The Kibitzer
by Tim Harding

The Vampire Gambit: Can We Bury It Now?

THERE are quite a few astonishing gambits around which don't
feature much in traditional opening books. For example, there is 4
g4 against the Gruenfeld Defence, 1 d4 Nf6 2 c4 g6 3 Nc3 d5 4 g4
(the Gibbins-Weidenhagen Gambit), the Shabalov Gambit against
the Semi-Slav (which also involves an early g2-g4 which can be
met by Nxg4), and of course the Winckelmann-Reimer Gambit
about which I've written in my newest book Four Gambits To Beat
the French (Chess Digest) and elsewhere.

Perhaps even more astonishing is the topic of this month's article,
the so-called Halsz Gambit where the deviation from recognised
paths comes as early as move three1 e4 e5 2 d4 exd4 and now 3
f4!? (see diagram) instead of the well-known 3 c3 (Danish Gambit)
and Centre Game (3 Qxd4, revived in recent years by some OTB
masters) and the rarer 3 Nf3. (See Diagram)

This article was far advanced when Hanon Russell pointed out that
Glenn Budzinski also wrote about the gambit a few months back at
The Chess Caf. [Glenn's article is available in The Chess Caf
Archives.] Sorry, Glenn, I'd overlooked that or I might have
chosen a different topic, but anyway I have a lot to add. The gist of
this article was originally to have been a section in my book
Winning At Correspondence Chess which the publishers cut out
because of length considerations.

Strictly speaking, the name Halsz Gambit is not right since why
give the name of a modern Hungarian postal player to a gambit that
has been around since the 1840s, if not earlier? However, no name
seems previously to have stuck to this bizarre idea 3 f4, and Halsz
has played numerous games with it over the past few decades, and
has even experimented with an analogous idea against the Sicilian
Defence1 e4 c5 2 d4 exd4 3 f4!?.

Maybe it should instead be called the Vampire Gambit because it is
dead yet refuses to stay buried. Can a reader finally produce the
conclusive refutation to put a stake through its heart?

You will not find 3 f4 in the Center Game mentioned in "Nunn's
Chess Openings", "BC02", "Modern Chess Openings" (12th
edition), "Open Gambits" (Botterill), "Unorthodox Openings
(Benjamin & Schiller), or "The Complete Book Of Gambits"
(Keene). Either the authors and editors did not know of the move
(probable) or they just dismissed it as obviously bad.

Even the second edition of ECO did not have it, and as for the
latest (third) edition I am mystified because Budzinski's article
refers to a supposed ECO recommendation which I cannot find in
my copy! If it exists, it should be under opening code C21, note 2
(pages 174-175) but the only moves I see mentioned there are 3
Nf3 and 3 Bc4 with 3 c3 as the main line and 3 Qxd4 in C22. So I
guess Budzinski was looking at the first edition, dating from the
1970s, where C21 was edited by Keres, but I passed on my copy of
that edition years ago.

Anyway, Paul Keres did mention the move in his
"Dreispringerspiel bis Knigsgambit", published by Sportverlag of
East Berlin in 1971; he did not think much of it and that, I believe,
would be the instant opinion of any grandmaster. However, the
gambit has won some games, as we shall see in a moment.

Budzinski seems to know little or nothing about Halsz and some
European articles about the gambit as he only refers to one
American source, the "Chess Horizons" magazine. Dr. Gyorgy
Halsz of Budapest rediscovered 3 f4 some years ago but he didn't
invent it. In fact, it dates from the uninhibited romantic days of the
last century - as the English chess magazine, "Kingpin" (issues 19
and 20) revealed, following up a 1990 survey in "Fernschach" by
German correspondence chess veteran Herman Heemsoth. More
recently in "Fernschach" (1996), Jerzy Konikowski wrote another
article on it. 

Most recently, both my own magazine "Chess Mail" and the
German "Kaissiber" have published a recent win by Halsz against
a Swiss correspondence chess international master, Dr Xaver
Steiner.

Halsz himself doesn't hold the master title but with a current
ICCF rating of 2312 based on over 300 games, he is evidently a
dangerous attacking player who wins a high proportion of his
games with White. Here is that recent game

Halsz-Steiner EU/M/1203, corr 1998
 1 e4 e5 2 d4 exd4 3 f4!? Bc5 

Keres wrote that "3...Bb4+, 3...Nc6 or 3...d5 are also good."

4 Nf3 Nc6 5 Bd3 

Halsz follows the main line from the book by Keres. Frank
Marshall was the source of this line, but it is not clearly Black's
best plan; perhaps Keres (lacking games to quote) found it easier to
cite this authority than looking for the very best defence as he
would surely have done had anybody played 3 f4 against him. 

Actually I was surprised to see Halsz adopt the 5 Bd3 and 6 Nbd2
move order, because in earlier years he tended to play the move a2-
a3 at move five or six, threatening to expand on the queenside with
gain of time. This is really a revival of an old idea seen in
F.Schwenkenberg-W. Vitzthum, Dusseldorf 1861, which went 1 e4
e5 2 d4 exd4 3 f4!? Nc6 4 Nf3 Bc5 5 Bd3 Nge7 6 a3 0-0 7 b4 Bb6
8 Bb2 a6 9 Nbd2 d5 10 e5 f6 11 0-0, White winning in 20 moves.
Of course Black's 8...a6 looks very passive; ...a5 is the normal
reaction to a3.

In the present game, Halsz plays a2-a3 slightly later.

5...Nf6 6 Nbd2 d6 

About this position, Keres wrote "And White has insufficient
compensation for the pawn". Frank Marshall was in fact the source
of this line. See what Halsz does to this book recommendation
now! However, 6...d5 seems more energetic and maybe Black
should play that move earlier.

7 a3 a5 8 0 0 0 0 9 e5 Ng4 10 Ne4 Bb6 11 Nfg5 h6 12 h3 f5 13
exf6 Nxf6 14 Nxf6+ Rxf6 15 Qh5 Ne7 (See Diagram) 

16 g4 

Halsz: "A typical intuitive sacrifice of which the consequences are
incalculable."

16...hxg5 17 Qh7+ Kf8 18 Bd2 Be6 19 fxg5 Rxf1+ 20 Rxf1+ Ke8
21 Qxg7 Kd7 22 g6 d5

If 22...Qg8 Halsz indicates the continuation 23 Qh7 Rf8 24 Bh6
Rxf1+ 25 Kxf1 d5 26 Bf5 Kd6 27 Bf4+ Kd7 28 Bg5 Bc5 29 Bxe7.

23 Bg5 Bc5 24 Bf6 Kc6 25 Re1 Qd6 26 Be5 Qd7 27 b4 axb4 28
axb4 Bxb4 29 Rb1 Ra4 30 Qf6 b5 31 Bf5 Nxf5 32 gxf5 Qe7 33
Qxe6+ Qxe6 34 fxe6 1 0.

Whether 3 f4!? is actually a good move must surely be doubtful,
since - unlike most gambits - it gives White no lead in
development for his pawn, but only a slight gain in space on the
kingside. Nevertheless, as a surprise weapon it is certainly more
dangerous than Keres thought, if carefully and ingeniously
followed up. 

One of Black's (psychological) problems is that 3 f4 looks so
obviously unsound that it is hard to decide which "refutation" to
adopt. Psychologically, opponents often don't play in a stable and
rational way when they believe their game is won from a very early
stage and they tend to underestimate their opponents.

In the earliest games I have seen, Black tried to exploit the fact that
White had not made the normal Danish gambit move, 3 c3, and
tried to hold the gambit pawn by 3...Bb4+. Budzinski says that this
check "might be the toughest nut for White to crack." (See
Diagram)

For example

a) 4 c3? dxc3 5 bxc3 (5 Nxc3 Nf6) 5...Bc5 and Black's bishop is
strongly posted (Marshall); 

b) 4 Bd2 and now

b1) 4...Nc6 5 a3 or 5 Bxb4; 

b2) 4...Qe7! tended to be successful5 Bd3 d5 6 e5 Nh6 7 Nf3
Bxd2+ 8 Qxd2 c5 9 0-0 Nc6 10 Na3 a6 0-1, 27 E.Pindar-
J.S.Kipping, Manchester 1861, but better is 8 Nbxd2 as Budzinski
says. However, there is another 19th century example: 5...Nf6 6 e5
Nd5 7 Nf3 Nxf4 8 0-0 Nxd3 9 cxd3 Bxd2 10 Nbxd2 c5 (0-1, 30)
A.Epiphanoff-J.Wallenrath, St Petersburg 1852. 

The modern treatment of 3...Bb4+ is

c) 4 Nd2 when Budzinski comments "Despite White's encouraging
results... Black's play has often left much to be desired in the few
available examples."

After 4 Nd2 play can go 4...Nf6 (instead of 4...Nc6 or 4...Qe7? as
played by Halsz's own opponents in games cited by Budzinski.) 5
Bd3 d5 6 e5 Ng4 7 a3! (instead of 7 Ngf3 c5 8 Ng5 c4 as in
A.Bartsch- H.Edighoffer, IECG email, 1998) 7... Ne3 8 Qe2 Bg4 9
Ngf3 Ba5 10 b4 Bb6 11 Nb3 0-0 12 Bb2 f6 13 Bxd4 Bxd4 14
Nbxd4 fxe5 15 fxe5 Nf5 16 Ne6 winning the exchange and soon
the game (Edinburgh Chess Club-Newcastle, intercity corr ).

In the postal game De Laat-Luuring (cited in the German periodical
Gambit Revue, 1990) Black tried instead to simplify by 6...Ne4 7
Ngf3 c5 8 0-0 Nxd2 9 Bxd2 Bxd2 10 Qxd2 Qb6 but after 11 b3
Qh6 12 Rae1 Nc6 13 Ng5 Bd7 14 e6 White retained an initiative
thanks to his lead in development. Black gave up the exchange by
14...fxe6 15 Rf3 0-0-0 16 Nf7 Qf6 17 Nxd8 Kxd8 and the game
was eventually drawn.

Nave defenders might suppose they could just defend the extra
pawn by 3...c5!? although this is liable to leave holes at d5 and d6
unless Black can rapidly follow up with ...d7-d5. The May 1899
American Chess Magazine attributed the gambit to James Jellett of
St Paul citing this 3...c5 line, not knowing of games played with 3
f4 in the 1860s or earlier. (See Diagram)

In ''Marshall's Chess Openings'' (1904), the American grandmaster
said 4 Bc4 was best e.g. 4...d6 5 c3 Nc6 6 Nf3 Bg4 7 0-0 Nh6 8 f5
''and White should win''. Modern exponents of the Halsz/Vampire
gambit seem to have overlooked this recommendation as they all
play 4 Nf3 instead.

After 4 Nf3 the 100-year-old article suggested 4...Nc6 5 Bd3 Nf6 6
0-0 Be7 7 Nbd2 0-0 8 Ng5 with interesting possibilities (Jellett),
but modern players prefer 4...d6 or 4...d5. For example

a) 4...d5 allows White to establish a kingside pawn majority by 5
e5; this position also often arises via 3...d5 4 e5 c5 5 Nf3 and I deal
with this line later in the article.

b) Surprisingly perhaps, the more flexible 4...d6 seems better
although it doesn't challenge d5 at once. Blank-A.Aasum,
corr,1975, continued 5 Bd3 Bg4 6 0-0 Nc6 7 a3 Be7 8 h3 Bxf3 9
Qxf3 Qc7 10 Bd2 Nf6 11 g4 0-0-0 12 e5 Nd7 13 exd6 Bxd6 14 b4
g6 15 c3 a6 16 Ra2 Rhe8 17 b5 axb5 18 Bxb5 dxc3 19 Nxc3 Nd4
20 Qd3 c4 21 Bxc4 Bc5 22 Kh2 Nb6 23 Bb5 Nb3 24 Nd5 Nxd2 25
Nxc7 Rxd3 0-1.

Thirteen years later, Dr Halsz improved on this in a postal game
against the same Norwegian opponent with 10 Nd2 Nf6 11 Nc4
and won a messy game 11...0-0-0? 12 Bd2 d5 13 exd5 Rxd5 14 g4
Rdd8 15 b4 cxb4 16 axb4 Bxb4 17 Bxb4 Nxb4 18 Rxa7 Qc6
(18...Nc6 19 Ra8+ Kd7 20 Bf5+ Ke8 21 Re1+ Kf8 22 Qa3+ Kg8
23 g5 Nd5 24 Re8+ Rxe8 25 Rxe8#) 19 Bf5+?! Kc7 (19...Kb8 20
Qa3!) 20 Qb3 Qc5 21 Ra5 Qe7 (21...b5 22 Bd3 Ra8 23 Rfa1 with
the initiative) 22 Re5 Qf8 23 Rb1 Nfd5 24 Be4 Qc5 (24...d3 25 c3
b6 26 cxb4 Nxf4 27 Qa3 with advantage) 25 Rxd5! Rxd5 26 Qxb4
1-0 (Halsz-Aasum). 

However, Black can improve at various points and Konikowski has
suggested that 9...Nf6!? 10 Nd2 0-0 gives Black the edge. So
Marshall's idea may indeed be critical.

The gambit appears to have re-surfaced in the 1920s. I found the
following Australian correspondence game in a history booklet
about CC in that country, "The First Fifty Years" (compiled by
H.W.M. Lunney, 1980). This was actually the first game to finish
in the competitions organised by the Australian correspondence
chess organisation.

G.W. Robertson - James McCrakett Australia, corr 1929
1 e4 e5 2 d4 exd4 3 f4 Bc5 4 Nf3 Nc6 

Black keeps coming back to this plan developing and defending the
extra pawn.

5 Bd3 Nf6 6 Nbd2 d6 

According to Frank Crowl in Australasian Chess Review, April
1930, "6...d5 seems more energetic, and if 7 e5 Ne4 whereby the
black N would occupy a commanding post." However, that isn't
clear as Halsz has won a game or two from that position.

Crowl said that this was "a favourite opening of Alexander
McDonnell; though not strictly sound it is hard to meet".
Unfortunately there are none of McDonnell games with the gambit
in the Oxford Encyclopaedia of Chess Games. Can any be traced?

7 a3 a5? 8 Qe2 0 0 9 0 0 Bd7 10 e5 Re8 11 Ng5! h6?

Immediately fatal (wrote Crowl). Some prospects of defence lay in
11...Bg4 and if 12 Bxh7+ Kf8, but by 12 Ndf3 White would retain
the better game. 

12 Nde4 Nxe4 (See Diagram)

13 Nxf7!! Kxf7 14 Qxe4 Rg8 15 Bc4+ 1-0

Another example of what can happen to Black if he plays passively
is B.Blank-K.Wothe, 14th German CC Ch1 e4 e5 2 d4 exd4 3 f4
Bc5 4 Nf3 Nc6 5 a3 a5 6 Bd3 Nf6 7 0-0 d6 8 Qe1 0-0 9 h3 Bd7 10
e5 Re8 11 Qg3 Nh5 12 Qh2 g6 13 g4 Ng7 14 Nbd2 f5 15 exf6
Qxf6 16 Ne4 Qf8 17 Qg3 Re7 18 Qh4 Ne8 19 f5 gxf5 20 Bh6!
Rg7 (If 20...Ng7 21 Nf6+ Kh8 22 Nh5.) 21 Bc4+ Kh8 22 Nfg5
Ne5 23 Be6 Bxe6 24 Nxe6 Qg8 25 Bxg7+ Nxg7 26 Nf6 1-0.

Halsz-Vitomskis, 28th European CC Championship 1984, saw the
gambit come unstuck but far from deterring Dr. Halsz from
repeating the experiment, he often played it on later occasions

3...Bc5 4 Nf3 Nf6 5 a3 a5 6 Bd3 Nf6 7 0-0 d5! 8 e5 Ne4 9 Nbd2 f5
10 exf6 Nxf6 11 Qe1+ Qe7 12 Qh4 0-0 13 Re1 Qd6 14 Ne5 Re8
15 Ndf3 Bd7 16 Ng5 [16 Bd2!? Konikowski, FS 1998/4] 16...h6
17 Bg6? [17 Ngf7 Qf8 18 Bd2 Re7 19 Nxc6 Bxc6 20 Ne5
Konikowski] 17...Rxe5! 18 fxe5 Nxe5! 19 Bd3 hxg5 20 Qxg5 Nf7
21 Qh4 Ne4-+ 22 Bf4 Qb6 23 Rf1 Re8 24 Qh5 Nf6 25 Qh4 Bd6 26
Bg5 Nxg5! 27 Qxg5 Re5 28 Qd2 Ng4 29 h3 Ne3 0-1.

Heemsoth gave this game as a good example of how to meet the
dangerous gambit. But in Halsz-Dr W. Wittmann, White varied
with 11 Re1+ Be7 (11...Ne7 may be better.) 12 h3 0-0 13 g4!?
Nxg4!? (13...h5 14 g5 Ne8 15 Nh4 with the initiative -
Konikowski) 14 hxg4 Bxg4 15 Nf1 Bh4 16 N1h2 Bxf3 17 Nxf3
Bxe1 18 Qxe1 Qf6 19 f5! Rae8 20 Qg3 Re3?! (20...h6 21 Bf4 Ne7
and Konikowski's 20...Ne7 are the critical tests here.) 21 Bxe3
dxe3 22 Qxc7 Nd4 23 Nxd4 Qxd4 24 Kg2 Qxb2 25 Rh1 h6 26
Qg3 e2 27 Qe3 Qxa3 28 Qe6+ Kh8 29 Qxe2 Qb4 30 Re1 a4 31
Qe7 Qg4+ 32 Kf2 Qf4+ 33 Ke2 1-0 (33...Qg4+ 34 Kd2 Qf4+ 35
Kd1).

The article in "Kingpin" recommends 3...d5 (avoiding Marshall's 4
Bc4!) 4 e5 c5 5 Nf3 Nc6, aiming for a blockade on the light
squares, maintaining the extra pawn. However, in the two
examples I have seen Black failed to carry the plan through
correctly.

A.J. Chavez-T. Casamayor, in the Von Massow Memorial (Cuban
CC 1991) continued 6 b3 Bf5 7 Bd3 Nge7?! (Simpler 7...Qd7
followed by 0-0-0) 8 0-0 Qd7 9 Ba3 Ng6? (9...Bxd3 10 Qxd3 Nf5
would put the onus on White to justify his gambit. Now White
develops threats.) 10 Ng5 Be7 11 e6! fxe6 12 Bxf5 exf5 13 Re1
h6? (Sacrificing the exchange by 13...0-0-0 looks the sensible thing
to do.) 14 Ne6! Kf7 15 Qh5 Nd8 16 Qxf5+ 1-0. In Halsz-
Sinovatsni (1990) Black met 6 b3 inventively by 6...Nh6 7 Bd3
Bf5 but after 8 0-0 instead of just developing his pieces, Halsz's
opponent unwisely acted the hero by 8...c4?! 9 bxc4 10 Bxc4 d3
leading to complications after 11 Bxd3 Bc5+ 12 Kh1 Bxd3 13
cxd3 Bd4 14 Nxd4 Qxd4 15 Qb3! and Black came off worst. 

Black can also follow up 3...d5 with the bishop check. After 1 e4
e5 2 d4 exd4 3 f4 d5 4 e5 Bb4+ was seen in Gimeno-Felez, 1993
Ibercaja Open , with a position that Budzinski compares to a
Falkbeer Counter-Gambit. The game went on 5 Nd2 Ne7 6 Nf3
Nf5 7 Bd3 c5 8 a3 Ne3 9 Qe2 and now Budzinski suggests the
improvement 9...Qa5.

The only other game in my database with 4...Bb4+ is somewhat
bizarre, but isn't the whole gambit bizarre?

Liban Van Damme - Gambini [C21] corr, 1972
1 e4 e5 2 d4 exd4 3 f4 d5 4 e5 Bb4+ 5 Nd2 c5 6 a3 Ba5 7 b4
cxb4 8 a4!? (White's moves 7-8 are an ingenious solution that
devalues Black's pawns.) 8...Nc6 9 Nb3 Bb6 10 Nf3 Bg4 11 Bb2
Bxf3 12 Qxf3 f6 13 Bd3 h6 14 e6 Qd6 15 Qh5+ Kd8 16 0-0 Rc8
17 a5 Bc5 18 Rfe1 b6 19 axb6 Bxb6 20 Kh1 g5?! (Asking for
trouble.) 21 fxg5 fxg5 22 Qf7 Rc7 23 Qf2 Qf4 24 Qxf4 gxf4 25
Bc1 f3 26 gxf3 (Black is still one pawn up but it is almost
worthless as White's connected passed pawns are more significant.)
26...Rg7 27 f4 Nge7 28 f5 Ng8 29 Bb2 Rg4 30 Rad1 Kc7 31 Bb5
Nge7 32 f6 Nf5 33 e7 Re4 34 Rxe4 dxe4 35 Bxd4 1-0

So how can Black put a stake through the Vampire gambit's heart?

One of many lines that ought to be good for Black in my opinion is
(1 e4 e5 2 d4 exd4 3 f4) 3...Nc6 4 Nf3 d5. This plan scores 4/5 for
Black in the games in my database and Black played some really
stupid moves in the one game he lost. (See Diagram)

Here in F.Schortgen-D.Mascarenhas, SEMI email 1997, White
played 5 Nxd4 which doesn't look right, After 5...dxe4 6 Bb5 a6 7
Bxc6+ bxc6 8 Be3 c5 9 Nb3 Qxd1+ 10 Kxd1 Nf6 11 Bxc5 Bxc5
12 Nxc5 0-0 13 Nc3 e3 White had his pawn back but where can his
uncastleable king go? Black eventually won.

So Blank, and all the other players who have faced this line,
pushed the e-pawn, which is the normal response in this gambit.

5 e5 Bc5 6 a3 Bb6

Black has also won a game with 6...Nge7 where White replied 7
b4?!.

After 6...Bb6 I have two examples involving master
correspondence players

a) Manfred Bauer- Jon Kristinsson, Scottish Magazine-10 yrs. A
1990, continued 7 b4 a6 8 Bd3 Nh6 9 0-0 Nf5 10 Qe1 h5 11 Kh1
Be6 12 Nbd2 Ne3 13 Rg1 Qd7 14 Nb3 Bg4 15 Nc5 Bxc5 16 bxc5
Nf5 17 Rb1 0-0-0 18 Rf1 h4 19 c3 h3 20 g3 dxc3 21 Qxc3 Nce7
22 Rf2 Qc6 23 Qe1 Qxc5 24 Ng5 Rhf8 25 Rfb2 b5 26 Qf1 f6 27
exf6 gxf6 28 Ne6 Nxg3+ 29 hxg3 Bxe6 30 a4 Nf5 31 Kh2 Rg8 32
axb5 Nxg3 33 Qe1 Rde8 34 bxa6 Bf5 35 Rb8+ Kd7 36 Rxe8 Rxe8
37 Bxf5+ Nxf5 38 Qd2 Rg8 39 Ba3 Qc4 40 Rg1 Re8 0-1

b) B. Blank-E.Hintikka, EU/M corr, circa 19906 a3 Bb6 7 Bd3
Nh6 8 0-0 Nf5 9 Qe2 h5 10 Kh1 Be6 11 b4 a6 12 Ng5 Nce7 13
Nd2 Qd7 14 Nb3 g6 15 Bd2 Kf8 16 Rab1 Re8 17 Nc5 Qc8 18
Ncxe6+ fxe6 19 a4 c6 20 b5 axb5 21 axb5 c5 22 c4 dxc3 23 Bxc3
c4 24 Bxf5 Nxf5 25 Rfd1 h4 26 Be1 Kg7 27 Bf2 Ng3+ 0-1.

Here is an alternative approach for Black. Anyone thinking of
playing the Halsz Gambit needs an improvement on the following
game; I wonder what Halsz himself thinks about this.

Vittorio Colo (ITA) - H. Ewald corr 1991
1 e4 e5 2 d4 exd4 3 f4 Nf6 

There has been very little experience with this Alekhine's Defence
style counter-attack against the white e-pawn; it is not mentioned
by Budzinski.

4 Bd3 

If 4 e5 d5 5 Qxd4 Qh4+ 6 g3 Nxg3 7 Nf3 Nf5 or 7 Qf2 Nf5 so
does White have anything better than 4 Qxd4 Nc6 when Black is
surely OK?

I also found a strange game Delmar-Rosen, Trenton Falls 1908, in
which White played 4 Bc4?! which was met by 4...b5 5 Bxb5 Nxe4
and Black eventually won. Presumably Black avoided 4...Nxe4
because of 5 Bxf7+ Kxf7 6 Qh5+ but there doesn't seem anything
to fear in that line.

4...Bb4+! 

This seems the right timing for the check. In other games in my
database, Black played 4...d5.

5 Bd2

If 5 Nbd2, then 5...d5!

5...Bxd2+ 6 Nxd2 d5 7 e5 Bg4 8 Ne2 Nh5 9 Qc1 Bxe2 10 Bxe2
Nxf4 11 Nf3 Nxe2 12 Kxe2 c5 13 Qf4 0-0 14 Rae1 Nc6 15 a3
Qc7 16 Kd1 f6 17 Qf5 fxe5 18 Qe6+ Qf7 19 Ng5 Qxe6 20 Nxe6
Rf2 21 Re2 Rxe2 22 Kxe2 b6 23 Nc7 Rd8 24 Rf1 Rd7 0-1.

There's a third possible refutation. Stefan Bcker, editor of the
entertaining and informative German theory and history magazine
"Kaissiber", has long advocated his Nordwalde Defence to the
King's Gambit, 1 e4 e5 2 f4 Qf6. (Nordwalde is the town where he
lives, near the city of Muenster in north-west Germany). then after
3 d4 (3 Nc3 Qxf4 is probably more critical.) 3...exd4 we have a
position that can also arise in the Halsz/Vampire Gambit, by

1 e4 e5 2 d4 exd4 3 f4 Qf6!? (See Diagram)

Stefan argues that this is the best answer to the Halsz Gambit. I
haven't seen any games where Black actually employed this move
order and I don't have any opinion about this variation, but I have
included a few games with the Nordwalde in my downloadable
file.

And this is not all. Halsz also plays an analogous gambit against
the Sicilian Defence1 e4 c5 2 d4 cxd4 3 f4!?. This "interpretation"
of the Smith-Morra Gambit (f4 instead of c3) seems utterly bizarre
at first sight, and even less likely to work than the main Halsz
Gambit since Black has not weakened his kingside. Yet this form
of the gambit can be effective too.

M.Frerichs-Gerwert Germany corr, 1995
1 e4 c5 2 d4 cxd4 3 f4!? Nc6 4 Nf3 Qb6 5 Bd3 Nf6 6 e5 Nd5 7
Na3 e6 8 Nc4 Qc7 9 0-0 Bc5? 

Hermann Heemsoth says Black should have played 9...b5! 10
Nd6+ Bxd6 11 exd6 Qxd6 12 Bxb5 Qc5! with advantage.

10 Ng5 g6?

10...b5 was the last chance.

11 Ne4 Ndb4 12 Ncd6+ Bxd6 13 Nxd6+ Kf8 14 f5 exf5 15 Rxf5!
1-0 (for if 15...gxf5 16 Qh5!).

As in the last two months, I posted on my website a collection of
games in the opening discussed in the current Kibitzer. These
games will be available for a few weeks, in PGN and in both new
and old ChessBase formats, at
http//www.chessmail.com/freegames.html. 