These
days, opening theory is everywhere. You can find
it in books and magazines, on videos, cassettes,
CDs, even on the Internet. This explosion of
information is exciting and provides many more
opportunities than existed ten or twenty years
ago. Still, we must keep in mind the limitations
of "theory." Many a player has learned
the hard way that theory often doesn't survive
a head-on collision with practice. I've played the sharp and dangerous King's Indian
Defense most of my life. As a consequence,
I've had my share of opening debacles. Perhaps
my worst ever occurred in the 1991 Iowa State
Championship, where my personal nemesis, Bob
Reynolds, showed that his preparation was better
than mine. I am hopeful that I've learned a thing or two
from my past mistakes, and the following game
is a case in point. My opponent walked into the
same line as the earlier Reynolds game. I had
a surprise waiting for him. Joel Mandeville (2150) - Randy Bauer (2300)
Rated Match, 1997
King's Indian Defense, Four Pawns Attack, Gunderam Gambit 1.d4 g6 2.c4 Bg7 3.Nc3 d6 4.e4 Nf6 Black's roundabout method for getting to the
King's Indian gives him certain advantages in
meeting irregular lines. For one thing, the Trompowski
bishop (1.d4 Nf6 2.Bg5) looks kind of funny out
there "pinning" the ghost of a knight.
It also allows black to accelerate his ...e5
break in lines where white plays an early Bf4.
The downside, if you care to call it that, is
that white may choose to switch to a kingside
structure with 2.e4. As I have a better record
with the Pirc than with the King's Indian, I
don't view that as much of a downside. 5.f4 
PLAYING FOR WORLD DOMINATION
The Four Pawns Attack is, in some respects,
white's most principled response to the King's
Indian. White says, in effect, "If you don't
want to stake a claim to the center, I will." While
the pawns look imposing, standing four abreast
and controlling or occupying literally the entire
center, it does come at a price. White's development
lags, and black will have his opportunities to
take pot shots at that center. The f-pawn may
also get in the way of the white pieces - the
queen bishop in particular may be left to "stare at
the backside" of the f4 pawn. For this reason,
white often seeks to liberate that piece with
either f4-f5 or e4-e5, intending to meet ...dxe5
with fxe5. 5... 0-0 6.Nf3 c5 These days, 6...Na6!? is all the rage, intending
to play ...e5 (even as a pawn sacrifice if necessary)
and bringing the queen knight to the active c5
square. Kasparov has played this way, and Joe
Gallagher recommends it in his BEATING THE ANTI-KING'S
INDIANS book. The move in the game is the long-time
favored approach. 7.d5 A playable option is 7.dxc5 Qa5 8.Bd3 Qxc5 9.Qe2
(white prepares to play Be3 so that he may castle)
9.Nc6 10.Be3 Qa5 with play akin to the Maroczy
Bind in the Sicilian Defense. 7...e6 Black strikes back in the center while the white
king is still in the vicinity. This tends to
limit white's options, as he can't get too aggressive. 8.Be2 exd5 9.e5!??! 
WHITE GOES FOR THE THROAT
All the punctuation is probably excessive, but
I think it conveys the fact that there is a genuine
disagreement among a variety of sources about
the viability of this move. As I mentioned in
the note to black's seventh move, it seems dangerous
to be picking a fight in the center while the
king is still uncastled. White hopes that the
time it takes black to win a pawn will be offset
by the open lines and chance for him to catch
up in development. This line, called the
Gunderam Gambit, has been recommended by GM Andy
Soltis in his 1.d4 repertoire book (both original
and revised version). It was also recently given
the stamp of approval by IM Rafael Klovsky and
NM Vadim Kaminsky in a 1997 article in Inside
Chess titled "Death to the King's Indian." On the other hand, IM Andrew Martin panned the
line in his WINNING WITH THE KING'S INDIAN book.
More recently, IM Anatoli Vaisser, arguably the
foremost expert on the Four Pawn Attack King's
Indian, said that the variation is currently
in crisis. 9...Ne4! Vaisser and Martin both give this move an exclam,
while Klovsky and Kaminsky say that the move
looks risky, since black's knight has no support
in the center. Black's idea is that 9...dxe5
would only open lines for white's pieces. The
knight also gives black some options with ...Qa5+,
and also might allow black to profitably play
...Nxc3. 10.cxd5 10.Nxd5 is the other option, but it doesn't
seem right to cede the black knight the nice
c6 square. One of white's advantages in King's
Indian lines is the spatial edge the central
pawns give. If white doesn't put his pawn on
d5, it's hard for him to claim much advantage
in central control. The key line after 10.Nxd5 is, naturally, 10...Nc6. Then
11.Qc2 f5 12.0-0 dxe5 13.fxe5 Nxe5 14.Bf4. Now
Vaisser suggests that both 14...Re8!? 15.Rad1
b6 and 14...Nc6 lead to an advantage for black.
Kovsky and Kaminsky quote the same line with
14...Nc6 as Vaisser but overlook a point where
black could have improved his play. While the
line certainly offers white some play, I don't
really think it's sufficient. 10...Nxc3 11.bxc3 Nd7 
WHITE TO MOVE
This is a key move for black, as he continues
to increase his development advantage and puts
more pressure on the white center. 12.e6 The stem game for this variation with 9...Ne4!
was Li Zunian-Gheorghiu, Dubai Olympiad 1986,
where 12.0-0 dxe5 13.fxe5 Nxe5 14.Be3 Nxf3+ 15.Bxf3
Qd6 left black a clear pawn up, and he won a
nicely played endgame. [Editor's Note: Actually, this isn't correct.
R Calvo Minguez - Diez del Corral, Malaga 1981
predated the Gheorghiu game. See the games' section
at the end of the article for the complete scores
of both.] 12...Bxc3+ 13.Bd2 Bxd2+ Black must be careful about his dark squares,
and 13...Bxa1? 14.exd7 gives white a clear advantage
because of the vulnerability of the black king. 14.Qxd2 fxe6 15.dxe6 Nf6 16.f5 gxf5 17.Qg5+
Kh8 18.Qxf5 Qe7 
DIFFERENT BOOKS, DIFFERENT ASSESSMENTS
After a relatively forced series of captures,
threats, and counter-threats, we've reached another
point of disagreement in theory. Soltis analyzes
up to 16.f5 and claims that white has play. Martin
analyzes up to 18...Qe7 and says it's "goodbye
to the e-pawn." Soltis may have reached
the same conclusion, because he abandons this
line in his revised second edition. Vaisser chooses
a different method, preferring 15...Nb6 to 15...Nf6.
He claims a black advantage after 14.0-0 Bxe6
15.Ng5 Bd5!. Obviously, white needs to try to maintain the
e6 pawn; if it falls, it's hard to see white's
compensation. Thus, white's next move is logical. 19.Ng5 Herein lies the tale. In 1991, when preparing
for the Iowa State Championship, I relied extensively
on IM Martin's book. As a result, I wasn't surprised
to see the position after 18...Qe7 on the board.
After all, that was Martin's line, and he pronounced
that the e-pawn was falling. When my opponent,
Bob Reynolds, played 19.Ng5, I naturally assumed
that the knight move couldn't be good. After
all, if it was, wouldn't Martin have included
it in his analysis? Thus, rather than taking a good, hard look at
the position, I simply replied 19...h6 after
just 5 minute's thought. After the only consistent
move, 20.h4! I then immediately blundered with
20...hxg5? when 21.hxg5+ Kg8 22.gxf6 Rxf6 23.Qh5!
put black in a losing position. After 23...Qg7
24.e7! it was all over (although I struggled
on to move 30). After the game, I spent more than my share of
time on the position after 19.Ng5 (which is,
of course, what I should have done during the
game!). While black may be able to get by with
a move like 19...Kg8, I found a more radical
solution. Imagine my surprise when the same position appeared
six years later... 19...Bxe6! 
WHITE DIDN'T EXPECT THAT!
This move obviously came as a shock to my opponent,
who went into a half hour think. Black is sacrificing
a piece, it is true, but all the rest of his
pieces become active while white's king is still
sitting in the center of the board. While it
may be some comfort to black that, even if things
don't work out, he has two pawns for the piece,
in reality, black believes that he will either
win his piece back or mate white - the pawns
don't matter all that much at the moment. In
fact, I'd worked all this out to a big advantage
for black in every line years ago. My only concern
now was to remember my earlier analysis! It's interesting to note that the couple of
computer software programs I've tried this position
on don't recommend 19...Be6 (although they find
better moves than the lemon I played against
Reynolds). However, once 19...Be6 is played,
they quickly conclude that it is a very strong
move. I suppose this is understandable, since
several lines require quite a few moves for black's
compensation to become clear. Needless to say,
I found this move without the aid of the computer,
but I have since checked my lines to make sure
they work on Fritz. 20.Nxe6? White obviously has to accept the sacrifice,
or he has absolutely nothing to show for his
two-pawn deficit. There are two ways to capture,
with the knight and with the queen, and they
each have their unique problems. According to my analysis, the move in the game
gives white fewer chances to complicate the issue.
After 20.Qxe6! Qg7, white can go down in flames
very quickly (because his king and queen are
stuck in the center and his knight attacked and
rook subject to attack). However, he has one
nice resource that complicates black's task considerably.
Typical lines run 20.Qxe6 Qg7 
WHITE'S IN TROUBLE
and now: A) 21.h4 protects the knight on its advanced
square. Unfortunately, it does nothing to help
with development. Now the logical 21...Rae8?!
gives black tons of play (I've worked these lines
out to a win) if the queen retreats. However,
with 22.0-0! Rxe6 23.Nxe6 white eliminates most
of the black forces. Granted, black is still
better in the resulting ending, but he can do
better. I think black should play 21...Nd5, which prevents
kingside castling while attacking the rook. The
knight also aims for the f4-square, with a double
attack on the queen and bishop. 21...Nh5 would
be similar, but in some lines it helps to keep
the bishop on e2. 
BLACK HAS TOO MUCH FOR THE PIECE
Thus, after 21...Nd5 white pretty much has to
allow the capture of the rook. After 22.Qxd5
Qxa1+ white has two choices: A1) 23.Qd1 Qxa2 24.Qd3 (this starts white's
thematic counterplay, which will be seen in the
following variation as well) 24...Qa1+ 25.Kd2
Qb2+ (It makes sense to confine the king to the
first rank to keep the white rook passive before
dealing with the mate threat) 26.Kd1 Qg7
27.Kc2 Rae8 28.Bf3 Rf4. Here I think that black,
with the white king still stuck in the center
and the rook out of play, has a big advantage
with his queenside pawns ready to roll and his
rooks engaged. A2) 23.Kd2 Qb2+ (black can't afford to take
the second rook, because 23...Qxh1 24.Qe4 is
an improved version of the A1 counterplay) 24.Kd1
Rae8 25.Qd3 Qa1+ 26.Kd2 Qxa2+ (the same method
as in A1 for gathering up another pawn) 27.Kc1
Qa1+ 28.Kd2 Qb2+ (the king is again sent back
to the first rank) 29.Kd1 Re7. Once again, black's
pieces are more active and the queenside pawns
the biggest long-term factor. Granted, white doesn't have to play 21.h4, which
doesn't help his development at all. Unfortunately,
the logical knight retreat has problems, and
attempts to get the king immediately out of the
center allow black to immediately regain his
piece with an extra couple of pawns and decent
development to boot. B) 21.Nf3 Rae8 22.Qc4 (22.Qh3 Nd5 23.0-0-0 Rxe2
24.Rxd5 Qa1 mate) 22.Re4 (22...Qxg2 also looks
good) 23.Qc2 Rfe8 24.Rf1 Nd5 25.Rc1 Ne3 26.Qd3
(26.Qd2 is met the same way, while 26.Qc3?? Nxg2+)
26.Nxf1 27.Kxf1 Rxe2 with a winning material
advantage. C) 21.0-0 (this immediately gives back the piece,
hoping to win back at least one pawn, but there
is a tactical problem) 21...Qxg5 22.Qxd6 Qe3+!
23. Kh1 (the problem is 23.Rf2 Qxf2+ 24.Kxf2
Ne4+ winning) 23.Rad8! (now white's back rank
is a real problem) 24.Qc7 Qxe2 25.Rxf6
Rd1+ with mate to follow. D) 21.0-0-0 (this also gives back the piece,
and white's king is still exposed) 21...Qxg5+
22.Kb1 (22.Rd2? Rae8 finishes things) 22.Rae8
23.Qc4 d5 24.Qc2 Qxg2 25.Rde1 Qe4 with a winning
advantage. 20...Rae8 21.Bc4 
BLACK TO MOVE
This seems like the best try. After 21.Nf4 Nd5!
22.Qxd5 Rxf4 there is no way to protect the bishop,
and 23.0-0-0 Qxe2 24.Rhe1 Rc4+ 25.Kb1 Rb4+ mates. 21.0-0-0 doesn't solve white's problems either.
After 21...Qxe6 22.Qxe6 Rxe6 23.Bc4 Re4 black
will emerge two sound pawns up. 21...Nd7! The impatient 21...d5? 22.0-0! Qxe6 23.Qxe6
Rxe6 24.Bxd5! at least gives white some chances
because of the superiority of the bishop versus
the knight (although black should still have
a clear edge). 22.Qh3 Of course, 22.Qg4 Ne5 breaks down the defense
of the advanced knight. 22...Rf4 The black pieces enjoy a "target rich environment".
Now the bishop, which is the key to holding the
knight, is attacked and the rook can also play
....Re4+ if necessary. On 23.Rc1 Nf8 piles more
pressure on the white knight. Note also that
23.Bb3 c4! rounds up the piece for black as well.
White decides that with his king stuck in the
center and no real way to save the knight, he
might as well attempt to get the queens off to
improve his long-term prospects. 23.Qc3+ Qf6 24.Qxf6+ On 24.Rc1 Nf8 again does the trick. 24...Nxf6 Black can be remarkably patient, as the knight
isn't going anywhere. 25.Rc1 b5! 
EFFECTIVELY ENDING THE GAME
This is better than 25...d5, which loosens the
c5-pawn. Now the piece finally falls, and white
tries to find the way that will give him the
best chances of survival. 26.Kd1 While white eliminates one of the black pawns
with 26.Bxb5, black's rooks get hyperactive after
26...Rxe6+ 27.Kd1 Rf2 26... Rxc4 27.Rxc4 bxc4 28.Re1 Ng4 Black doesn't play this just to give a check;
rather he is going to put his knight on the strong
d3-square. 29.h3 Nf2+ 30.Kc2 Nd3 31.Re2 d5 Now, with the d3-knight controlling c5, the
pawns can start rolling. 32.g4 Kg8 33.h4 h6! 
ACCURATE TO THE VERY END
A final little finesse. Now black threatens
...Kf7, which forces a rook trade. The knight
ending, given the black mass of passed pawns,
is easily won. 34.Kd2 Kf7 35.Nc7 Rxe2+ 36.Kxe2 d4 37.Nb5 Nc1+
38.Kd1 Nxa2 39.Nxa7 c3 40.Nb5 d3, 0-1. The pawns cannot be stopped. When I asked my opponent about the line after
the game, he told me he was simply following
Soltis' analysis! Of course, I couldn't pass
up the opportunity to tell him, "trust,
but verify." Silman: Here is a sampling of theoretically
important games in this line (E77). The results
lead to a clear verdict - White should avoid
9.e5 like the plague! WHITE TRIES 10.Nxd5 P San Segundo Carrillo - J
Howell
Gausdal, 1986
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.f4 0-0 6.Nf3 c5 7.d5 e6 8.Be2 exd5
9.e5 Ne4 10.Nxd5 Nc6 11.Bd3 f5 12.exf6 Nxf6 13.0-0 Nxd5 14.cxd5 Nd4 15.Ng5
Qf6 16.Nxh7 Kxh7 17.Qh5+ Bh6 18.g4 Bf5 19.Bxf5 Nxf5 20.Qh3 Qd4+ 21.Rf2 Rae8
22.gxf5 Re1+ 23.Kg2 gxf5 24.Qf3 Rg8+ 25.Kh3 Qf6 26.Rg2 Rg1 27.Rxg8 Kxg8 28.Qh5
Qg7 29.Qxf5 Qg2+ 30.Kh4 Qxh2+ 31.Qh3 Qxh3+ 32.Kxh3 Bxf4 33.Bxf4 Rxa1 34.Bxd6
c4 35.a3 Rd1 36.Bb4 Rxd5 37.Kg4 Re5 38.Kf4 Re2 39.Bc3 Kf7 40.Kf3 Re6 41.Kf4
Ke7 42.Be5 Kd7 43.Ke4 Kc6 44.Kd4 Rxe5 45.Kxe5 Kc5 46.a4 a5 47.Ke4 c3 48.bxc3
Kc4, 0-1. N Povah - J Howell
Lloyds Bank, 1994
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.f4 c5 6.d5 0-0 7.Nf3 e6 8.Be2 exd5
9.e5 Ne4 10.Nxd5 Nc6 11.Qc2 f5 12.exd6 Qxd6 13.0-0 Be6 14.Qb3 Bxd5 15.cxd5
Nd4 16.Nxd4 Bxd4+ 17.Kh1 Nf2+ 18.Kg1 Rae8 19.Bc4 Kh8 20.Qxb7 Qf6 21.g3 Ne4+
22.Kh1 Nd6 23.Qa6 Qe7 24.Bd3 c4 25.Qa4 Bg7 26.Bb1 Qe2 27.Qd1 Qxd1 28.Rxd1 Re2
29.a4 Rfe8 30.Bd2 Bxb2 31.Ra2 c3 32.Rxb2 cxd2 33.Bc2 Nc4 34.Rbb1 Re1+ 35.Kg2
Ne3+ 36.Kh3 Nxc2 37.d6 Rd8, 0-1. B Friesen - I Glek
Vlissingen , 1997
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.f4 c5 6.d5 0-0 7.Nf3 e6 8.Be2 exd5
9.e5 Ne4 10.Nxd5 Nc6 11.0-0 dxe5 12.b4 Nxb4 13.fxe5 Nxd5 14.cxd5 Nc3 15.Qc2
Nxe2+ 16.Qxe2 Qxd5 17.Rd1 Qc6 18.Rd6 Qa4 19.Bg5 Bg4 20.Rc1 b6 21.Bf6 Bxf6 22.exf6
Qf4 23.Rcd1 Rae8 24.Qf2 Re6 25.Rd7 Rxf6 26.Rxa7 Bxf3 27.Qxf3 Qxf3 28.gxf3 Rxf3
29.Rb7 Rf6 30.Rb1 Ra8 31.Rb2 Ra4, 0-1. THE CRITICAL 10.cxd5 R Calvo Minguez
- Diez del Corral
Malaga, 1981
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.f4 c5 6.d5 0-0 7.Be2 e6 8.Nf3 exd5
9.e5 Ne4 10.cxd5 Nxc3 11.bxc3 Nd7 12.0-0 dxe5 13.Ng5 h6 14.Ne4 f5 15.Nd6 e4
16.Rb1 Nb6 17.Nxc8 Rxc8 18.c4 Bd4+ 19.Kh1 Rc7 20.Rb3 Nc8 21.Rh3 Rh7 22.Qa4
a6 23.Bd2 Nd6 24.Rb1 Qe7 25.Rb6 e3 26.Be1 g5 27.fxg5 hxg5 28.Rxh7 Kxh7 29.Qb3
Rf6 30.Kg1 Kg7 31.a4 Rh6 32.a5 Be5 33.Bc3 Bxc3 34.Qxc3+ Kf7 35.Rb3 Rh4 36.g3
Rd4 37.Qxe3 Re4 38.Bh5+ Kg8 39.Qd3 g4 40.h3 Re1+ 41.Kf2 gxh3 42.Bf3 Ra1 43.Qc3
Ra2+ 44.Rb2 Rxb2+ 45.Qxb2 Nxc4 46.Qc1 Ne5 47.d6 Qf6 48.Qd1 Nxf3 49.Kxf3 Qf7
50.Kf4 Qd7 51.Qb3+ Kg7 52.Qc3+ Kg6 53.g4 Qa4+ 54.Ke3 Qd4+ 55.Qxd4 cxd4+ 56.Kf2
h2 57.Kg2 Kf7 58.g5, 1-0. Of course, black was just winning this game.
For example, 42.Ne4+ would have won on the spot. Li Zunian - Gheorghiu
Dubai Olympiad, 1986
1.d4 Nf6 2.c4 c5 3.d5 d6 4.Nc3 g6 5.e4 Bg7 6.f4 0-0 7.Nf3 e6 8.Be2 exd5
9.e5 Ne4 10.cxd5 Nxc3 11.bxc3 Nd7 12.0-0 dxe5 13.fxe5 Nxe5 14.Be3 Nxf3+ 15.Bxf3
Qd6 16.Qd2 Be5 17.h3 b6 18.Bh6 Re8 19.Rae1 Ba6 20.Be2 Bxe2 21.Rxe2 c4 22.Ref2
Re7 23.Kh1 Rae8 24.Bg5 Rd7 25.Re2 Qxd5 26.Qe3 Re6 27.Qc1 f6 28.Bh6 Qc5 29.Qe1
Rde7 30.Rd2 Rd6 31.Qe4 Qc6 32.Qe2 Rxd2 33.Qxd2 Qd7 34.Qf2 Qd3 35.Re1 Kf7 36.Re3
Qd1+ 37.Re1 Bg3 38.Qxf6+ Kxf6 39.Rxd1 Re1+, 0-1. D Rozakis - K Nikolaidis
Aegina, 1996
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.f4 0-0 6.Nf3 c5 7.d5 e6 8.Be2 exd5
9.e5 Ne4 10.cxd5 Nxc3 11.bxc3 Nd7 12.e6 fxe6 13.dxe6 Nf6 (13.Nb6!? also
has a good reputation) 14.f5 gxf5 15.Ng5 Ne4 16.Nxe4 fxe4 17.Qd5 Qe7 18.Bc4
Bxc3+ 19.Ke2 Bxa1 20.Bg5 Bf6 21.Rf1 Bxg5 22.Qxg5+ Qxg5 23.e7+ Be6 24.Bxe6+
Kg7, 0-1. |
