GM Analysis and GM Evaluation

I encountered this position while working on the Slav. It was featured in the October 1983 Chess Life (p.26), in an 'Opening Forum' column by GM Leonid Shamkovich.

White played 15.Nd2, threatening to trap the Queen with Nd2-c4. Black countered with 15...a6, but White achieved a won game with 16.Bxc6 Bxc6 17.Nc4 Qf5 18.Bd6 e3 19.Nxe3 Qxb1+ 20.Qxb1 Bxd6 21.Qxb6. The players on first board were the national champions of their respective countries.

Radio Match USA-USSR (Game 2) 1945
Denker, Arnold

Botvinnik, Mikhail
(After 14.Bc8-d7)
[FEN "r3k2r/p2b1ppp/1pn1p3/qB6/3PpB2/bQP2N2/P4PPP/1R2K2R w Kkq - 0 15"]

Shamkovich proposed 15...O-O, and if 16.Nc4, then 16...Nxd4! wins. He gave several alternatives for White.

• 16.Qc4 a6 17.Bxc6 b5, and
  
  
• 16.Be3 Rfc8 17.Nc4 Nxd4 18.Bxd4 Qxb5. Now there are again two main lines.
  
  
  • 19.Nxa3 Qg5 20.O-O e5 21.Be3 Qg6 22.Kh1 Be6
    
    
  • 19.Qxb5 Bxb5 20.Nxa3 Bd3 21.Rc1 f6
    

  Here Shamkovich remarked, 'In both cases, I believe Black has adequate compensation for the piece. He has two strong Pawns, he can attack White's weak c-Pawn, and he won't be challenged by White's misplaced Knight at a3 for a while.'

I confirmed the GM's nice analysis, which was prepared before computers were available for tactical checking. It is useful to note that the GM and modern chess software arrive at the main variations. The GM's evaluation was another matter. With only two Pawns for a piece, the computer values the position at a Pawn down. Who is right, the GM or the computer?

To play through the complete game see...

Mikhail Botvinnik vs Arnold Denker, Radio Match USA-USSR (2) 1945
http://www.chessgames.com/perl/chessgame?gid=1032118

...on Chessgames.com.