Pittsburgh Chess League, Round 1, 9/16/90
Latrobe vs. Transarc, board 1
Mark Eidemiller vs. Bruce Leverett

1 d4    Nf6
2 c4    c5
3 d5    g6

      This is the ``au courant'' way of playing the Benoni.  But you won't
      find it in BCO-2 or MCO-13!  I have been having trouble with the White
      side of this for several years and decided to try the Black side.

4 Nc3   Bg7
5 e4    d6
6 Nf3   O-O
7 h3

      Of course 7 Be2 is OK (as Tom Magar and Tom Martinak played against me),
      but this move is sharper -- also very ``au courant.''

7       Na6
8 Bd3   e5

      But this is very old-fashioned.  For excitement, Black should play
      8 ... e6.  After the text move, White can prevent a serious
      counterattack on either side of the board, so Black is limited to
      arduous defensive maneuvering.

9 g4!

      The logical followup.  In the post-mortem, we considered 9 dxe6 fxe6
      10 e5 to be dangerous, but it accomplishes less than nothing:
      10 ... dxe5 11 Nxe5 Nh5!, or 11 Ne4 b6.

9       Bd7
10 Be3  Kh8

      White has the possibility of Qd2 and Bh6, and Black wants to be able
      to cut across this with ... Ng8.

11 Qe2  Rb8
12 Nd2  Nc7
13 f3?!

      Missing an important opportunity.  After 13 a3 a6 14 b4, Black would
      be forced to defend with 14 ... b6, leaving White with attacking
      chances on both sides of the board.

13      a6
14 a4

      The difference is that now, Black is in no danger on the queenside.
      A draw has become a likely outcome.

14      b6
15 Kd1!?

      Odd-looking, but it is natural for White to connect his rooks while
      leaving them both in their corners.

15      Qe8
16 Kc2  Ng8
17 h4   f5
18 g5

      This consolidates White's space advantage on the kingside, but at
      the same time, I felt that I could easily cope with whatever threats
      White could work up there.  Black could have defended against 18 h5
      by 18 ... fxg4 19 fxg4 gxh5, and now 20 Rxh5 is met by 20 ... Nf6,
      so White has nothing better than 21 gxh5, with the h-file still closed.

18      Ne7
19 b3   b5
20 axb5

      20 a5 would put an end to all activity by Black or White on the
      queenside and in the center.  That was safest and perhaps best.

20      axb5
21 cxb5??

      White should have sat tight.  21 Ra7, which I feared at the time,
      doesn't achieve anything:  21 ... b4 22 Rxc7? bxc3 23 Kxc3 (or
      23 Nb1) Nc8!, trapping the rook and threatening 24 ... Qd8, winning
      the exchange.

21      Nexd5!

      The routine 21 ... Nxb5 gives Black a miserable defense after
      22 Nxb5 Bxb5 23 Bxb5 Rxb5 24 Nc4, etc.  Oddly enough that was what
      I was planning to play.  I didn't see 21 ... Nexd5 until my opponent
      had played his 21st move.

22 Nxd5

      Equally critical is 22 cxd5.  Play would be similar to the game:
      22 ... e4 23 fxe4 Qe5 24 Ndb1, and now 24 ... fxe4 followed by
      25 ... Nxb5 looks convincing.

22      Nxd5
23 exd5 e4

      This sacrifice is always something to watch for in King's Indian-like
      positions.  White has to capture, but then his own pawn on e4 becomes
      his enemy.

24 fxe4 Qe5
25 Rab1 f4!

      In the post-mortem, my opponent said that he had seen the piece
      sacrifice when he played his 21st move, but missed this move.

26 Bf2  f3!
27 Qf1?

      No one considered that White could play 27 Nxf3 Qc3+ 28 Kd1 Bg4;
      but now 29 Be1 traps the queen!  The best I can find for Black is
      29 ... Rxf3 30 Bxc3 Rxd3+ 31 Bd2 Bxe2+ 32 Kxe2 Rd4 33 Bc3 Rxe4+
      34 Kd3 Rg4.  I think that Black has winning chances in the rook
      endgame, but it's not easy.  Also, 27 Qe3? Qc3+ 28 Kd1 Bd4 wins.

27      Ra8
28 Be3

      There isn't anything good.  28 Rg1 (to keep the Black bishop out
      of g4) is answered by 28 ... Qc3+ 29 Kd1 Ra2 30 Nc4 (or 30 Be3) Bh3!

28      Bg4
29 Bf2  Ra2+
30 Kd1  Qc3
31 Be1

      Or 31 Nc4 Re2.

31      f2+
32 Be2  Qc2 mate