Lessons From Bregenz

IN round one of the Bodensee Seniors I was White against Kurt Vogel (1643).

The game began as a Scandinavian, 1.e4 d5, but I mixed things up by switching into the Blackmar-Diemer Gambit with 2.d4!?

There are 2,034 examples of this transpositional trick in ChessBase's 2026 Mega database, although, on the face of it, the trick rebounds on White as 2.d4!? scores 48%, seven percentage points lower than the mainline 2.exd5.

Then again, of all the moves after 1.e4 d5 that appear at least 100 times in Mega26, 2.d4!? records the second-highest score, so it has something going for it.

Not only that, but it has been played by grandmasters, although German GM Matthias Wahls, in a 1997 article for ChessBase, wrote: "The vast majority of Scandinavian Defence supporters are aware of the possibility of becoming involved in the BDG."

He furthered this by giving d4 a question mark, explaining that he had been taught a pawn is worth "roughly two to three tempi," whereas in the BDG White "only receives a single tempo for the sacrifice."

With all due respect, I am not sure Wahls is right about this.

After 2...dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3, the following position is reached 3,407 times in Mega26.

It is Black to move, but I am showing the board from White's viewpoint

Black has accepted the sacrifice by taking on f3 - until then it was still possible to decline the gambit, for example by 4...e3 (585 examples in Mega26).

After Nxf3, White has developed two pieces, against Black's one, and has opened diagonals for both white bishops, whereas Black has yet to open a diagonal for the f8 bishop.

I make that an advantage of two development tempi, and it will be three once White castles kingside, as the king's rook will land on the half-open f file.

True, it is Black to move, but even so, by my reckoning, White has the "roughly two to three tempi" that Wahls was taught is necessary for a positional pawn sacrifice.

Then again, show the diagram position to Stockfish18 and Dragon1, and they reckon Black has the upper hand, which rather supports Wahls' opinion.

Black should continue, according to the engines, with 5...c6, 5...Bf5 or 5...g6!? The last move is interesting as it was Diemer's choice when Bogoljubow played the gambit against him in 1952 at Baden-Baden. ChessBase gives the game as a White win, but Black made the last move, the final position is dead drawn, and no explanation is given as to what happened, so it is anyone's guess as to how the game ended.

Be that as it may, the main line in Mega26 runs 5...Bg4 6.h3 Bh5 (6...Bxf3 is also popular) 7.g4!? Bg6 8.Ne5 e6 9.Bg2 c6, reaching the following position.

White has developed three pieces, including a knight to the fifth rank, which, according to Tarrasch, counts as two development tempi

Black has developed two pieces, but it is White to move, and White can still castle kingside, developing the king's rook to a half-open file. Alternatively, White might castle long, in which case the queen's rook
will be developed to a half-closed file - one on which the black queen currently resides.

Again White seems to be getting "roughly two to three tempi" for a pawn.

Naturally, counting tempi is far from being the only way to evaluate opening play, and, in the diagram, after 6...c6, White has more space in the centre and on the kingside, but also a lot of holes.

This time the engines have a higher regard for White's play, rating the position as giving equal chances (Stockfish18) or at best giving Black a slight edge (Dragon1).

LESSON: as is often the case in chess, there is plenty of room for divergence of opinion, and the objective merits of a move, especially if it involves a gambit, may be very different from the practical results.