CHAPTER SIX

Two Knights Defence

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 this line is quite unusual should mean
that many exponents of White's side
will be caught unawares.
Instead of 9.Nf3 White can also try
Steinitz's 9.Nh3, which is another
move that Bobby Fischer rehabilitated.
But against this I think that Malaniuk's
9...g5 followed by 10...Bg7 (Grischuk-
Malaniuk( Game 40)) is a strong plan,
shutting the knight out of the game and
setting about advancing the kingside
pawns. In my correspondence game
with Azevedo Pessoa (Azevedo Pes-
If White plays 1.e4 e5 2.Nf3 Nc6 soa-Davies( Game 41)) my opponent
3.Bc4, both 3...Nf6 4.Ng5 and 3...Bc5 tried another unusual move that the
4.b4!? - or 4.c3 Nf6 5.d4 - allow White Dutch GM John Van der Wiel has
to sharpen the struggle and involve his played in several games, namely
opponent in having to know some 8.Qf3. I felt that Black had some in-
theory. itiative but White finally managed to
My choice of 3...Nf6, the Two force a draw by perpetual check. If
Knights Defence, was made because Black wants more he could examine
of its relative pugnacity. White cannot 11...Bb7!? instead of 11...Be6.
easily create an equal position in Finally there is 6.d3, which was re-
which it is difficult for Black to play cently given a run out in the corres-
for a win. pondence game Spitz-Piccardo( Game
After 3...Nf6 White's 4.Ng5 effec- 42). Black obtained good counterplay,
tively wins a pawn, but the lost time although there may be more to be said
gives Black good compensation after in this complex line.
4...d5 5.exd5 Na5. Rather than force Black to play a
In Alekseev-Yemelin( Game 39) we promising gambit, White can try to
see what is essentially the main line( seize the initiative with 4.d4. After
9.Nf3), but with my recommendation 4...exd4 5.e5 I like the unusual but
being the slightly unusual 10...Bc5 sound 5...Ne4!?, which has also been
rather than 10...Bd6, and then Yeme- the choice of strong grandmasters such
lin's 12...Nb7!?. Black seems to be as Romanishin. In Kozakov-Jonkman(
doing quite well here, and the fact that Game 43) White recovers the pawn
106
 Two Knights Defence

with 6.Bd5 Nc5 7.Bxc6 but gives 9.Nh3 is dealt with in Grischuk-
Black excellent light square play. Malaniuk( Game 40).
V.Gurevich-Romanishin( Game 44) 9...e4 10.Ne5 Bc5!
features the more testing 7.0-0, but
10...Bd6 has been more popular in
even so Black gets counterplay with
the past but I like the text. Black gets
the clever 7...Be7 8.Qe2 0-0 9.Rd1
White to play 11.c3 as after 11.d4
Qe8!. White can also try 5.0-0 instead
exd3 12.Nxd3 Qc7 there is an effective
of 5.e5, but this was essentially put out
development for White with 13.b3!, as
of commission in Karaklajic-
played by Kasparov and Morozevich.
Jovanovic( Game 45) with 11...Bxh2+.
Finally we come to 4.d3, which often 11.c3
leads to similar positions to the Closed
Variation of the Ruy Lopez. Dolma-
tov-Romanishin( Game 46) features a
good way to treat this line for Black,
playing ...Rad8 before retreating the
knight to c6 and just letting White cap-
ture on e6 if that's what he wants to do.

Game 39
E. Alekseev–V.Yemelin
St-Petersburg, 2000 In 'Play the Open Games as Black',
Emms suggested that the simple
11.0–0
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5
d5 5.exd5 Na5 6.Bb5+ c6 7.dxc6
bxc6 8.Be2

might be better, suggesting that

Black still had to justify his play after
11...Qd6 12.Ng4 Bxg4 13.Bxg4 h5
14.Be2( 14.Bh3 is powerfully met by
8.Qf3 is covered in the game Azeve- 14...g5!) 14...Ng4 15.g3. I can see his
do Pessoa-Davies( Game 41). point in that 15...Nxf2 16.Rxf2 Bxf2+
8...h6 9.Nf3 17.Kxf2 h4 18.Qg1 hxg3+ 19.hxg3
Introducing what is really the main defends for White and he has two
line of the 4 Ng5 Two Knights. pieces for a rook.; Black's best may be
to avoid an immediate attack on the
knight with 11.0–0 0–0!, after which
107
Play 1.e4 e5!

12.c3 Qc7 13.d4 exd3 14.Nxd3 leads 16.Qc2 Nd5 17.b4 Nb7 18.Na3(
to similar positions to Sutovsky- 18.Bf3 Be6 19.Na3 Bxe5 20.fxe5
Postny in the next note. Qxe5 was also equal in Herbrech-
11...Qc7 12.f4 tsmeier-Read, Correspondence 1985)
and now 18...a5!( as Emms pointed
As I mentioned in the previous note,
out, 18...Nxf4 19.Nxf7! is good for
12.d4 is not very good here because
White) 19.Nac4 f6 20.Nxd6( or
12...exd3 13.Nxd3 Bd6
20.Nd3 Bf5) 20...Nxd6 21.Nd3 Bf5
looks quite good for Black.
13.b4
White has tried two other moves
here: 13.d4 exd3 14.Qxd3 0–0

makes a development with b2-b3 and

Bb2 ineffective. Sutovsky-Postny, Tel
Aviv 2001, continued 14.Nd2 Bf5
15.b4 Nb7 16.Nc4 Rd8 17.Be3 0–0
18.Bxa7 Nd5 19.Bd4 Rfe8 20.Nxd6
Nxd6 21.0–0 Nb5 and Black had dan- still leaves White with the problem
gerous threats. of getting his king safe. Fine-Steiner,
12...Nb7!? Washington 1944, went 15.Nd2 Nd6
16.Bf3 Bf5 17.Qe2 Nd5 18.Nb3 Bb6
19.c4 Nb4 20.c5 Nc2+ 21.Kf2 Bxc5+
22.Nxc5 Nxa1 and White was fighting
for survival.
13.Qa4

I really like this move, calmly im-

proving Black's worst place piece be-
fore proceeding with direct action.
12...Bd6 has been played more often,
for example 13.d4 exd3 14.Qxd3 0–0
15.0–0( 15.Nd2 Bxe5 16.fxe5 Qxe5 attacks the c6-pawn: 13...Nd8(
17.0–0 was equal in Harding-Read, 13...Bd7!? is also interesting but
Correspondence 1992) 15...Rd8 doesn't seem to have been tried) 14.b3(
14.Na3 0–0 15.Nc2 Re8 16.b4 Bd6
108
 Two Knights Defence

17.Nd4 Bxe5 18.fxe5 Qxe5 19.Nxc6 14...exd3 15.Qxd3 0–0 16.0–0 a5

Qg5 20.0–0 Bh3 21.Rf2 was Mednis- 17.Bf3 Bxe5 18.fxe5 Qxe5 19.Bxc6
Van Hoorne, Antwerp 1955, when axb4 20.Qb5
21...Nxc6! 22.Qxc6 e3! 23.dxe3 Ne4
would have been good for Black)
14...Ne6!( 14...a5 15.Ba3 Ba7 16.Rf1
Be6 17.d4 exd3 18.Nxd3 Nb7 was
quite intricate and complex in Fink-
Pinkus, New York 1946; but 14...0–0
15.Ba3 Bxa3 16.Nxa3 seems to take a
lot of the steam out of Black's posi-
tion) 15.Qxc6+( 15.d4 exd3 16.Nxd3
Bd6 17.Bf3 Bb7 gives Black excellent
compensation) 15...Qxc6 16.Nxc6
Nxf4 17.d4 exd3 18.Bf3 Bd6 19.Nb4
Rb8 20.Bxf4 Bxf4 21.Nxd3 Bd6 and was equal and soon drawn in Med-
the bishop pair plus White's pawn nis-Spassky, Antwerp 1955.
weaknesses give Black adequate com- 14...Nd5 15.0–0 Nxf4 16.Rxf4 Bxe5
pensation for the pawn. 17.Rxe4 0–0 18.Nc4?
13...Bd6 After this Black is clearly better.
White has to defend his h2 pawn with
18.Rh4.
18...Bxh2+ 19.Kh1 Nd6 20.Nxd6
Bxd6 21.d4 Bf5 22.Re3 Rae8 23.Bf3
Bg3 24.Bg4 Be4

13...Bb6 seems very reasonable here

too, for example 14.Na3( 14.Qa4 Bd7
15.Na3 0–0 16.Nac4 Nd6 17.Nxb6
axb6 18.Qb3 Be6 19.Qb1 b5 gave
Black good play for the pawn in Med-
nis-F.Anderson, Montreal 1956) 25.Bd2
14...Nd6 15.Nac4 Nxc4 16.Bxc4 0–0 The last chance to stay on the board
17.d4?! exd3 18.Qxd3 Ng4! 19.Qe2 was with 25.Kg1. Now the storm
Re8 20.Bxf7+ Qxf7 21.Nxf7 Rxe2+ clouds gather around White's king.
22.Kxe2 Kxf7 and Black had two
25...f5 26.Bh5?! Qf4 27.Rxg3 Qxg3
pieces for a rook in Szymanski-Sliwa,
Krakow 1953. With White being the exchange
down and under attack, the rest re-
14.Na3
quires little comment.
This looks too ambitious. White
28.Qg1 g6 29.Bd1 Qh3+ 30.Qh2
should probably play 14.d4, when
Bxg2+ 31.Kg1 Qxh2+ 32.Kxh2 Be4
109
Play 1.e4 e5!

33.Bxh6 Rf7 34.Bd2 Rh7+ 35.Kg3 his kingside pawns whilst menac-
Rh1 36.Rc1 Rg1+ 37.Kf2 Rg2+ ing...g5-g4; indeed White cannot castle
38.Kf1 Rxd2 39.c4 Bg2+ 40.Kg1 here because 10...g4 wins the knight.
Re1+ 0–1 10.d3 Bg7!
Once again a move that I like. Black
Game 40 has also tried to bring the knight on a5
back into play with 10...c5 11.Nc3
A.Grischuk–V.Malaniuk
Nc6, but after 12.Bf3 Qd7 13.Ne4
Russian Team Championship, 2001
Nxe4 14.Bxe4 g4( Veinger-Bronstein,
Rishon Le Ziyyon 1991) Bronstein
told me that White could have gained
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 the advantage with 15.Qxg4!( rather
d5 5.exd5 Na5 6.Bb5+ c6 7.dxc6 than the 15.Ng1 f5 16.Bxc6 Qxc6
bxc6 8.Be2 h6 9.Nh3 17.f3 c4! of the game) 15...Qxg4
16.Bxc6+ Bd7 17.Bxa8. Shredder
doesn't agree just here but I trust
Bronstein's intuition more. In the vari-
ation 17...c4 18.0–0 cxd3 19.cxd3 Rg8
20.Kh1 Qe2 21.Be3 Bxh3 22.gxh3
Qxb2 23.Rfb1 Qe2 24.Be4 Shredder's
assessments have gradually changed
from giving Black a clear advantage to
equal. But Black's king is in serious
trouble here.
11.Nc3
Steinitz's move, which was later used This is probably best. In Mutu-
by Bobby Fischer, Nigel Short and Deseatnicov, Kishnev 2001, White
Gata Kamsky. played 11.Be3, but there followed
White avoids the loss of time inhe- 11...0–0 12.Nc3 Nb7 13.Ng1 Nd5
rent in 9.Nf3 e4, but puts the knight on
a very strange square.
9...g5!

14.h4( 14.Nxd5 might improve but

after 14...cxd5 15.h4 g4 16.Bxg4 f5
17.Bh5 f4 18.Bc1 Bf5 Black gets tre-
Black has tried several other moves mendous pressure for his pawns)
here such as 9...Bd6 and 9...Bc5, but I 14...Nxe3 15.fxe3 Nd6, and Black had
think this is the best. Black advances more than enough compensation.
110
 Two Knights Defence

11...0–0 12.Ng1 Nb7 13.Nf3 25...Nxe1 is a more rational move,

Pilgaard-Dervishi, Bergamo 2004, after which 26.Bd2 cxd4 27.Rxe1 Bb7
went 13.Bf3 but after 13...Qc7 14.h3 simply favours Black.
Nd6 15.Nge2 Nf5 16.0–0 Nh4 17.Ng3 26.Bg2 f4
Rb8, The immediate 26...Nf3+ can be
answered by 27.Bxf3 gxf3 28.Be3, for
example 28...Qg4 29.Qc6 e4 30.Qe6+
Kh7 31.Kh1 Qh3 32.Rg1, when it is
difficult to see how Black can continue
the attack.
27.Bd5+
If White wants to put the bishop on
e4 he should do so immediately.
27.Be4 Nf3+ 28.Kh1 Qh5 29.Bxf3
gxf3 30.gxf4 is much better than the
game because Black's king is still sit-
White was under serious pressure. ting on the g-file.
13...Nd5 14.0–0 Nd6 15.Ne4 f5! 27...Kh8 28.Be4?!
28.Bd2 is a preferable follow-up.
28...Nf3+ 29.Bxf3

Now it is clear that the bishop on g7

is beautifully placed. It both supports
the centre and helps cover the king.
The point behind playing 27 Be4 ra-
16.Nxd6 Qxd6 17.Nd2 g4 18.Re1 ther than the check is that now White
Ba6 19.Bf1 Qg6 20.g3 cannot play 29.Kh1 because it runs
20.Nb3 is well met by 20...Rad8, into 29...Qh5 30.Bxf3 fxg3!( Black has
when 21.Nc5 Bc8 gets all Black's time for this because White has no
pieces working nicely. bishop check) 31.fxg3 gxf3 etc.
20...Rad8 21.c4 Nb4 22.d4 Rxd4 29...gxf3 30.Bd2 Qg4 31.Kh1
23.Qa4 c5 24.Nb3 Nc2 25.Nxd4 Or 31.Qc6 fxg3 followed by
Nxd4?! 32...Qh3.
Black is tempted by the idea of deli- 31...fxg3 32.Qxa6 Qh3 0–1
vering mate, but maybe he shouldn't
have been. A terrific win by Malaniuk.

111
Play 1.e4 e5!

0 and Black had a menacing attacking

Game 41 position in Steinitz-Chigorin, World
F.Azevedo Pessoa–N.Davies Championship, Havana 1892.
Correspondence, 2004 8.Ba4? h6 9.Nf3 e4 10.Qe2( after
10.Ne5 Qd4! Black wins a piece for
inadequate compensation) 10...Bd6
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5
11.Nd4( 11.d3 0–0 12.dxe4 Nxe4
d5 5.exd5 Na5 6.Bb5+ c6 7.dxc6
13.0–0 Nc5 14.Bb3 Ba6 is good for
bxc6 8.Qf3!?
Black according to Emms) 11...Qb6
12.c3 0–0 13.b4?! Bxb4 14.cxb4 Qxd4
favoured Black in Horwitz-Pindar,
Manchester 1961.
8...h6
Black has also tried 8...Be7 and
8...Rb8, but the text seems like the
most reliable continuation.
9.Ne4 Nd5 10.Nbc3!

This is the kind of move that can be

very awkward in practical chess be-
cause it is sharp and Black may not
know it.
Besides the standard 8 Be2 there are
three other moves here: the dodgy, the
ugly and the bad:
8.Bd3 Nd5!

White must develop at top speed here

or risk being crushed by Black's ad-
vancing e- and f-pawns.
10.Ba4 Be7 11.0–0( or 11.Ng3 0–0
12.Nf5 Bc5 13.d3 Qf6 14.g4 Nf4
15.Bxf4 exf4 16.Nd2, as in Capitaine-
Legrand, Plancoet 2003, and now
16...Bxf5 17.gxf5 Qd4 18.Bb3 Rae8+
is simply very good for Black) 11...0–
0 12.d3( 12.Re1 f5 13.Ng3 e4 14.Qd1
Ba6 offered Black excellent compen-
9.Ne4 f5 10.Ng3 Nf4 11.Bf1 Bc5 sation for the pawn in Stenzel-Brooks,
12.c3 Bb6 13.d4 Ng6 14.Bd3 0–0 Chicago 1994) 12...f5 13.Nec3 Bb7
15.b4 Nb7 16.Bc4+ Kh8 gave Black a 14.a3 c5 15.Qh5 Nb6 16.Bxh6? was
strong initiative in Castaldi-Keres, played in Van Weersel-Solleveld,
Stockholm Olympiad 1937. Vlissingen 2003, and now 16...Qd6!
8.Bf1?! h6 9.Nh3 Bc5 10.d3 Qb6 would have been strong because
11.Qe2 Bg4 12.f3 Bxh3 13.gxh3 0–0– 17.Bg5( 17.Bd2 Nxa4 18.Nxa4 Qc6

112
 Two Knights Defence

wins a piece) 17...Bxg5 18.Qxg5 Rf6

leaves White facing many threats.
10...cxb5 11.Nxd5 Be6

This is a new move, re-centralizing

the knight before committing myself to
any concrete action.
Gikas-Balashov, Lugano 1988, had
At the time I played the game I liked gone 13...Qd7 14.Ng3 h5 15.c3( 15.d3
keeping the bishop here because it might be an improvement) 15...Nc6
covers the crucial f5-square. I was also 16.Rd1 h4 with a very messy position.
influenced by the fact that Balashov 14.d3 Qd7 15.Ng3
has been Spassky's second for many I also looked at 15.Qh5 but thought
years and played 11...Be6 after that Black was doing well after
Spassky had tried 11...Bb7. Van der 15...Be7 16.f4 g6 17.Qe2 f5 18.Nc3
Wiel-Spassky, Reggio Emilia 1986, Nd4 etc.
went 11...Bb7 12.Ne3 Qd7( in
15...h5 16.c3 h4 17.Ngf5 h3 18.g3
P.Watson-L.Cooper, England 2002,
Black played 12...Be7 13.d3 g6, but And here I examined 18.g4, conclud-
then 14.0–0 would have prepared ing that Black stood well after 18...g6
White to meet 14...f5 with 15.Nxf5 19.Ng3 Be7 in view of White's com-
gxf5 16.Qh5+ etc) 13.d3 Nc6 14.0–0 promised kingside.
0–0–0 15.c3 g6( Van der Wiel-Van 18...Ne7!
Kooten, Hoogeveen 2004, continued
instead 15...Qe6 16.Rd1 g6 17.Qf6
Qxf6 18.Nxf6 Bg7 19.Ne4, with Black
having inadequate compensation for
the pawn in the endgame; as this was a
later game it also implies that Van der
Wiel had an improvement ready had
Black played like Spassky) 16.a4 b4
17.Nf6 Qe6 18.Nfd5( 18.a5!? might be
the improvement Van der Wiel had in
mind) 18...f5 19.c4 Nd4 20.Qh3 g5
21.Re1 Rg8 and Black had dangerous Eliminating White's advanced
attacking chances. knight. 18...g6 is not good due to
12.Ne3 Rc8 13.0–0 Nc6 19.Ng4 Qd8 20.Nf6+! Qxf6 21.Nd6+
Ke7 22.Nxc8+ Bxc8 23.Bg5! Qxg5
24.Qxc6, with a position that I tried to
(next diagram) make work for Black but without suc-

113
Play 1.e4 e5!

cess. The poor king position prevents exd4 15.Ne4 Bf4 16.Kf1 a draw was
the efficient reorganization of forces. agreed.
19.g4 8.Qe2
After 19.Nxe7 Bxe7 20.Qe4 Black
can simply castle, 20...0–0, when
21.a4 b4 generates excellent play.
19...g6 20.Nxe7 Bxe7 21.Qe4 0–0
I saw that this could lead to a forced
draw but was unable to find a good
alternative.
22.Qxe5 Qxd3 23.Nf5! Bxf5 24.Bh6
f6 25.Qxe7 Rf7 26.Qe3 Qxe3 27.Bxe3
Bxg4 28.f3 Bf5 ½–½
This leads to positions in which
Black has compensation for a pawn.
Game 42 David Bronstein introduced a piece
P.Spitz–V.Piccardo sacrifice in this position with 8.dxe4,
Correspondence, 2004

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5

d5 5.exd5 Na5 6.d3

although it is not now thought to give

White enough compensation. After
8...Nxc4 9.Qd4( 9.Qe2 Nb6 10.c4
Bb4+ 11.Kf1 0–0 12.a3 Re8 13.e5 Bf8
14.h3 c6 15.Be3 Nxc4 16.Qxc4 Nxd5
Morphy's old move, which hasn't was good for Black in De Zeeuw-
been seen much since the century be- Timmerman, Dutch Team Champion-
fore last. Still, you might have it ship 1992) 9...Nd6 10.Nc3 c6 11.0–0
played against you by an enterprising cxd5 12.e5 Nf5 13.Qd3 Ne4 14.Nxd5
GM like Ian Rogers. Nc5, White had inadequate compensa-
6...h6 7.Nf3 e4 tion for the piece in L.Bronstein-
Rogers-Jonkman, Wijk aan Zee Ra.Garcia, Mar del Plata Zonal 1969.
2003, featured 7...Bd6, and after 8.Nc3 Note that Luis Bronstein is not a rela-
0–0 9.Qe2 Bg4 10.h3 Bxf3 11.gxf3 c6 tion of the famous Russian GM with
12.dxc6 Nxc6 13.Be3 Nd4 14.Bxd4 whom he shares the same surname.
8...Nxc4 9.dxc4 Bc5

114
 Two Knights Defence

to prove after 12.Nc3) 12.Nc3 bxc4

13.Qxc4 Qd6 14.0–0 Ba6 15.Nb5

10.c3
Preparing to put the knight on d4, al-
though this uses valuable time.
White has tried a few other moves was Bird-Chigorin, London 1883,
here but in every case it looks as if when simply 15...Qxd5 16.Qxd5 Nxd5
Black has a strong initiative: would have left Black with a clear
10.Nfd2 0–0 11.Nb3 Bg4 plus.
10.Bf4 0–0 11.Nfd2 Bg4 12.Qf1 c6
13.h3 Bh5 14.g4 Bg6 15.Nc3 Bb4
16.dxc6 bxc6 17.0–0–0 Bxc3 18.bxc3
Qa5

12.Qf1 Bb4+ 13.c3( 13.Nc3 c6 14.h3

Bh5 15.g4 Bg6 16.dxc6 bxc6 17.Bd2
e3! 18.fxe3 Bxc3 19.bxc3 Bxc2 left
White's position in tatters in Luckis-
Keres, Buenos Aires 1939) 13...Be7 put White's king in desperate trouble
14.h3 Bh5 15.Be3( 15.g4 Bg6 16.Be3 in Van der Weide-Medina Garcia,
Nd7 17.N1d2 Ne5 18.0–0–0 b5 Amsterdam 1967.
19.cxb5 Nd3+ 20.Kb1 Qxd5 was good 10...b5!?
for Black in Salwe-Marshall, Vienna 10...b5 has been regarded as the
1908) 15...c6 16.g4 Bg6 17.dxc6 bxc6 theoretical continuation but the simple
18.N1d2 Qc8 and White's king was 10...0–0 might also be good. Wad-
short of a safe home in Popa-Godena, dingham-Hebden, British Champion-
Verona 2005. ship, Southport 1983, continued
10.h3 prevents anything landing on 11.Nd4 Bg4 12.Qc2 Bxd4 13.cxd4 b5
g4 and creates 'luft' for the knight. 14.cxb5( or 14.b3 c6) 14...Qxd5
Now 10...0–0 11.Nh2 b5( 11...Nh7 15.Qc5 Rfd8 16.Qxd5 Rxd5 17.Be3
was played in one game by the legen- Rxb5 18.b3 Nd5 and Black won the
dary Paul Morphy, but Black has it all endgame.

115
Play 1.e4 e5!

11.cxb5

The point of Black's play - with the

bishop on a6 White will lose the capa-
This looks like a new move, but it bility to castle. Of course castling isn't
isn't one that is particularly threatening everything and there are times in this
for Black. game that Black might have wished he
11.b4 Be7 12.Nfd2 Bg4 13.f3( had more pawns.
13.Qe3 bxc4 14.0–0 Qxd5 15.Re1 Bf5 17.a7 Ba6 18.a4 Bd3 19.Na3 Rxa7
16.f3 Qd3 was also very good for 20.Nab5 Ra6 21.Nxd6 Qxd6 22.Ne2
Black in Torres-Uralde, Argentina Rfa8 23.0–0
2001) 13...exf3 14.gxf3 Bh5 15.cxb5
So White gets castled anyway and
0–0 16.0–0 Re8 gave Black a danger-
Black wins back the material. It is al-
ous initiative in Grob-Keres, Dresden
ready looking rather equal.
1936.
23...Rxa4 24.Rxa4 Rxa4 25.Re1 Nd5
11...0–0 12.Nd4 Qxd5 13.Be3 Bg4
26.Nc1 Ra8 27.b4
14.Qd2
27.Nxd3 Nxe3! 28.Rxe3 Ra1+
29.Re1 Rxe1+ 30.Qxe1 exd3 gives
Black chances in the queen ending
thanks to the passed d-pawn.
27...Nxe3 28.Qxe3 Qd5 29.Qf4 c6
30.Nxd3 exd3 31.c4 Qd7 32.Rd1 Rd8
33.Qe5
Preparing the advance of the b-pawn.
33...Qd4 34.Qxd4 Rxd4 35.b5 Rxc4
½–½

14…a6!?
Game 43
There's a case to be made for
14...Bxd4 15.Bxd4( or 15.Qxd4 Qxb5) M.Kozakov–H.Jonkman
15...e3! 16.fxe3 Ne4 when Black's Lvov, 2001
pieces become very active.
15.bxa6 Bd6 16.h3 Bc8 1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Bc4
Nf6 5.e5 Ne4!?
(next diagram)
(next diagram)

116
 Two Knights Defence

0 as unclear; Black might contemplate

14...f5!? on his next move and then
transfer the queen to h5 via e8) 9...Bc5
10.c3 0–0 11.Bc2 f6! 12.cxd4?! Ncxd4
13.Nxd4 Nxd4 14.Rxd4 Bxd4 15.Qd3
Bxf2+! 16.Kxf2 fxe5+ 17.Kg1 Bf5 and
Black went on to win in Voigt-
Mikhalchishin, Dortmund 1992.
6.0–0

This has been the choice of many top

players, and it does seem much simp-
ler than the popular 5...d5.
6.Bd5
White has two other lines:
6.Qe2 Nc5

6…d5( 6...Be7 might leave White a

bit better after 7.c3 d5 8.Bb3 Nc5
9.cxd4 Nxb3 10.Qxb3, as in Kuprei-
chik-Beliavsky, USSR Championship,
Frunze 1981) 7.exd6( after 7.Bb5
Black can transpose to the 5...d5 line
by playing 7...Bc5, but I like Black's
treatment in the game Pachman-
7.0–0( 7.c3 d3 8.Qe3 Be7 9.b4 Ne6 Gligoric, Leipzig Olympiad 1960:
10.Bxd3 d6 11.exd6 Bxd6 12.0–0 0–0 7...Bg4 8.h3 Bxf3 9.Qxf3 a6 10.Bxc6+
13.Bc2 Nf4 was equal in Abergel- bxc6 11.Nd2 Ng5 12.Qd3 Ne6 13.f4
Fressinet, Port Barcares 2005) g6 14.f5 gxf5 15.Qxf5 Qd7 16.Nf3 h6
7...Ne6!( an improvement on the older 17.Bd2 c5 18.Nh2 h5 19.Rf2 0–0–0
7...Be7 that makes it difficult for 20.Qxf7 Be7 and a complex, double-
White to get the pawn back) 8.Rd1( edged situation arose in which Black's
8.Bxe6 dxe6 9.Rd1 Be7 10.Nc3 0–0 chances were certainly not worse)
11.Qe4 f5 12.Qe2 a6 13.Be3 Qe8 to 7...Nxd6 8.Bd5 Nf5 9.Re1+ Be7
complex play in Castany Pampalona- 10.Bxc6+ bxc6 11.g4 Nh6!( 11...Nd6
Narciso Dublan, Spain 1999; while 12.Nxd4 Qd7 13.Bg5 f6 14.Bf4 Kf7
8.c3 d5 9.exd6 Bxd6 10.Bg5 Be7 15.Nc3 Re8 16.h3 Bf8 17.Qf3 Bb7
11.Bxe7 Qxe7 12.Re1 0–0 13.cxd4 18.Na4 was good for White in Svesh-
Qf6 leaves Black very comfortably nikov-Arkhipov, Elista 1994)
placed according to analysis by Emms) 12.Qxd4( 12.Bxh6 gxh6 13.Nxd4 c5
8...d5! 9.Bb3( Mikhalchishin sug- 14.Qf3 Bxg4 15.Qxg4 cxd4 16.Qg7
gested 9.Bb5 as an improvement, giv- Rf8 17.Na3 Qd7, intending...0–0–0,
ing the variation 9...Bc5 10.c3 Bd7 was fine for Black in Filipovic-
11.Bxc6 Bxc6 12.cxd4 Bb6 13.Nc3 0– Blagojevic, Neum 2004; and 12.Bg5

117
Play 1.e4 e5!

Be6 13.Bxe7 Qxe7 14.Qxd4 Nxg4 Bxc5 13.Bb3 d5 14.a3 Na6 as in Hec-
gave Black counterplay in Sveshni- tor-Nunn, Vejle 1994.
kov-Smikovski, Togliatti 2003) 7...dxc6 8.Qxd4
12...Bxg4 13.Bxh6( 13.Qe4 is well
8.Nxd4 Ne6 9.Be3 Nxd4 10.Qxd4
met by 13...Bxf3 14.Qxf3 Qd5 15.Qe2
Qxd4 11.Bxd4 Bf5 12.c3 0–0–0 was
Nf5, intending to castle) 13...Qxd4(
excellent for Black in A.Minasian-
Black might also consider 13...Bxf3!?
Mainka, Candas 1992.
14.Qxg7 Kd7 with a wild position)
14.Nxd4 gxh6 15.Nxc6 Be6 16.Nxe7 8...Bf5
Kxe7 with an even-looking endgame,
Black's active pieces compensating for
the doubled h-pawns.
6.c3 d5 7.exd6 Nxd6

Black stands very well here, with the

bishop pair in an open position.
9.Qc3
9.Bg5 Be7 10.Bxe7 Qxe7 11.Qc3 0–
is good for Black thanks to the gain 0 12.0–0 Rad8 13.Nbd2 Na4 14.Qc4
of time on the bishop on c4. Nb6 15.Qb3 Rd5 was good for Black
6...Nc5 7.Bxc6?! in Upton-Mikhalchishin, Mexico 1977.
This trade is quite harmless. Besides 9...Ne6 10.Be3 Qd5! 11.Nbd2 Be7
7.0–0 and 7.Qe2( see the next game)
there is 7.c3, but then 7...dxc3 8.Nxc3
Be7 9.Be3 0–0

12.a3
With the maniacal intention of cas-
tling long. White should admit the fact
leaves White with rather nebulous that his opening has been a disaster
compensation for the pawn, for exam- and play the modest 12.0–0.
ple 10.Qe2 Nb4 11.0–0–0 c6 12.Bxc5

118
 Two Knights Defence

12...0–0 13.0–0–0 c5 14.Nb3 Qc6

15.Rd2 Rfd8 16.Rhd1 Rxd2 17.Rxd2
a5!

Threatening both a fork on e2 and

the simple capture of White's e-pawn.
White opts for a radical solution but
finds himself in even more trouble.
Threatening to drive White's knight 27.Bxd4?! cxd4 28.Qxc7 Bxa3+
to the miserable a1–square with 18...a4 29.Kb1 Qh1!
whilst angling for a queenside pawn
The pin is deadly. And Black's back
storm. Of course White can take it, but
rank isn't a problem because he can
that opens lines in front of his king.
drop the bishop back to f8.
18.Nxa5 Qa6 19.Nb3
30.Qc4 Bh3 31.e6 fxe6 32.b4 Qxf1
19.Nc4 loses a knight after 19...b5. 33.Qc8+ Kf7 34.Qd7+ Kg6 0-1
19...Qf1+ 20.Rd1 Qxg2 21.Ne1 Qc6 White is running out of checks and
he's about to lose the knight on e1.

Game 44
V.Gurevich–O.Romanishin
Herson, 1989

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.d4

exd4 5.e5 Ne4!? 6.Bd5 Nc5 7.0–0

The movements of Black's queen

remind me a little of snooker, her ma-
jesty bouncing off several cushions
before finishing back where she
started. Meanwhile Jonkman's advan-
tage has increased as White's pawn
structure is wrecked and Black has an
open a-file.
22.Nd2 Rd8 23.Nf1 Rxd1+ 24.Kxd1
Bg4+ 25.Kc1 Qd5 26.b3 Nd4!
7.Qe2 Be7 8.0–0 transposes back in-
to the game.; 7.Bxc6?! was seen in
(next diagram)
Kozakov-Jonkman( Game 43).
119
Play 1.e4 e5!

7...Be7 8.Qe2 White stands somewhat better after

White can also play 8.Nxd4, for ex- 9...d6 10.Bxc6 bxc6 11.Nxd4 Bd7
ample 8...Nxd4( the risky 8...Nxe5!? 12.Bf4; while 9...Ne6 10.c3! dxc3
was tried in Cohen-Flear, Isle of Man 11.Nxc3 d6 12.Ne4 gives him pressure
1994, after which 9.f4 Nc6 10.Nf5 Bf6 for the sacrificed pawn.
11.Qg4 Kf8 12.Ne3 d6 13.Qd1 Ne7 10.Bf4
eventually saw Black consolidate; this Trying to prevent Black from effec-
might be worth trying if you want to tively moving his d-pawn.
try and win as Black, although it does Alternatives don't promise much:
look very dangerous) 9.Qxd4 10.Bxc6 d3!( better than 10...dxc6
11.Nxd4 Bd7, which is also not bad)
11.cxd3 dxc6 12.d4 Ne6 followed
by...b6,...Bb7 and...Rd8 is very good
for Black.
10.Na3 Nb4 11.Bc4 d3 12.cxd3 d5
13.exd6 Bxd6 14.Qxe8 Rxe8 15.Nb5
Bg4 and the position was about even
in Khelminitsky-Malaniuk, Sibenik
1990.
10...Ne6

9…0–0( 9...c6 10.Bc4 0–0 11.Rd1

b5 12.Be2 Ne6 13.Qe4 f5 14.Qd3 Qc7
15.f4 left Black shy of equality in Er-
wich-Romanishin, Hoogeveen 2004)
10.Be3 d6 11.Qc3 Nd7( Emms' sug-
gestion of 11...Na4 is interesting as the
knight can be dropped back to b6)
12.f4 Nb6 13.Bf3 c6 14.Rd1 d5 and
Black had equalized in Fahrni-
Schlechter, Baden-Baden 1914.
8...0–0 9.Rd1 Qe8! This is a logical move, but a year lat-
er Romanishin had a change of heart.
Khelminitsky-Romanishin, Sibenik
1990, varied with 10...b6 11.Bxc6 d3
12.cxd3 dxc6 13.d4 Ne6 14.Bg3 Bb7
15.Nc3 Rd8!, intending...Rd7 followed
by...Qa8 and...Rfd8, with strong pres-
sure against d4 and the possibility of
opening up for the light-squared bi-
shop with...c6-c5.
11.Bg3 Kh8?!
The ever-ambitious Romanishin
A clever move, getting the queen wants to move his f-pawn and get his
away from the gaze of White's rook on queen into play on the kingside, but
d1.
120
 Two Knights Defence

this does leave the cramping pawn on Not 16.Nb3? b5! 17.Qxb5 Rb8
e5 intact. 18.Qc4 Rb4 and Black wins either the
Objectively stronger is 11...d6! queen or the bishop on d5.
16...a6?!
Probably not the best. Here Gurevich
and Schneider give 16...Na5 17.Qb5
a6 18.Qxa5 as good for White, but
their analysis seems wrong. Black can
play the amazingly calm 18...d6!!(
their 18...c6 should be met by 19.Re5
Bxe5 20.Bxe6 etc.), when 19.Ne5(
19.Ne4 b6 20.Nxf6 Rxf6 21.Qd2 Qxd5
is simply good for Black) 19...Bxe5
20.Bxe5 dxe5 21.Bxe6 Bxe6 22.Qxe5
after which 12.exd6 Bxd6 13.Bxd6 is equal.
cxd6 14.Bxe6 Bxe6 15.Nxd4 Nxd4 17.a5 Ne7?!
16.Rxd4 Bxa2! 17.Nc3 Qxe2 18.Nxe2 17...Ncd8 can be answered by
Rfe8 19.Nc3 Be6 20.Nb5 would lead 18.Bxe6 Nxe6 19.Bxc7 d5 20.Qb4,
to an equal position. However, I doubt preventing the capture of the bishop
that Black could win such a game. because the rook on f8 would hang.;
12.Nbd2 But maybe 17...Qf5 was worth consi-
If 12.Na3 Black could play 12...Nb4 dering.
13.Bc4 a6! 14.Nxd4 b5 with double- 18.Bxe6 dxe6 19.Qxc7 Nc6 20.Qb6
edged play. I don't agree with the as- Qd5
sessment of Gurevich and Schneider Or 20...Bd8 21.Bc7.
who claim that Black is slightly better
- I would say it's equal after 15.Qe4 21.Qb3 Bd7?! 22.Nc4?!
Rb8 16.Be2.
12...f5 13.exf6 Bxf6 14.Re1 Qg6

White should play the cold-blooded

22.Qxb7, when 22...Rad8( or 22...e5
23.Nxe5!) 23.Qb3 leaves Black miser-
14...Qh5 would have been strongly ably placed.
met by 15.Bxe6 dxe6 16.Bxc7; but
14...d3 15.cxd3 Bxb2 16.Rab1 Bc3 22...Rad8 23.Nfd2 Bc8 24.Nb6 Qxb3
was worth considering. 25.Nxb3 Nb4 26.Re2 Nd5
15.Qc4 Qh5 16.a4 26...d3 doesn't help much after
27.cxd3 Nxd3( or 27...Rxd3 28.Nc5)
121
Play 1.e4 e5!

28.Bc7, forcing Black's rook to leave Bb6 leaves White down but not com-
the d-file. pletely out.
27.Nc4 Bd7 28.Nd6 36...Nxg3+ 37.hxg3 d3! 0–1
28.Rd2 is also good. This time 37...d3 38.Ra2 dxc2
28...Bc6 29.Nc5 39.Nd3 can be answered by 39...Rxf2!,
which is why White had to put the
Gurevich and Schneider give
king on f1.
29.Rxe6! Nb4 30.Re2 d3 31.cxd3
Nxd3 as unclear, but 32.Rd1 Nb4(
32...Nxb2? 33.Rxb2 Bxb2 34.Nf7+)
Game 45
33.Nc5 still looks horrible for Black.
N.Karaklajic–S.Jovanovic
29...e5 30.Ndxb7?! Palic, 1996
This allows Romanishin to finally
get some counterplay. 30.Nde4 main-
tains White's edge. 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.d4
30...Bxb7 31.Nxb7 Rb8 32.Nc5 Rxb2
33.Nd3?
White is falling apart. 33.Bxe5 is the
best, when 33...Nc3 34.Ree1 Bxe5(
not 34...Rxc2? 35.Bxf6 gxf6 36.Ne6,
hitting f8 and d4) 35.Rxe5 Rxc2 36.f3
h6 leaves the outcome in the balance.
33...Nc3! 34.Rxe5
White has no good move. 34.Rd2 is
answered by 34...Rb5; 34.Ree1 by
34...Rxc2 35.Bxe5 Bxe5 36.Rxe5 Rd2; The more sober 4.d3 features in the
and 34.Nxb2 by 34...Nxe2+ 35.Kf1 game Dolmatov-Romanishin( Game
Nxg3+ 36.hxg3 e4 46).
34...Bxe5 35.Nxb2 Ne2+ 4...exd4 5.0–0
White completes kingside develop-
ment and offers a second pawn. But
Black can take it and still reach a nice
position.
5...Nxe4!

36.Kh1?
36.Kf1 is the best chance, when
36...Nxg3+ 37.hxg3 d3 38.Ra2 dxc2
39.Nd3 Rc8 40.Nc1 Rc5 41.Ke2 Bc7
42.Kd3 Bxa5 43.Rxc2 Rxc2 44.Kxc2
122
 Two Knights Defence

6.Re1 11.Nb5( 11.Ne4 Qxd4 12.Qxd4 Rd8)

White has also tried 6.Nc3?! but then 11...Rc8 12.Bf4( or 12.Nxa7 Bc5
simply 6...Nxc3 7.bxc3 d5 8.Bb5 Be7 13.Rf4 Rd8 as in Lembeck-Klasmeier,
9.Ne5( 9.Nxd4 Bd7) 9...Bd7 10.Nxd7 German Bundesliga 1987) 12...Bc5
Qxd7 11.cxd4 a6 12.Ba4 b5 13.Bb3 13.Re4 Bb6 14.Qe2 0–0 15.Be5 Qg6
Na5 Black a good pawn up in left Black a pawn up for nothing in
S.Sokolov-Grodzenski, Correspon- Mindeguia Guruceaga-Estemera,
dence 1976. Pamplona 1995.
Equally unimpressive is 6.Bd5, when 7...Qxd5 8.Nc3 Qh5
6...Nf6 7.Ng5( 7.Bg5 Be7 8.Bxf6 Bxf6
9.Re1+ Ne7 10.Nxd4 0–0 and Black
had a sound extra pawn in Venkatara-
manan-Gokhale, Calcutta 1994)
7...Nxd5 8.Re1+ Be7 9.Qh5 g6
10.Qf3( or 10.Qh6 d6 11.Qg7 Rf8
12.Nxh7 Be6) 10...0–0 11.Qxd5 Nb4
12.Qb3 d5 does not give White what
I'd describe as an attacking position.
6...d5

9.Nxe4
After 9.Rxe4+ Be6 10.Nxd4 Qxd1+
11.Nxd1 0–0–0 White gets into trouble
on the d-file.
9...Be6 10.Bg5 Bd6

7.Bxd5
White also has the Canal Variation
with the amazing 7.Nc3?!, but then
7...dxc4 8.Rxe4+ Be6 9.Nxd4 Nxd4
10.Rxd4 Qf6

11.Bf6
This meets with a neat refutation
which essentially puts this opening out
of business as a winning attempt.
White has to try something else, but
what? Here are the alternatives:
11.Nxd6+ cxd6 12.Bf4 Qd5

(next diagram)

123
Play 1.e4 e5!

13.c3( 13.Qd2 0–0 14.b3 Qc5 and White was a pawn down with
15.Rac1 Bg4 and White had to regret kingside weaknesses in Aleksic-
not trying to recapture the pawn more Pavlovic, Becici 1993.
directly in Sorri-Ornstein, Helsinki 11...Bxh2+!!
1990) 13...Kd7!( Black's king is per-
fectly safe here) 14.Nxd4 Nxd4
15.Qxd4( 15.cxd4 g5 16.Be3 h5
17.Qa4+ b5 18.Qa5 h4 19.Rac1 Rhc8
20.a4 b4 21.Rxc8 Qxa5 22.Rxa8 Bd5
was also good for Black in Nystrom-
Hector, Stockholm 2001) 15...Qxd4
16.cxd4 Bd5 17.Bd2 a5 18.a3 b5 19.f3
Rhg8 20.Kf2 g5 also left White fight-
ing for the draw in N.Thomas-Hector,
Bled 2002.
11.c4 0–0 12.c5 Be5 13.Nxe5 Qxd1
14.Raxd1 Nxe5 15.Rxd4 Nc6 16.Ra4 A stunning blow which secures
Rfe8 17.Bf4 Bd5 Black an advantageous endgame.
After 11...0–0 12.Nxd6 cxd6
13.Bxd4 White gets his pawn back
without suffering any harm.
12.Nxh2
White has no choice. Both 12.Kf1
Bc4+ 13.Re2 0–0 and 12.Kh1 Be5+
13.Bh4 0–0 leave Black in a winning
position.
12...Qxd1 13.Raxd1 gxf6 14.Nxf6+
Kf8! 15.Nf3 Rd8 16.Ng5 Bf5
White's knights look dangerous but
was better for Black in COMP Deep they don't have a permanent outpost.
Junior-Smirin, internet match( rapid) As a result they inevitably get driven
2002. back.
11.h4 h6 12.Nxd6+ cxd6 13.Bf4 Qd5 17.Rd2 Kg7 18.Nge4 h5! 19.f3 Rh6
14.Rc1 0–0 15.Qd2 Rfe8 16.c3 dxc3
20.g4
17.Qxc3 Bg4
Or 20.Nxh5+ Rxh5 21.g4 Ne5 etc.
(next diagram)
124
 Two Knights Defence

20...hxg4 21.fxg4 Bxe4 22.Nxe4 Ne5

23.Kg2 Rdh8 24.Rxd4?
Losing the exchange. The only way
to keep the game going was with
24.Nf2, although Black could then
keep his extra pawn with 24...Nc6.
24...Rh2+ 25.Kg3 R8h3+ 26.Kf4 Nf3
27.Red1 Nxd4 28.Rxd4 Rxc2 29.Rd7
Rxb2 30.Rxc7 Rb4 31.Re7 Ra3
32.Kg5 Ra5+ 33.Kf4 Rxa2 34.Kf3
Ra3+ 0–1 8.c3( 8.a5 a6 9.c3 Ng8 10.Qb3 f5
11.exf5 Rxf5 12.Nbd2 Nf6 13.Be6
Bxe6 14.Qxe6 Qd7 15.Qxd7 Nxd7
Game 46 was equal in Tkachiev-Fressinet, Cap
S.Dolmatov–O.Romanishin d'Agde 2002) 8...Bg4 9.h3 Bh5
USSR Championship, Minsk, 1979 10.Nbd2 d5 11.exd5 Nxd5 12.Nf1 f5
13.Ng3 Bg6 14.Nxe5 Nxe5 15.Rxe5
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.d3 Nb6 16.Qf3 f4 17.Nf1 Bd6 18.Re1
Be8!, intending 19...Bc6, which gave
him a good game in Grosar-I.Sokolov,
Bled 1991.
5...0–0 6.0–0 d6 7.c3 Na5 8.Bc2 c5
9.Nbd2

This quiet move has become the

most popular way of playing the Ital-
ian Game for White. It can lead to po-
sitions much akin to the Closed Ruy
Lopez.
4...Be7 5.Bb3 Instead, 9.b4 cxb4 10.cxb4 Nc6
5.0–0 0–0 6,Bb3 leads back to the 11.b5 Na5 12.Bb2 Bg4 13.h3 Bxf3
game, but White has an independent 14.Qxf3 Re8 15.Qe2 Qb6 is fine for
alternative in 6.Re1( 6.Nc3 d6 7.a3 is Black.
well met by 7...Be6) 6...d6 7.a4( 7.a3 9.a3 Nc6 10.b4 a6 11.Nbd2 Re8
is similar) and with the bishop ready to 12.Bb2 Bf8 13.Re1 Nh5 14.Nf1 g6
drop back to a2, the...Na5 plan is no 15.Ne3 was Bauer-Hebden, Andorra
longer effective. Consequently Black Zonal 1998, and now Hebden after-
should switch plans here and play wards thought he should play 15...Bg7
7...Kh8 with the idea of a later...f7-f5.
(next diagram) 9...Qc7 10.Re1 Be6 11.Nf1 Rad8

125
Play 1.e4 e5!

Mark Hebden advised me that this 13.Qe2

was the best way to play for Black, Exerting indirect pressure against e5.
delaying the return of the knight to c6. After 13.h3 Black can play 13...d5!,
He also said that when White plays for example 14.exd5( 14.Qe2 d4 15.c4
Ng5 you just let him eliminate the bi- g6 16.Bh6 Rfe8 17.Qd2 Bf8 18.Bxf8
shop. I'm sure he's right, though I don't Rxf8 19.Qh6 Nb4 was fine for Black
fully understand the reason for delay- in the game Dolmatov-Timman, Ams-
ing the knight retreat. terdam 1980) 14...Nxd5 15.Qe2( Yu-
The immediate 11...Nc6 seems play- dasin assessed both 15.Ng5 Bxg5
able, for example 16.Bxg5 f6 17.Bd2 Rfe8; and 15.Ba4
f6 16.d4 cxd4 17.cxd4 Nb6 as unclear)
15...f6 16.Bd2 Kh8 17.a3 Rfe8 18.d4
cxd4 19.cxd4 Bf8 20.Qe4( 20.dxe5
Nxe5 21.Nxe5 Qxc2 favours Black)
20...Bg8 21.dxe5 Nxe5 22.Ba4 Nxf3+
23.Qxf3 ½–½, Yudasin-Beliavsky,
USSR Championship, Moscow 1988.
13...Rfe8
In Turner-Hebden, British Cham-
pionship, Hove 1997 Black played the
unusual-looking 13...Kh8, after which
12.Ne3( 12.Ng5 Bg4 13.f3 Bd7 14.f4 14.h3 h6 15.Nh4 Nh7 16.Nhf5 Bg5
Bg4 15.Nf3 exf4 16.Bxf4 was a tad 17.Bxg5 hxg5 18.Ne3 g6 19.Bb3 Bc8
better for White in J.Howell-Finegold, 20.Nd5 was good for White.
London 1988) 12...Rad8 13.Ng5 h6 Romanishin's move makes more
14.Nxe6 fxe6 15.Bb3 Qd7 16.Nf5 Kh8 sense to me, lending indirect support
17.Nxe7 Qxe7 18.f4 b5 19.fxe5 Nxe5 to e5 and 'shadowing' White's queen.
20.d4 Ng6 21.Qe2 c4 22.Bc2 and a 14.Ng5 Bg4 15.f3 Bc8 16.Bb3 Rf8
draw was agreed in Sandipan- 17.f4
Ganguly, Visakhapatnam 2004.
Having put the pawn on f3 this is the
12.Ng3 Nc6 only logical way to continue.
17...Bg4 18.Nf3 exf4 19.Bxf4 Ne5
(next diagram) 20.d4
Claiming space in the centre, but the
d-pawn is also vulnerable.

126
 Two Knights Defence

20...Ng6 21.Bd2 cxd4 22.cxd4 Qb6 40.Ne2?!

23.Qf2 Be6 24.d5 After this I prefer Black. White
should try 40.Be2, when 40...Rc2
41.b3 Rb2 42.Bc4 holds things togeth-
er. Less good is 40.Kf2 in view of
40...Bg4 41.Ke1 Rc1 42.Ne2 Rb1
when White is in trouble.
40...Rc4 41.Rxd6 Rxe4 42.h3 Bb5
43.Rd8+ Kg7 44.Kf2 f5
Black could also play 44...Nd3+ first
and after 45.Kf3 support the rook with
45...f5. I don't see much difference to
the game.
If White has to play this he's got 45.b3?
nothing - with the use of the e5-square
for his knight Black has little to fear in Now White is losing. His last attempt
the endgame. The only try for some- to stay on the board was with 45.Rb8
thing concrete was with 24.Nf5, but Nd3+ 46.Kf1 Nxb2 47.Rxb5 Nxd1
then 24...Bxb3 25.axb3 Rde8 lines up 48.d6 Kf6 49.d7 Ne3+( 49...Ke7
against e4. 50.d8Q+ Kxd8 51.Rd5+) 50.Kg1 Ke7
51.Ng3 Re6 52.Rxb4 when the clear-
24...Qxf2+ 25.Kxf2 Ng4+ 26.Kg1 ance of all the queenside pawns draws.
Bc8 27.Bc3 N4e5 28.Bd4 b6 29.a4
45...Nd3+ 46.Kf3 Ne5+
Or 29.Nxe5 Nxe5 30.Rec1 Bg5
31.Rc7 Rd7 with equality. Black repeats moves to gain time on
the clock.
29...Bf6 30.Nxe5 Bxe5 31.Bxe5 Nxe5
32.a5 g6 33.axb6 axb6 34.Rec1 Bd7 47.Kf2 Nd3+ 48.Kf3 Ne1+ 49.Kf2
35.Ra6 Rb8 36.Ba4 Rfc8 37.Rc3 b5 Nxg2! 50.Ng3 Rd4 51.Bf3
38.Bd1 b4 39.Rxc8+ Rxc8 51.Be2 would have been more stub-
born, although Black still seems to be
winning after 51...Rd2 52.Kf3 Bxe2+
53.Nxe2 g5 etc.
51...Rd2+ 52.Kg1 Nf4 53.Rb8 Nxh3+
54.Kh1 Nf2+ 55.Kg1 Nh3+ 56.Kh1
Ng5 0–1

127
Play 1.e4 e5!

Summary

4.Ng5 leads to very complex positions in which Black gets ongoing compensa-
tion for the sacrificed pawn. 4.d4 doesn't cause Black much trouble after either
4...exd4 5.e5 or 5.0-0 Nxe4, while 4.d3 is similar to a Closed Spanish.

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5

4.d3 - Game 46
4.d4 exd4
5.0-0 - Game 45
5.e5 Ne4 6.Bd5 Nc5(D)
7.Bxc6 - Game 43
7.0-0 - Game 44
4...d5 5.exd5 Na5(D) 6.Bb5+
6.d3 - Game 42
6...c6 7.dxc6 bxc6 8.Be2
8.Qf3 – Game 41
8...h6(D) 9.Nf3 - Game 39
9.Nh3 - Game 40

6…Nc5 5…Na5 8…h6

128