Continuing with the world class engine-to-engine competition TCEC Season 7, I discovered in the previous post Komodo - Stockfish Superfinal Openings, that the Gruenfeld Defense, Exchange Variation (D85) was one of only two hand-picked openings (out of 32) that had been battered +2-0=0. Since I'm an occasional Gruenfeld player myself, this is particularly worrying, and I decided to look at the two games in more depth.
Explaining a game between two engines is a nearly impossible task. If you have ever looked at a long tablebase endgame, the two sides seem to maneuver back-and-forth randomly for move after move until a decisive position appears on the board. That is how most engine-to-engine games look to me, so I'll just mention a few key positions.
The first eight moves for both sides were dictated by the organizers: 1.d4 Nf6 2.c4 g6 3.Nc3 d5 4.cxd5 Nxd5 5.e4 Nxc3 6.bxc3 Bg7 7.Be3 c5 8.Qd2 O-O. Among GM-level players, the moves 8...Qa5 and 8...cxd4 are both more popular than 8...O-O. I don't really know why that is, but it's a sign that the Gruenfeld is not yet hopeless for Black.
Both games then followed the same path for the next few moves: 9.Rc1 Qa5 10.Nf3 Bg4 11.d5, but here they diverged. Komodo captured 11...Bxf3, while Stockfish continued developing with 11...Na6 .
The following diagrams show a few key points in both games. The left diagram for each game shows the position after move 20. The right diagram shows the move that first wins material.
The winning strategy in both games looks similar. The passed, protected d5-Pawn and the extra space give White an advantage in maneuvering. White uses the resulting initiative to attack on the various open and semi-open files. Black eventually cracks under the pressure and loses the equivalent of a Pawn. The extra material is enough to win the game.
To play through the games, see
- Stockfish vs Komodo; TCEC Season 7 - Superfinal (2014); g.7, and
- Komodo vs Stockfish; TCEC Season 7 - Superfinal (2014); g.8