The Kramnik-Shirov Counterattack: Playing Black with Confidence

The first thing to notice about this position is that Black has willingly traded a bishop for a knight on c3. Why? Because the resulting doubled c-pawns give Black a long-term structural target. After White recaptures — most commonly with bxc3 — Black's next moves aim to seize space in the centre. The engine's best continuation is bxc3 e4 Nd4 Nf6, where Black pushes forward with ...e4, kicking the f3-knight to d4, and then develops the king's knight. This isn't just a random sequence: Black gains central space and tempo on the white knight. You're not worse here — you're fully equal with a clear plan.

White has two serious ways to take back on c3: bxc3 or dxc3. Across over 73,000 games in the Lichess database, bxc3 has been played in 54,159 games (White scores 50.2%), while dxc3 appears in 18,853 games (White scores 49.9%). Statistically, they are remarkably similar in results — both give White roughly half the points. But they lead to very different pawn structures. With bxc3, White has a double c-pawn and an open b-file; with dxc3, the centre becomes fluid and White's pawns are healthier but the dark-squared bishop is less active. Your strategy as Black shifts depending on which recapture White chooses. Against bxc3, push e4 and seize space. Against dxc3, focus on developing quickly and pressuring the centre.

While the main recaptures are perfectly playable, some tempting alternatives are catastrophic. The database reveals four moves that are outright blunders for White, and each one hands you a huge advantage if you spot it. Here's what to look out for: - b3 (60 games, White scores 13.3%): Loses roughly six pawns according to the engine. White should have played dxc3 instead. - Nxe5 (42 games, White scores 7.1%): Loses roughly five pawns. That greedy knight grab on e5 is a trap you'll happily let White fall into. - e3 (35 games, White scores 17.1%): Loses roughly 3.4 pawns. A timid move that gives away White's edge. - g3 (29 games, White scores 17.2%): While not listed with precise loss numbers like the others, its terrible score speaks for itself. If White plays any of these, you're already significantly better. The engine says dxc3 was the better move in each case, so when White avoids the main lines, you should be alert for a chance to capitalise.

When White plays the most popular move bxc3, the engine's best response is e4, pushing that pawn forward with tempo. The knight has to move, and Nd4 is the natural square. From there, you continue developing with Nf6. Your knights are well-placed, your pawn on e4 controls key central squares, and you have the bishop pair (since White lost their dark-squared bishop in the trade). Your plan is straightforward: finish development (Be7 or Bd6, O-O), target the doubled c-pawns as a long-term weakness, and keep that central space advantage. White's doubled pawns on the c-file are a structural liability in the endgame, so don't rush to trade pieces early — use your space and bishop pair to create threats first.