It arose in my round five game at the Thuringia seniors' championships, in which I was White against Erich Krüger (2033).
There was a time when theoreticians more-or-less condemned 5...exf6, certainly regarding it as inferior to 5...gxf6.
Reuben Fine in The Ideas Behind The Chess Openings (Bell, 1948) wrote that "Black has no compensation for his minority of pawns on the queenside; the sole reason why he can so often draw in practice is that the white majority is so hard to exploit."
Fine insisted: "Regardless of how White continues, Black's game remains too passive."
Raymond Keene in The Openings In Modern Theory And Practice (Bell & Hyman, 1979) said of 5...exf6: "Black's kingside is extremely solid, and the deterioration in his pawn-structure is not likely to cause him too many problems, since reduction to an instant pawn ending is unlikely."
But he added: "Black can normally hold the draw, but should hardly hope for more unless White overpresses."
Even a relatively-modern specialist book on the opening, for example Joe Gallagher's Starting Out: The Caro-Kann (Everyman, 2002), while calling ...exf6 more solid than ...gxf6, claimed "it's not thought of too highly these days."
Gallagher explained: "Variations where Black accepts a disadvantage and just tries to hang on for the draw are not very popular nowadays."
So why has ...exf6 leapt to prominence?
Wikipedia reckons the variation "offers Black rapid development, though also ceding White the superior pawn-structure and long-term prospects (Black has to be cautious that the d-pawn is now a potential passed pawn in the endgame)."
 |
| No one can argue against White's pawn-structure being superior in an ending, but does the position really offer Black "rapid development"? |
I decided it would be interesting, and instructive, to look at what normally happens in the Tartakower, using ChessBase's 2026 Mega database as the reference point.
White usually continues 6.c3, which occurs 56.8% of the time (11,480 games out of 20,229).
Immediately, White appears to have given away the advantage of the first move, but why?
The point is White would like to play Bd3, but in the diagram position that loses a pawn to ...Qxd4.
White gets some compensation for the pawn, but the transaction is unlikely to prove profitable.
White could instead defend the pawn with 6.Nf3, which occurs 18.6% of the time, but that is not without its problems as the move makes the knight vulnerable to a later pin by ...Bg4, eg the main line in Mega26 after 6.Nf3 continues 6...Bd6 7.Bd3 0-0 8.0-0 Bg4, when Black has equalised, according to Stockfish17.1 and Dragon1.
After 6.c3 Bd6 7.Bd3 0-0 White usually plays 8.Qc2, which occurs 63.8% of the time (3,051 games out of 9,164).
After 8...Re8+ White plays 9.Ne2 an overwhelming 97.3% of the time (4,779 games out of 4,912).
Black then has to do something about his attacked h pawn, and he nearly always plays 9...h5, which gets the pawn off prise, as Nigel Davies is fond of saying, AND prepares to weaken White's castled position, assuming White castles short, which is usually the case.
Indeed, in 50.9% of the games (2,309 out of 4,539) White immediately plays 10.0-0.
 |
| Both sides having castled, it is a good moment to take stock |
*****
*****
*****
*****
Despite Wikipedia boasting of Black's "rapid" development, it is White who has a slight edge in this department - both sides have castled, but White has in addition developed queen, bishop and knight, while Black has only developed rook and bishop.
True, it is Black to move, and Black could level the number of pieces developed, but instead Black normally plays 10...h4.
White in 1,213 games out of 1,790 (67.8%) defends the kingside with 11.h3, and Black replies 90% of the time with 11...Nd7.
White's 12.Bd2 occurs in 523 games out of 1,153 (45%) - but d2 it is not a great square for the bishop, although where else is it to go?
After 12...Nf8 (99.1%) 13.Rae1 (88%) Black's most common move, but by no means a consensus choice, is 13...Bc7 (147 games out of 497 - 29.6%), moving the king's bishop for the second time, while leaving most of the queenside undeveloped.
 |
| After 13...Bc7 Black has developed just two pieces, the king's rook and bishop, against White's five pieces |
Clearly, whatever Black is relying on in the main line of the Tartakower Variation, it is not rapid development.
The above position occurs 149 times in Mega26, with quite a few high-profile games, the highest-rated being Abhimanyu Mishra (2610) - Nodirbek Abdusattorov (2771), Fide Grand Swiss (Samarkand, Uzbekistan) 2025.
That continued 14.f4 Qd6 15.c4 b6 16.Rf3 Ne6 17.Be3 Bd7, bringing about the following position.
 |
| White has at least a slight edge, and possibly the upper hand, according to the engines |
*****
*****
*****
*****
After 18.Red1 g6 19.c5 bxc5 20.dxc5 Qe7 Black was still to complete his development (the queen's rook is unmoved), and White took the chance to start an attack with 21.f5 Ng5 (21...gxf5 was better, according to the engines) 22.Bxg5 fxg5 23.fxg6 f6.
Now the engines reckon 24.Nd4 gives at least the upper hand. Instead 24.Bc4+ Kg7 25.Bf7 Red8 26.Rdf1 Qe5 27.Qc3 Bc8 resulted in an unclear middlegame, with equal chances, according to the engines.
They actually rate the position as completely even, and the game more-or-less stayed that way until being agreed drawn, with only kings left on the board (!), at move 69.
For the present, at least, the Tartakower Variation is here to stay.
No comments:
Post a Comment