Lessons From Mallorca: Danger Of Analogous Positions

IN my round-one game at the Colònia de Sant Jordi U2400 I played the Richter-Rauzer Variation of the Sicilian.

The game began 1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 d6 6.Bg5.

This was the first time I have reached the position, with either colour.

My theoretical knowledge was virtually zero - indeed I knew no more than the basic idea behind the move, which is summed up in Wikipedia thus: "The move 6.Bg5 was Kurt Richter's invention, threatening to double Black's pawns after Bxf6 and forestalling the Dragon by rendering 6...g6 unplayable."

Ignore the first part of that sentence - Richter was at most four years old when the move was first played.

However the basic idea behind the move - an attempt to make the Dragon "unplayable" - is correct, although not everyone agrees 6...g6!? 7.Bxf6 is all that bad for Black.

Indeed it has been played by plenty of grandmasters, and is a speciality of some, including 2700+ Russian star Daniil Dubov, albeit it is a line he likes to play at fast time-controls.

I knew nothing of this, but I knew Black's commonest reply to 6.Bg5 is 6...e6, so I was slightly surprised when my opponent chose 6...Qb6!?

There are 1,778 games in ChessBase's 2024 Mega database with the position after 6...Qb6!?

*****

*****

*****

*****

*****

White's normal continuation, 7.Nb3, occurs 1,458 times in Mega24, and is the top choice of Dragon1.

Stockfish17 for a long time suggests 7.Be3!?, which occurs 112 times, but scores 60% - nine percentage points more than 7.Nb3.

However, given enough time, Stockfish17 comes to marginally prefer Dragon1's choice.

I played 7.Bxf6!?, which, like the other two moves, has also been played by grandmasters, and occurs 136 times, although scoring only 47%.

The engines reckon my choice is dubious - after all, it gives up the bishop-pair and strengths Black's centre, albeit weakening the black kingside.

After 7...gxf6 I felt more-or-less obliged to play 8.Nb3, although the engines prefer the rare 8.Bb5!?

Now I do not believe 7.Bxf6!? is a mistake - my choosing it places me in the company of at least two 2600+ players.

However, my reason for choosing it was a mistake, in as much as I played a move that I knew (or thought I knew) was good after 6...g6!?, but in a position in which my opponent had played something completely different.

Instead of analysing the position, I played on the general principle of wanting to punish an opponent who had not chosen the main line.

LESSON: general principles make a useful starting point for considering a position, but analysis must prevail.

Sant Jordi Round One

Spanton (2013) - Luis Abadía Pascual (1836)

Sicilian Richter-Rauzer

1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 d6 6.Bg5 Qb6!?

The most popular moves in ChessBase's 2024 Mega database are 6...Bd7 and especially 6...e6.

How should White respond?

*****

*****

*****

*****

*****

7.Bxf6!?

The mainline in Mega24 runs 7.Nb3 e6 8.Qd2 Be7 9.0-0-0 0-0 10.f3 Rd8 11.Be3 Qc7, with what Stockfish17 and Dragon1 reckon is a game with equal chances.

7...gxf6 8.Nb3 Be6!?

The engines reckon this is a positional mistake, preferring 8...f5, as played in Orest Popovych - Bent Larsen, Canadian Open (Toronto) 1968, which continued 9.Qh5!? fxe4 10.Bc4 Ne5 11.Bb5+ Kd8!?, when the engines reckon 12.Be2 would have given White full compensation for a pawn.

9.Nd5 Bxd5 10.exd5 Ne5 11.f4?

White has a winning game after 11.Bd3!? or 11.c3, according to the engines, eg 11.Bd3!? Nxd3 12.Qxd3, when I guess they like White's extra space and superior pawn-structure.

11...Qe3+ 12.Qe2 Qxf4 13.Qb5+

The engines suggest 13.g3!?, continuing 13...Qb4 14.c3 Qg4!?, and now 15.Qb5+ or 15.Nd4, claiming Black is only slightly better.

13...Nd7 14.Qxb7 Qe3+ 15.Be2 Rb8 16.Qc6 Rg8 17.g3 Rg5 18.Rf1?!

Probably better is 18.Rd1.

18...Re5 19.Rf2?!

Probably better is 19.Qc4, but 19...Bh6 leaves Black well on top.

Black to play and gain an advantage worth about 2.75 pawns, according to the engines

*****

*****

*****

*****

*****

19...Bh6

This looks natural, especially as it is an example of 'inviting everyone to the party', but winning is the engines' 19...Qe4! White has no good answer to the threat of ...Qh1+, eg 20.Rd1?! Qh1+ 21.Kd2 Bh6+ 22.Kd3 Qe4+ 23.Kc3 Qb4+ 24.Kd3 Re3#. Best, according to the engines, is 20.Qc3, but simply 20...Qxd5 gives a winning advantage, and Black probably has even better moves.

20.Kf1!

This equalises by making Bf3 possible in answer to 20...Qe4.

20...f5?!

The engines like 20...Qe4 or 20...Qb6.

White to play and gain an advantage worth about 1.75 pawns, according to the engines

*****

*****

*****

*****

*****

21.Rd1

Black has no good answer to 21.Na5!, eg 21...Qc5 22.Qxc5 and 23.Nc6, or 21...Qd2 22.Nc4 Qxd5 23.Nxe5.

21...Qb6  22.Bd3 e6 23.Qxc6 Nxb6!? 24.dxe6 fxe6 25.Nd4?!

This allows a strong reply, so White probably needs to play 25.c4!?, albeit the engines prefer Black.

Black to play and gain at least the upper hand, according to the engines

*****

*****

*****

*****

*****

25...Nd5!

Black ignores the threat of a fork at c6 to set up the threat of an even stronger fork at e3.

26.Re1?

The engines give 26.Rb1, but reckon 26...Ne3+ or, probably even stronger, 26...Kd7 with ...Ne3+ to come, gives at least the upper hand.

26...Ne3+ 27.Kg1 Rxb2?

Black lets himself be distracted by the apparently easy win of a pawn, whereas 27...Ng4 (27...Nd5 is also strong) wins, eg 28.Rfe2 Be3+ nets at least the exchange.

28.Rfe2 Ng4 29.Nxe6! Be3+ 30.Kg2 Rxe6

Possibly better is the engines' 30...Bd2!? 31.Rxe5 dxe5 32.Re2 Bh6, when the threat of 33...Ne3+ prevents 33.Bxf5?

31.Bxf5 Re5 32.Bxg4 Bd4 33.Be6

This allows a sharp reply, whereas 33.Bf5 h6 34.a4 gives White at least the better side of equality, according to the engines.

33...Rxc2! 34.Rxc2 Rxe1 35.Bf5 h6 36.Bg4 Bc5 37.h3 Kf7 38.Bh5+ Kf6 39.g4!?

Getting a pawn off a dark square, so it cannot be a target for Black's bishop, but this lets the black king into the white kingside.

39...Kg5 40.a4 Kh4 ½–½

If anyone is better in the final position, it is Black, but LAP was relieved to have made the first part of the time control (40 moves in 90 minutes, with 15 minutes to finish, and a 30-second increment throughout), and anyway, as he pointed out, the game should now be drawn.