Opening Theory: Piece Sacrifice in the Velimirovic

by  Bruce Leverett  and Ken Davenport


We are going to discuss a line of the Velimirovic Attack of the
Sicilian Defense.  One sign of an opening line that is challenging is 
that one of the players of the game quoted here,  IM Maurice Ashl
ey, plays this line as both white and black.

The Velimirovic Attack of the Sicilian Defense was described in a
book by Bruce Leverett published by Chess Enterprises in 1983. Since 
then, many
new ideas and lines have been tried. In the July 1994
 issue of  Chess Life IM Maurice Ashley played white in this opening
against Deon Solomons. After  1 e4  c5  2 Nf3 d6   3 d4  cxd4   4 
Nxd4 Nf6 5 Nc3  Nc6  6 Bc4 e6  7 Be3 Be7  8 Qe2 O-O  9 O-O-O  Qc
7  10 Bb3 a6  11 g4 Nd7  the following position arose.


r1b2rk1
1pqnbppp
p1npp3
8
3np1p1
1BN1B3
PPP1QP1P
3KR3R

The game continued 12  g5  Nc5  13  Kb1 (an uninspired move. Probably
13  h4  or 13  Rhg1 is better)  ... Bd7  14  Rhg1 b5  15  Rg3.  
Solomons gives Rg3 a '?', noting that it is difficult to see how
white should proceed after b5.  White needs something better a little
earlier.

That something better may be the potential piece sacrifice 12  Nf5.
Solomons calls this line highly speculative, giving 12  Nf5  Nc5  13  
Nxg7  Nxb3+  14  axb3  Kxg7  15   g5  Bd7  16  Qh5  Kg8  17
 f4, and noting that white may not have enough compensation for the
piece. Nick DeFirmian in MCO-13 (1990) doesn't mention 12 Nf5.  
Kasparov and Keene in BCO-2 (1989) give 12 Nf5 exf5 as a main varia
tion, with 12 ... Nc5 in a footnote.  The footnote continues 13 Nxe7+
Qxe7 14 Rd2 b5 15 f3 +/-.  This evaluation seems rather optimistic, 
since Black
can probably hold.  An improvement is needed, so
how about  13 Nxg7 ?

Continuing with 13  Nxg7  Nxb3+  14  axb3 Kxg7 15  g5 Bd7 16  Qh5 Kg8
17  f4 Rfc8 (threatens indirect mate on c2 and therefore pins white's 
c3 knight).


r1r3k1
1pqbbp1p
p1npp3
6PQ
4PP2
1PN1B3
1PP4P
2KR3R
 After  17 ... Rfc8
Note that 17 ... f5 allows 18 g6, and 17 ... Be8 allows 18 f5, in
either case clearly winning for White.  Then  18 Rhg1 Bf8.  So far, 
logical play by both sides. The play so far, and some of the foll
owing variations, were worked out by Ken Davenport with assistance
from some other inmates of SCIP.

One variation may now go 19 Rg3 Bg7 20 Rh3 Ne7 21 Qxh7+ Kf8 22 Bd4
Bxd4 (otherwise 22. ...  Bxc3  23 bc  Qxc3 etc)  23 Rxd4 e5  24 Qh6+ 
Ke8 (or 24
... Kg8 25 Qh8) 25 Qh8+ Ng8 26 Qxg8+ Ke7 27 Nd5+ Ke6
 28 f5 #.

To avoid this debacle, Black must vary.  22 ... e5 suggests itself,
but White can reply 23 f5!  Now 23 ... f6 24 gxf6 Bxf6 25 Be3 looks 
pretty horrible, so instead Black accepts the sac with 23 ... e
xd4, and White sacs some more with 24 f6!  This is a crucial
variation but there may be better for black earlier. Consider 19 ... 
Ne7 20 Rh3
h6 -- a familiar idea in the Velimirovic, but not good eno
ugh here.  After 21 gxh6 White threatens 22 Rg1+ Kh8 (or 22 ... Ng6
23 Rxg6+ fxg6 24 Qxg6+ Kh8 25 Rg3) 23 Qxf7.  And 21 ... Ng6 and 21 
... Be8 both
allow 22 f5.

Black's best defense is 19 ... Bg7 20 Rh3 Nb4.  Black appears to
escape with at least a draw, but perhaps no more.  A crucial 
variation is 21 Qxh7+ Kf8 22 Bd4 e5 23 f5 exd4 24 f6 Bxh3 25 Qxg7+ Ke8.
Now on 26 Qg8+ Kd7 27 Qxf7+ Kc6 28 Qc4+ Kb6 29 Qxb4+ Ka7, or 29 Qxd4+
Ka5 (but not 29 ... Qc5 30 Na4+), Black wins.  But on 26 Qh8+ Kd7 27 
Qxh3+, it becomes very tricky.  It's a draw by perpetual che
ck after 27 ... Kd8.  But if Black tries for more with 27 ... Kc6,
he's taking chances.  28 Rxd4 threatens both 29 Rxb4 and 28 Rc4+, and 
28 ... Qa5 doesn't help since 29 Rc4+ still wins a rook. 28 ..
. Kb6 29 Rxb4+ Ka7 or 28 ... Rh8 29 Rc4+ Kb6 30 Rxb4+ Ka7 leaves
Black up the exchange, but White has compensation, at least at first 
glance.

Having seen these lines, we can go back and look at 20 ... Ne7.  We
don't think it's as good as 20 ... Nb4, but the reader can work out 
the details
and maybe come up with some new ideas.

19 f5 is not an improvement.  After 19 ... Ne5 Black is threatening
20 ... exf5 (because the knight on c3 is now pinned).  So 20 f6 Ng6 
21 Rg3 e5 and now 22 Rh3 drops the exchange.  After 22 Rd2 Be6
23 Nd5 Bxd5 24 Rh3 h6 25 gxh6 Kh7 26 exd5 a5, Black is secure and his
counterplay is starting.
Many of the above lines were suggested by Bruce Leverett. But this
was not the last word. Davenport and Leverett played a game at SCIP 
on May 7, starting from the diagram, at a time limit of game/25.
 This game continued  18 Rhg1 Bf8  19 f5!? Ne5  20 Kb1! Bc6!  21 Rd4
b5 (the start of a plan that is much too slow) 22  Rg3 a5  23 Rh3 h6  
24 Rd2 b4
 25 Rg2! bxc3  26 gxh6+ Kh8  27 h7 (inaccurate--Bl
ack could fight now with 27 ... Ng6-- better was 27 Qg5)  27 ... Bg7
28 Rxg7! Kxg7  29 Qh6+ Kh8  30 Qf6 mate.  Davenport's 19 f5 and 20 
Kb1 add some
life to the discussion. In the post-mortem, it see
med that Black could refute it:  21 ... exf5 (instead of 21 ... b5)
22 exf5 Nf3, winning the exchange. But in this position Davenport 
suggests  23 g6!

r1r2bk1
1pq2p1p
p1bp2P1
5P1Q
3R4
1PN1Bn2
1PP4P
1K4R1



Now Davenport suggests the following variation (among others) as a
sample of the strength of White's attack: 23 ... fxg6 24 fxg6 Nxg1  
25 Rf4 (threatening 26 Rf7)  25 ... hxg6 26 Qxg6+  Bg7 (or Qg7)
27 Qe6+ Kh8 28 Rh4+, etc.

It is possible to bypass this variation by declining to take the rook
(22 ... Nf3 and 24 ... Nxg1).  In this case Black would have to 
strengthen his
play earlier; possibilities include 21 ... Qd7 (or 21 ... exf5 22 
exf5 Qd7)
or even 21 ... Bg7.

Further analysis by SCIP inmates and visiting master Franc,ois 
Margot, however,
seems to rehabilitate the acceptance of the rook.  After 23 ... fxg6
24 fxg6 Nxg1 25 Rf4, Margot suggested 25 ... Bf3, and now if 26 Qf5 
Bg7
27 Qe6+ Kh8 28 Rf7, Black repulses the attack by 28 ... Re8 (Robert 
Gandy).
After the less forcing 26 Qg5, Davenport's 26 ... h6 is adequate to 
safeguard
the king.  These variations are important because the resulting 
positions seem
substantially clearer than the variations in which Black declines the 
rook.

At this point, the overall evaluation is favorable for Black; that 
is, it
appears that it's up to White to find improvements to justify the 
piece sac
13 Nxg7.  Readers are invited to try to find improvements for either 
side.
It can be fun to make a creative contribution to the analysis of an 
exciting
opening!