Lessons From South Wales VII

IN round seven at the South Wales International I had white against a Fide master rated 2289.

The full game can be seen at S7 but here I want to concentrate on the opening, which featured a line of the currently fashionable Tartakower Variation of the Caro-Kann.

That starts 1.e4 c6 2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nf6 5.Nxf6+ exf6.

How should White continue?

*****

*****

*****

*****

*****
I played what is easily the commonest continuation in ChessBase's 2022 Mega database, 6.c3. This may look slow but is likely to be needed at some point to support the d pawn, to let White possibly set-up a queen-bishop battery on the b1-h7 diagonal and to make protection of b2 easier.

The vast majority of games in Mega22 continue 6...Bd6 7.Bd3 0-0, and now 8.Qc2 is most popular but Magnus Carlsen is among those who have preferred 8.Ne2.

After 8...Re8 9.0-0 the continuation 9...Nd7 is probably considered best theoretically, although 9...Qc7, 9...Na6, 9...Bg4, 9...h5?!, 9...f5 and 9...g6 have also been tried by grandmasters.

The most popular reply to ...Nd7 is 10.Bf4, when Magnus Carlsen (2863) - Rasmus Svane (2613), Lichess Online Blitz 2020, saw 10...Nf8 11.Bxd6 Qxd6 12.Re1 with a position Stockfish15 and Komodo13.02 reckon is equal (but 1-0, 51 moves). I played 10.Ng3, when 10...g6 covered the f5 and h5 squares.

My 11.Qf3!? has been played by a 2446 but is not liked by the engines. Stockfish15 reckons 11.Bh6!? gives White a slight edge after 11...f5 12.Qd2, while Komodo13.02 prefers 11.Ne4 Be7 12.Re1, again claiming a slight edge for White.

After Qf3 the game saw 11...Nb6 12.Ne4 Be7, which also occurred in Aleksandra Dimitrova (2246) - Anna Gvanceladze (2238), Moscow Women's Championship 2017.

My 13.Re1 may be a novelty and is preferred by the engines to the Moscow game's 13.Bf4, but after 13...Be6 the engines do not like 14.Bd2, after which 14...f5 15.Ng3 a5 saw Black get a queenside initiative. Instead they reckon 14.Bf1!? is good, when 14...f5 15.Nc5!? Bxc5 16.dxc5 Na4 17.b3!? Nxc5 18.c4 is a promising pawn sacrifice, according to the engines, who prefer White's bishop-pair to Black's extra but doubled pawn.

CONCLUSION: the Tartakower Variation of the Caro-Kann has been around more than 120 years and is still evolving. It gives Black easy development at the cost of an inferior pawn-structure. White's superior structure is an endgame asset, but there is much play left before any ending is reached.