The Kibitzer 
by Tim Harding

How Stands the "Faj"?

In 1995 I wrote a book, "The Fighting Fajarowicz" for Chess
Digest, in which I took a thorough look at one of the byways of
opening theory, the Fajarowicz Variation of the Budapest Defence.
This arises by 1 d4 Nf6 2 c4 e5!? 3 dxe5 Ne4!? 

In the regular Budapest, White cannot hold the extra pawn except
by playing really anti-positional moves like 4 f4. However, with
3...Ne4 Black forgoes the opportunity of recovering the pawn in
many lines and makes a true gambit of it - although sometimes
White does not make a big effort to hold on to the extra material. 

I never claimed that it was an "industrial strength" defence like the
Nimzo-Indian or Slav, suitable for use in space programs and other
"mission-critical applications" but I did claim that the Faj could
give White some unexpected problems if played intelligently by a
creative player who had done his research. When a player prepares
his opening repertoire, the Fajarowicz just the kind of line that may
be "skipped" in the belief that it will never come up, and then
suddenly White finds he has to start thinking at move 4!
Half-remembered variations jumble together in the mind and an
inferior path can easily be chosen. 

The Faj book has about 370 games as study material, probably the
largest collection of games assembled at that time with the line.
Now, however, I find I have roughly 1200 games in my database,
although this figure no doubt includes a few undetected doubles
and some of the N.N. versus computer Internet blitz games of little
merit which creep in if you buy one of those CD databases that
contain a million games or so. However, there certainly seem to
have been a lot of games played with the Faj in 1995-7, both OTB
and correspondence, so maybe my book had something to do with
this.

The book in fact attracted more interest than I expected, including
a critical reassessment of the variation in two issues of the German
quarterly "Kaissiber" this year. So I decided it was time to Kibitz
the opening again and see whether the Faj had finally been refuted.

Not many of the new games have featured grandmasters on the
black side but here is an exception: 

H. Olsen - S. Conquest Reykjavik open 1996 1 d4 Nf6 2 c4 e5 3
dxe5 Ne4 4 a3 Nc6 5 Nf3 d6 6 exd6 Bxd6 7 Nbd2 Bf5 8 g3

A new move which only seems to encourage Black to attack. 

White should prefer 8 Nxe4 as in N. Sheehan - J.Gibson, Dublin
(league) 1997:

8 Bxe4 9 e3 Qf6 10 Bd3 Bxd3 11 Qxd3 0 0 0 12 Qc2 h5 (A new
move) 13 0 0 g5 14 c5 g4 15 Nd4 Nxd4 16 exd4 Bf4 17 Bxf4
Qxf4 18 Rfd1?! c6 19 d5 cxd5 20 Qc3 d4 21 Qc4 Rh6 22 g3 Qf6
23 Rd3 h4 24 Rad1 hxg3 25 fxg3?! Rxh2! 26 R1d2 Rh1+ 0 1. It
will be checkmate in two moves. The winner of that game is an
Irish friend who has played the Faj for years; many of his earlier
games are in my book. The problem with this game is that White
missed 18 c6! as Gibson pointed out afterwards; Black could be in
a bit of trouble then.

Now back to Stuart Conquest's game:

8...Bc5 9 e3 Qf6 10 Bg2 0-0-0 11 0-0 h5 12 Nd4 

White returns the pawn but the tactics don't work for him.

12...Nxd2 13 Bxd2 Nxd4 14 exd4 Bxd4 15 Qf3 c6 16 Be3 Bxb2 17
Ra2 Bd4 18 Rd2 Bxe3 19 Rxd8+ Rxd8 20 fxe3 Bg6 

Conquest heads for a winning ending: standard strategy for GMs to
defeat less experienced players in international opens without risk.

21 Bh3+ Kc7 22 Qxf6 gxf6 23 Rxf6 Rd3 24 e4 Rd6 25 Rf4 Rd1+
26 Kf2 Rd2+ 27 Kg1 Kd6 28 Bc8 b6 29 Rf6+ Kc7 30 Bf5 Rd6 31
Rxd6 Kxd6 32 Bxg6 fxg6 33 Kf2 Ke5 34 Ke3 g5 35 Kd3 b5 36 c5
a5 37 Ke3 b4 38 axb4 axb4 39 Kd3 g4 40 Kc4 Kxe4 41 Kxb4 Kd4
0-1.

Chess Cafe proprietor Hanon Russell also got my book and tried
the Faj for himself in an ICCF email master tournament, but he
wasn't so lucky. I tend to believe that in correspondence play the
move 3 Ne4 is less likely to be viable than over-the-board, since
surprise and uncertainty are important. 

O. Ejermo - H. Russell

EM/M/A012, 1997

(Notes based partly on comments by Hanon Russell)

1 d4 Nf6 2 c4 e5 3 dxe5 Ne4 4 a3 Nc6 5 Nf3 d6 6 Qc2!

White varies from the previous game; note the early move order
differences. Buecker gives this queen move an exclamation mark
and it does indeed look like one of the crunch variations for the
Faj.

6...d5 7 e3 Bg4 8 cxd5 Qxd5 9 Bc4 Qa5+ 10 b4

This has always been played but "Kaissiber" editor, Stefan Buecker
suggests another treatment for White 10 Nbd2! with the idea
10...Bxf3 (10...Bf5 11 Bxf7+!?) 11 gxf3 Nxd2 12 Bxd2 Qxe5 13
Ba6! and Black is lost e.g. 13...Nd4 (13...0-0-0 14 Qxc6) 14 Qa4+
c6 15 Bxb7.

10...Bxb4+ 11 axb4 Qxa1 12 Qxe4 Bxf3 13 gxf3

In another game from the same section, White instead played 13
Qxf3?! (not in my book), but after 13...0 0 14 Qe4 Qxe5, Black
was fine.

13...Qxe5 14 b5

In the book I say that 14 Bb5 (untested) may be best but I suggest a
way for Black to handle it, so possibly the pawn advance is
strongest after all. The material imbalance makes the whole line
hard to assess. 

14...Qxe4 15 fxe4 Ne5 16 Be2 

After the game, Mr Russell wrote me: "I have come to the
conclusion that this position may be much more dangerous for
Black than evaluated in your book. After several hours of analysis,
I rejected your suggestion to castle and re-position the Knight to
c5. The pressure that White can mount on g7 is just too strong. I
decided that queenside castling was OK (or at least the best for
Black in the given position), provided I did not open the game
prematurely, which would make White's Bishops very difficult to
handle."

16...0 0 0 17 Rg1

Here 17 0-0 was played in Gustafsson-Gibson (in my book), which
White won. 

17...g6 18 Bb2 Rhe8 19 Bd4! 

A very good move, otherwise Black will play....Nd3+, and one of
the Bishops will be exchanged. 

19...Kb8 20 f4 

Another aspect of the position makes itself felt. The White central
pawn mass is mobile and quite a problem for Black. 

20...Nd7 21 Nc3 Nb6 22 Kf2 

All of a sudden Black seems to be running out of decent moves. 

22...Re6 23 h4 Ree8 24 h5 c6 25 bxc6 bxc6 26 Rb1 Kc7 27 e5 Re7
28 Bc5 Re6 29 Ne4 gxh5 30 Bd6+ 1-0.

So if 6...d5 is unacceptable, what can Black play instead? He has
these choices:

a) A different move 6;

b) A different move 5 (instead of 5...d6);

c) Avoid the line altogether by adopting different line at move 4,
depending on whether White plays 4 Nf3 or 4 a3. I recommend the
latter course.

First let us dismiss option a). After 1 d4 Nf6 2 c4 e5 3 dxe5 Ne4 4
Nf3 Nc6 5 a3 d6 6 Qc2! the alternative is 6...Nc5 which Tseitlin
and Glazkov, in their book on the Budapest a few years ago,
described as "the only move enabling Black to organise
resistance". However, it is not very attractive. After 7 b4 Ne6
(7...Nd7 8 exd6 Bxd6 9 Qe4+) 8 Bb2 dxe5 9 e3! f6 theory goes 10
Bd3 g6 11 Nc3 Bg7 12 Rd1 Qe7 13 Nd5 Qf7 14 c5 0 0 15 Bc4ñ
(so far a 1979 postal game Neumann-Schmitz) 15...Ncd8! 16 h4 c6
17 Nc3 Qe7 18 Ne4 Nf7 followed by ...Rd8, Black would have
retained chances of gradually neutralising Black's strong pressure
(Tseitlin & Glazkov). The "Big Book of Busts" claims a refutation
by 16 Bxe5 fxe5 17 Ng5 but after Buecker's suggestion 17...Qd7,
Black seems OK since Nf6+ "wins" the black Queen at too high a
price.

There are two problems with this line. The first, again pinpointed
by Buecker, is that 10 Nc3 gives a clear advantage to White. The
Fritz5 program assesses 10...Qd7 as the best reply but with almost
a pawn's worth of positional advantage to White. Buecker's
analysis in "Kaissiber" instead runs 10...g6 (or 10...Be7 11 Rd1
Bd7 12 Bd3 g6 13 h4) 11 Rd1 Bd7 (not 11...Qe7? 12 b5 Ncd8 13
Nd5 Qf7 14 Nxf6+) 12 c5 Bg7 13 Bc4 Qe7 (13...Qc8 14 0 0 Nf8
15 Rd2 Be6 16 Bxe6 Nxe6 17 Nd5 a6 18 Qc4! with a strong
initiative) 14 Nd5 Qf8 (14...Qd8 15 0 0 0 0? 16 Nf4!) 15 0 0
0 0 0 16 Rd2 followed by doubling rooks on the d-file with a
clear advantage to White.

The second problem is that after 10 Bd3 g6 White can instead play
11 h4! as in Nieves Garcia-Rogers, Las Palmas 1995. Grandmaster
Rogers pointed out in "Kaissiber" that although the game ended in
a draw, this was probably only because of his 2260-rated
opponent's fear of their rating difference. After 11...Bg7 White
could have won the e-pawn right away by 12 h5 f5 13 hxg6 hxg6
14 Rxh8+ Bxh8 15 b5 etc. and even the quieter way White played
was very strong. In the final position when he offered a draw,
Garcia was the exchange up for a pawn with winning chances.

However, this is not the end of the story, because Niels Joergen
Jensen (author of an earlier monograph about the Faj) and Roger
Thomsen wrote from Copenhagen to say that Black has an
improvement. After 6 Qc2 Nc5 7 b4 Ne6 8 Bb2 dxe5 Black would
probably be OK if (instead of 9 e3) White played 9 Nxe5 Nxe5 10
Bxe5 a5! So their solution for Black is to play 8...Nxe5! (instead of
8...dxe5) and after 9 Nxe5 dxe5 etc. they reach the same position,
which is analysed on page 27 of the July-September issue of
"Kaissiber". 

Jensen and Thomsen now consider 11 c5, 11 b5 and 11 bxa5?! but
further on the same page, Buecker recommends 11 Nd2! axb4 12
axb4 Rxa1+ 13 Bxa1 c5 14 b5 and he also thinks that instead of 10
Bxe5 a good line for White is 10 e3, to meet 10...f6 by 11 h4
followed by Bd3 and h5. It appears that Black is not "saved" in this
line yet and practical tests of the Danish idea are needed. If
Buecker is right, then 5...d6 is busted.

What about option b? There is nothing better for Black at move 5 if
White adopts the move order 1 d4 Nf6 2 c4 e5 3 dxe5 Ne4 4 Nf3
Nc6 5 a3. It is true that there have been a lot of games with 5...a5
but there is nothing very encouraging or new for Black, especially
after 6 Qc2, while 5...Qe7 fails to 6 Qc2 Nc5 7 b4 Ne6 8 Bb2
according to Tseitlin and Glazkov. The inevitable conclusion is
that after 4 Nf3 the move 4...Nc6 is too routine and Black must do
something else, i.e. look for an option c).

The alternatives against 4 Nf3 are 4...Bb4+ and 4...b6. The
problem with the bishop check is that it rules out a lot of the more
fascinating and tricky lines that Black would like to reach and
which make the Faj attractive in the first place, e.g. 4 Nf3 Nc6 5
Nbd2 Nc5 6 g3 d6!?. Generally, after 4 Nf3 Bb4+, two pairs of
minor pieces get exchanged on d2, Black regains his pawn at e5
and White acquires a nagging positional edge.

The move I wanted to recommend was 4...b6 but I could not find a
good line for Black after 5 Qd5 Bb7 6 Qxb7 Nc6 7 Nd4!! Bb4+ 8
Nc3 0-0 (Sapierzynski-Berg, Poland-Sweden corr 1992) 9 a3!

Here, Buecker comes to Black's rescue for a change and proposes
5...Bb4+ (instead of 5...Bb7). There is not room to repeat all his
analysis here; anyway, this is a fascinating line for home analysis:
will White's queen be trapped in the corner after eating a black
rook? The main line goes 6 Nc3 Nxc3 7 Qxa8 Ne4+ 8 Bd2 Nxd2 9
Nxd2 Nc6 with a critical position. Black threatens 0-0, Qe7 and
Ba6, and if 10 g3 0-0 11 Bg2 Qe7 12 Bxc6 Bxd2+ 13 Kxd2 Qb4+
Black has a favourable position, so 10 c5!? seems better with a
possible continuation mentioned in "Kaissiber" being 10...0-0 11
e3 (11 Rc1!?) 11 ..Qe7 12 Rd1 Ba6 13 Qxf8+ Kxf8 14 Bxa6. Is
White better here? Can Black play differently?

If White wants to reach the line we discussed above, but avoid
...Bb4+ lines, he often plays 4 a3 first. I don't trust 4...Qh4 which
is superficially attractive because 5 g3! is hard to meet. Against 4
a3, too, Black has tried 4...b6!?, which is cited approvingly by
Buecker.

However, if Black does want to meet 4 a3 by 4...d6 then after 5
Nf3 he has a credible alternative to 5...Nc6, namely 5...Bf5!? as
suggested in the second edition of Borik's book on the Budapest;
not then 6 Qc2? Ng3 while if 6 exd6 Bxd6 Black has a lead in
development for the pawn, he said. There have been several games
with 5...Bf5 since I prepared my book. Here is one of them.

E. Gleizerov - Dausch, Cappelle la Grande open 1995

1 d4 Nf6 2 c4 e5 3 dxe5 Ne4 4 a3 d6 

Black can also play 4...Nc6 first, to meet 5 Qc2 by 5...d5, but then
White simply plays 5 Nf3 and gets our main line above.

5 Nf3 Bf5 6 Nbd2 

Also critical are 6 g3 and 6 Bf4 (Raetsky-Dausch, Bad Zurzach
1995).

6...dxe5 7 Nxe4 

Although 7 Nxe5 may be just about playable it is understandable
that White avoids continuations like 7...Nxf2 8 Qa4+ (8 Kxf2
Qd4+) 8...c6 9 Nxf7 Qe7 (or 9...Kxf7 10 Kxf2 Bc5+ 11 Ke1) 10
Kxf2 Qxf7 11 Nf3 where Black is likely to have the safer king in
the long run.

7...Qxd1+ 8 Kxd1 Bxe4 9 Nxe5 Bd6 10 Nd3 Nc6 

Despite the simplification, Black has compensation for the pawn.

11 Bd2 

11 e3 0 0 0 12 f3 Bg6 13 Kc2 Rhe8 14 b4 Bxb4 15 axb4 Nxb4+
16 Kc3 Nxd3 17 Bd2 a6 ought to have been good for Black in
Ward-Dausch, Copenhagen open 1995, but Black allowed a
combination and White eventually won.

11...0 0 0 12 Bc3 Be5 13 Kd2 Rhe8 14 Bxe5 Rxe5 15 Kc3 Bxd3
16 exd3 Nd4 17 b4 b6 18 g3 c5 19 Bh3+ f5 20 Rhe1 Rde8 21 Rxe5
Rxe5 22 Bf1 g5 23 h3 f4 24 g4 Kc7 25 Rb1 Re7 26 Kd2 Nf3+ 27
Kc3 Nd4 28 Kd2 Nf3+ drawn by repetition.

Another line discussed in "Kaissiber" is 1 d4 Nf6 2 c4 e5 3 dxe5
Ne4 4 Qc2 following Hertneck-Trapl, Bundesliga 1991: 4...Bb4+ 5
Nd2 d5 6 Nf3 and now 6...Bf5?! This is only a sideline, and not
critical to the ultimate value of the "Faj" because Black has the
playable alternative 6...Nc6, but the resulting position is
fascinating.

After 6...Bf5 7 Qb3 Nc6!? (instead of Trapl's choice 7...Bxd2+) 8
cxd5 Nc5 9 Qc4 b5 10 Qf4 Qxd5 11 Qxf5 Ne4 Tseitlin and
Glazkov wrote that Black's threats are difficult to meet but I said
that depends on what I called a "fantasy variation which may be
flawed" - and Buecker seems to have found the flaw.

Tseitlin and Glazkov gave the line 12 a3 Rd8 13 axb4 Nxb4 14
Rb1? Qc4? but I wrote in my book that 14...Nc2+! forces mate in
six. A disbelieving German reader wrote to say he didn't see the
mate: evidently he didn't have a computer!

I pointed out that 14 Nd4!! Qxd4 15 e3 Nc2+ 16 Kd1! Qb4! leads
to a draw (and this has since occurred in an IECC email game
Beck-Soberano, 1996). However, Buecker says White can do
better than this, in two ways:

a) 13 Qf4 (instead of 13 axb4) 13...g5 14 Qe3 Bxd2+ 15 Bxd2 g4
16 Qd3!? gxf3 17 Qxd5 Rxd5 18 Bf4.

b) 12 e3 (instead of 12 a3) 12...g6 13 Qf4 0-0-0 14 Bxb5! Nxd2 15
Ba6 and 16 Bxd2.

Is it possible to revive this line for Black?

Incidentally, Mr Buecker was seeking further biographical
information on Fajarowicz and if possible a photograph of him.
Can any reader help?

Finally, "Kaissiber" 4/1997 mentions that an American player may
have been the true originator of the variation! Hugh Myers and
Edward Winter both found in "British Chess Magazine" January
1919 (unavailable to me), on pages 1-2, a mention of the move by
the noted theoretician Stasch Mlotkowski (an American of Polish
extraction who often wrote in "BCM"). 

Mlotkowski said that S.R. Barrett of Philadelphia had played
3...Ne4 against him 15 years previously!