]6k1/2qr1p1p/1p2rbp1/p1pp1n2/P7/1Q1PP1PP/1P1B1PB1/3RR1K1 b - - 7 27

Here Nirvana played 27. ... Nxg3

And that was a mistake. Even Nirvana finds it's a mistake 2 moves later.
http://tcec.chessdom.com/archive.php?se=10&st=1&ga=1

[tpoppins]

By chance I came across the following position in a game Houdini(Tactical) played versus Critter

[d]r1b2rk1/pp1nqpbp/4pnp1/2Pp4/2P2B2/5NP1/PPQNPPBP/R4RK1 b - - 0 11

Here Critter chose 11. ... e5 instead of regaining it's pawn. Thus allowing Houdini to chose the positional exchange sacrifice 12. Nxe5 Nxe5 13. cxd5

There seems to be a confusion of terms here. Neither the example above nor the one below is an exchange sacrifice.

Here is another one from current TCEC 10

[d]6k1/2qr1p1p/1p2rbp1/p1pp1n2/P7/1Q1PP1PP/1P1B1PB1/3RR1K1 b - - 7 27

There is also this one (12. .. b6). The compensation is very long-term and probably outside even an engine's horizon, but Topalov won the game. I am pretty sure it was posted before:

[D] r6r/pp1kpp1p/4bnp1/2p5/2P5/2P3P1/P3PPBP/1RB1K2R b K - 0 12

And this one is a modern classic. To play a move like 12...b6 in such a nearly-level endgame one must have a fine positional sense and a supreme confidence in own technique (not to mention a set of rather large and hairy cojones ;).

Here Nirvana played 27. ... Nxg3

Here is an example of a human exchange sacrifice.

14. Ra2!?

[d]r2qr1k1/p1pn1ppp/1p1p1n2/8/2bPP3/P1P1PQN1/R5PP/1BB1K2R b K - 0 14

Not bad, but a rather tame move compared to 22.Ra3 in Tal--Gligoric, Candidates qf2 (1) Belgrade 1968:

[d]r3rbk1/2p2pp1/p6B/1p1qp3/P2P2R1/1nP2N1P/5PP1/R2Q2K1 w - - 0 22

although that wasn't an exchange sacrifice either. Here are two real classics, then:

[d]3rq1k1/4rppp/2n3b1/pp2P3/2pP1QB1/P1P1R3/1B4PP/4R1K1 b - - 0 25
Reshevsky--Petrosian, Zuerich (2) 1953
25...Re6

[d]2r2bk1/R4p2/3p1Pp1/3Np3/4P2p/3r3P/6PK/1R6 b - - 0 41
Tal--Fischer, Candidates (4) Curacao 1962
41...Rxd5

There is also this one (12. .. b6). The compensation is very long-term and probably outside even an engine's horizon, but Topalov won the game. I am pretty sure it was posted before:

[D] r6r/pp1kpp1p/4bnp1/2p5/2P5/2P3P1/P3PPBP/1RB1K2R b K - 0 12

And this one is a modern classic. To play a move like 12...b6 in such a nearly-level endgame one must have a fine positional sense and a supreme confidence in own technique (not to mention a set of rather large and hairy cojones ;).