The best player of the late 1500s and early 1600s was Gioachino Greco, who showed this
idea in his writings:
[Event "sacrifice on h7 by B (Greek Gi"]
[Site "sacrifice on h7 by B (Greek G"]
[Date "1792.??.??"]
[Round "83"]
[White "greco"]
[Black "Anon"]
[Result "1-0"]
[ECO "C00"]
[Annotator "Regis,Dave"]
[PlyCount "21"]
1. e4 e6 2. d4 Nf6 3. Bd3 Nc6 4. Nf3 Be7 (4... Bb4+ 5. c3 Ba5 6. h4 O-O 7. e5
Nd5 8. Bxh7+ Kxh7 9. Ng5+ Kg8 10. Qh5 Re8 11. Qxf7+ Kh8 12. Qh5+ Kg8 13. Qh7+
Kf8 14. Qh8+ Ke7 15. Qxg7#) 5. h4 O-O 6. e5 Nd5 {[#] the simplest setting for
the sacrifice - sometimes known as the Greek Gift after El Greco's pioneering
analysis} 7. Bxh7+ Kxh7 8. Ng5+ Kg8 (8... Bxg5 9. hxg5+ Kg6 10. Qd3+ (10. Qh5+
Kf5) 10... f5 11. gxf6+ Kf7 12. Rh7 Rg8 13. Qg3 Qf8 14. Bh6 Ke8) (8... Kg6 9.
h5+ Kh6 (9... Kf5 10. Qf3+ (10. g4#) 10... Nf4 11. Qxf4#)) 9. Qh5 Bxg5 10. hxg5
f5 (10... f6 11. g6) 11. g6 1-0
Neat, eh? This sacrifice is known to chessplayers the world over as the
Greek Gift(*). What makes it work? There's a sort of recipe:
1.Black Knight missing from f6
2.White can bring the Queen to h5 or d3
3.The Queen has support in attack, such as:
a.a Knight that can get to g5 (without being swapped off by Be7xg5)
b.a Rook can get to h3 (or similar)
4.Black can't cover h7 with ...Bf5 or ...Nd7-f6
Actually, in Greco's example, we haven't got the Knight in support (3a),
as it gets swapped off, but the swap brings in the Rook (3b).
A typical variation:
[Event "sacrifice on h7 by B (Greek Gi"]
[Site "sacrifice on h7 by B (Greek G"]
[Date "1792.??.??"]
[Round "83"]
[White "greco"]
[Black "Anon"]
[Result "1-0"]
[ECO "C00"]
[Annotator "Regis,Dave"]
[PlyCount "29"]
1. e4 e6 2. d4 Nf6 3. Bd3 Nc6 4. Nf3 Bb4+ 5. c3 Ba5 6. h4 O-O 7. e5 Nd5 8.
Bxh7+ Kxh7 9. Ng5+ Kg8 10. Qh5 Re8 {This next bit is important} 11. Qxf7+ Kh8
12. Qh5+ Kg8 13. Qh7+ Kf8 14. Qh8+ Ke7 15. Qxg7# 1-0
The Knight and Queen together can't force mate all by themselves, they
need a little bit of help from a pawn in both cases (Pg5 or Pe5).
We've seen what happens if the King tries to retreat, but what if it is
brave and comes forward?
[Event "sacrifice on h7 by B (Greek Gi"]
[Site "sacrifice on h7 by B (Greek G"]
[Date "1792.??.??"]
[Round "83"]
[White "greco"]
[Black "Anon"]
[Result "1-0"]
[ECO "C00"]
[Annotator "Regis,Dave"]
[PlyCount "18"]
1. e4 e6 2. d4 Nf6 3. Bd3 Nc6 4. Nf3 Be7 5. h4 O-O 6. e5 Nd5 {[#] the
simplest setting for the sacrifice - sometimes known as the Greek Gift after
El Greco's pioneering analysis} 7. Bxh7+ Kxh7 8. Ng5+ Kg6 9. h5+ (9. Qd3+ f5
10. exf6+ Kxf6 11. Qf3+ Kg6 12. h5+ Kh6 13. Qe4) 9... Kh6 (9... Kf5 10. Qf3+ (
10. g4#) 10... Nf4 11. Qxf4#) 1-0
It's harder to calculate if the King comes forward, but that one wasn't too bad.
Try again:
[Event "sacrifice on h7 by B (Madrid)"]
[Site "sacrifice on h7 by B (Madrid)"]
[Date "1971.??.??"]
[Round "?"]
[White "Markland, PR."]
[Black "Klundt, K."]
[Result "1-0"]
[ECO "B22"]
[PlyCount "33"]
1. e4 c5 2. c3 Nf6 3. e5 Nd5 4. d4 cxd4 5. cxd4 d6 6. Nf3 Nc6 7. Nc3 Nxc3 8.
bxc3 e6 9. exd6 Bxd6 10. Bd3 Qa5 11. O-O {!?} Qxc3 12. Rb1 O-O {?} 13. Rb3 Qa5
{[#] (i) bN missing from f6 (ii) wN can go to g5 without losing the attack
(iii) wQ can get to g4/h5 (iv) other pieces can support the attack} 14. Bxh7+
Kxh7 15. Ng5+ Kg6 16. Rh3 Bd7 17. Ne4 (17. Ne4 {1-0} f6 18. Nxd6 f5 19. Qh5+
Kf6 20. Bg5#) 1-0
[Event "attack: king hunt (cf. sacrifi"]
[Site "attack: king hunt (cf. sacrif"]
[Date "1983.??.??"]
[Round "?"]
[White "Stein"]
[Black "Langeweg (Plovdiv Echt)"]
[Result "1-0"]
[ECO "C55"]
[PlyCount "29"]
1. e4 e5 2. Nf3 Nc6 3. Bc4 Bc5 4. c3 Nf6 5. d4 exd4 6. O-O Nxe4 7. cxd4 Be7 8.
d5 Nb8 9. Re1 Nd6 10. Bd3 O-O 11. Nc3 Ne8 {[#] the classic Bxh7 sacrifice
suggests itself, but White first uses another characteristic sacrifice of a
pawn} 12. d6 cxd6 {the White pieces now have the d5 point for launching, while
Black will find it very hard to unravel even if the attack falters} (12... Nxd6
13. Bxh7+ Kxh7 14. Rxe7 Qxe7 15. Nd5 Qd8 16. Ng5+ Kg6 17. Nf4+ Kf6 18. Nh7+ Ke7
19. Qe2+ Ne4 20. Qxe4+ Kd6 21. Qd5+ Ke7 22. Qe5#) 13. Bxh7+ {usually this is
impossible when the bB is on e7} Kxh7 14. Rxe7 Qxe7 15. Nd5 {( black resigns )}
(15. Nd5 Qd8 16. Ng5+ Kg6 17. Nf4+ Kxg5 18. h4+ Kf6 19. Qd4+ Kf5 20. Qd5+ Kf6
21. Qg5#) 1-0
A bit harder, that one: White needed the support of the Nc3 and had to
remove the Be7.
Here's a famous example, which only just works:
[Event "sacrifice on h7 by B (hard)"]
[Site "sacrifice on h7 by B (hard)"]
[Date "1930.??.??"]
[Round "?"]
[White "colle"]
[Black "o'hanlon (nice)"]
[Result "1-0"]
[ECO "D05"]
[PlyCount "39"]
1. d4 d5 2. Nf3 Nf6 3. e3 c5 4. c3 e6 5. Bd3 Bd6 6. Nbd2 Nbd7 7. O-O O-O 8. Re1
(8. Qe2) (8. e4) 8... Re8 9. e4 dxe4 10. Nxe4 Nxe4 11. Bxe4 cxd4 {[#]} (11...
Nf6) 12. Bxh7+ Kxh7 13. Ng5+ {[#]} Kg6 (13... Kg8 {This is the only real
alternative. Analysis by Euwe and Kramer suggests that White's attack is worth
a draw but no more. The variations and ideas are very typical and worth
playing over.} 14. Qh5 {[#]} Nf6 (14... Ne5 15. Rxe5 (15. Qh7+ Kf8 16. Ne4 Ng6
17. Nxd6 Qxd6 18. h4 Ke7 19. h5 Rh8 20. Bg5+ Ke8 (20... Kf8 21. hxg6 Rxh7 22.
gxh7 {wins}) 21. Qxg7 Rxh5 22. Qf6 Qe7 {and Black is better}) 15... Bxe5 16.
Qxf7+ Kh8 17. Qh5+ {(Euwe) and White has nothing better than perpetual check} (
17. b3 $1 {(Ed. Lasker) idea Ba3} Qe7 (17... Bd6 18. Qh5+ Kg8 19. Qh7+ Kf8 20.
Qh8+ Ke7 21. Qxg7#) (17... Bf6 18. Qh5+ Kg8 19. Qh7+ Kf8 20. Ba3+ Be7 21. Qh8#)
18. Qh5+ Kg8 19. Qh7+ Kf8 20. Qh8#) (17. Qg6 Kg8 18. b3 $1) 17... Kg8 18. b3 $1
) (14... Qf6 15. Qh7+ (15. Re4) (15. Nh7 Qg6) 15... Kf8 16. Ne4 Qe5 17. cxd4 (
17. Bg5 f6 18. Nxf6 (18. Nxd6) 18... Nxf6) 17... Qd5 {?} (17... Qxd4 $2 18.
Qh8+ Ke7 19. Bg5+ {wins} Nf6 20. Qxg7 Be5 21. Rad1 Qxb2 22. Rd2) (17... Qxh2+
18. Qxh2 Bxh2+ 19. Kxh2 {is about level}) (17... Qa5) 18. Qh8+ Ke7 19. Qxg7 {
with strong attack} Kd8 20. Bg5+ f6 21. Nxf6) 15. Qxf7+ Kh8 16. Re4 Nxe4 (16...
Bxh2+ 17. Kxh2 Nxe4 18. Qh5+ Kg8 19. Qh7+ Kf8 20. Qh8+ Ke7 21. Qxg7+ Kd6 22.
Nf7+ {wins the Q}) 17. Qh5+ Kg8 18. Qh7+ Kf8 19. Qh8+ Ke7 20. Qxg7#) (13... Kh8
14. Qh5+ (14. Nxf7+) 14... Kg8 15. Qxf7+ Kh8 16. Qh5+ Kg8 17. Qh7+ Kf8 18. Qh8+
Ke7 19. Qxg7# (19. Rxe6#)) 14. h4 (14. Qg4 f5 (14... Nf6)) 14... Rh8 {[#]} (
14... f5 15. h5+ Kf6 16. Qxd4+ Be5 17. Qh4 g6 18. f4 {+-}) 15. Rxe6+ Nf6 {[#]}
(15... fxe6 16. Qd3+ Kf6 17. Qf3+ Kg6 18. Qf7+ Kh6 19. Nxe6+) 16. h5+ Kh6 (
16... Rxh5 17. Qd3+ Kh6 18. Qh7#) 17. Rxd6 Qa5 18. Nxf7+ Kh7 19. Ng5+ Kg8 20.
Qb3+ 1-0
Good enough on the day, but did it work against the best defence?
Masters argued about that one for decades; the American master Edward
Lasker** finally showed a forced win using the Ba3 idea.
All the different King moves need to be calculated; sometimes they are
all easy to calculate, sometimes one gets really tough, sometimes they
are all tough!
* Timeo Danaos et dona ferentes, wrote Virgil, giving the words
to Laocoon: "I fear the Danaans, even when bearing gifts". (Often
paraphrased as: "Beware the Greeks bearing gifts".) After ten years of
war, the ancient Greeks (the Danaans) parked a huge wooden horse outside
the city walls of Troy, and apparently disappeared; once the Trojans brought the
horse into the city, Greek soldiers emerged from hiding in the horse and
opened the gates to allow in the Greek army.
** (not the World Champion Emanuel Lasker)