Upsetting Chess

IN the first round of the English 65+ championship there were a few minor upsets when heavily outrated players drew with their betters, but no major upsets, ie no one managed to beat a stronger opponent.

By contrast, the first round of the English 50+ championship saw no minor upsets, but there was a major one when the top-seeded grandmaster lost.

Mark Hebden (2451 ECF/2383 Fide) - Natasha Regan (2072 ECF/2056 Fide)

Queen's Gambit Accepted

1.d4 d5 2.c4 dxc4 3.Nf3 Nf6 4.e3 e6 5.Bxc4 a6

This is second in popularity in ChessBase's 2025 Mega database, albeit a long way behind 5...c5.

6.0-0 Nbd7!?

In Mega25 the text scores three percentage points better than the more popular 6...b5 and 6...c5, which nevertheless are preferred by Stockfish17 and Dragon1.

How should White respond?

*****

*****

*****

*****

*****

7.a3!?

The engines prefer the much more popular 7.a4, which seeks to prevent Black's normal queenside expansion in the QGA.

7...b5 8.Ba2 c5 9.Nc3 Bb7

The game has transposed to a fairly well-known position, with 256 examples in Mega25

*****

*****

*****

*****

*****

10.Qe2 Qb6 11.Rd1 Be7!?

Deviating from Hebden (2485) - Matthew Sadler (2480), Hastings Challengers 1991-2, which saw 11...Rd8 12.e4 cxd4 13.Nxd4 Ne5 14.Bf4 Ng6 15.Be3 Bc5 16.Nc2 0-0 17.Bxc5 Qxc5, with a slight edge for Black, according to the engines (0-1, 34 moves).

12.e4 cxd4?!

The same response to e4 as in Hebden-Sadler, but this time the engines reckon it is dubious, preferring 12...b4 or 12...c4.

13.Nxd4 Ne5 14.Bf4 Ng6!?

This is an improvement on the known move 14...Nc6?, when the engines point out the strength of 15.Nxc6 Bxc6 (15...Qc6 is even worse, they reckon) 16.Nd5!, eg 16...exd5 17.exd5 Bb7 18.d6 etc, or 16...Bxd5 17.exd5 exd5 18.Bxd5 Nxd5 19.Rxd5 Rd8 20.Re5 Rd7 21.Re1 Qd8 22.Re4!? Rb7!? (22...f6 23.Bd6) 23.h4!? f6 24.b4!? Kf7 25.Qa2+ Kf8 26.Qe6 Rd7 27.Qa6 etc.

15.Be3 Qc7 16.Rac1 Qe5!?

The black queen is well-placed in the centre of the board, if it can escape harassment from White's pieces and pawns

*****

*****

*****

*****

*****

17.f3 h5!?

Going for it, but the engines prefer 17...0-0.

18.a4?

Hebden apparently underestimates the danger he is in on the kingside. However, while the engines agree the text is bad, and agree White should be slightly better, they struggle to decide exactly how the slight edge should be realised, but perhaps best is 18.Qf2, allowing for the possibility of Qg3 in some lines.

18...b4 19.Nb1 Bd6 20.g3

This is Dragon1's top choice, at least for a while, whereas Stockfish17 suggests 20.Kf1!?, which is obviously not what White wants to be playing.

20...h4 21.g4!?

Perhaps 21.Qg2 is a little better, but Black is on top.

21...Nxg4!?

By no means the only good move.

22.fxg4 Qxe4 23.Qf3

How should Black proceed?

*****

*****

*****

*****

*****

23...Bxh2+!
After 23...Qe5 24.Qxb7! Black has nothing better than 24...Qxe3+, when 25.Kf1 gives a completely equal position, according to the engines.

24.Kg2??

White had to try 24.Kxh2 Qe5+ 25.Bf4 Nxf4 26.Qxb7, when an immediate discovered check is only good enough for a draw However the engines continue 26...0-0!, which Dragon1 reckons is winning, although for quite some time Stockfish17 gives Black only a slight edge after 27.Qc7!? The position is very murky, but best may be 27...Qe3 28.Qb7 Qh3+ 29.Kg1 Qxg4+ 30.Kf1 Qh3+ 31.Kg1 Qg3+ 32.Kh1 Rad8, when the position is still unclear, but probably favours Black, although it is hard to say by how much.

The text overlooks, among other things, that it puts the g4 pawn en prise.

24...Qxg4+ 0-1

Mariánské Lázně Lesson 8: Chess Is Deep

WHAT would you play with white in the following position - it arose in my round-eight game - which started as an Exchange Variation of the French, but quickly became more typical of the Queen's Gambit Accepted?

There are 2,058 examples of this position in ChessBase's 2024 Mega database

*****

*****

*****

*****

*****

*****

Almost three-quarters of the games saw the move I played, 7.0-0, which scores a very respectable 56% in Mega24.

But Stockfish16 and Komodo14.1 prefer 7.Qe2+!?, which occurs 347 times and scores an excellent 62%.

After the main continuation, 7...Qe7 8.Qxe7 Bxe7, queens are off the board, which must reduce White's attacking chances in this IQP position.

Play usually proceeds 9.0-0 0-0 10.Re1 Bd6!? 11.Nc3

Black scores just 25% after 11.Nc3 in Mega24, although the engines reckon White is at best only slightly better

*****

*****

*****

*****

*****

It is not easy, at least for me, to understand what is going on.

I guess the answer is that, as I discovered in my game, with queens on the board, White's position in general, and the isolated queen's pawn in particular, come under strong pressure.

With queens off the board, White is unlikely to develop a mating attack, but the IQP will probably come under little or no pressure, and meanwhile White enjoys a central space advantage with lots of pieces remaining.

LESSON: general considerations, such as keeping queens on the board when playing with an IQP in order to generate attacking chances, take second place to the specific circumstances of a position.

Mariánské Lázně Round Eight

View from the rear of my hotel

FACED my second German of the tournament.

Spanton (1736) - Thomas Müller (1925 - Fide ID: 4652401*)

French Exchange/Queen's Gambit Accepted

1.e4 e6 2.d4 d5 3.exd5 exd5 4.Nf3 Bd6!?

This beats 4...Nf6 for popularity in ChessBase's 2024 Mega database, and scores six percentage points more. One reason for not playing ...Nf6 quickly is that the black king's knight may go to e7 to support ...Bf5, the idea being to swop off White's good bishop (assuming it goes to its usual square of d3).

5.c4!?

More popular is 5.Bd3, but the text becomes attractive after Black spends a tempo on ...Bd6. The point is that after 4.c4 a fairly often recommended reply is 4...Bb4+!?

5...dxc4!?

This is third in popularity behind 5...Nf6!? and especially 5...c6.

6.Bxc4

What's in a name? ChessBase has been calling the opening "C01: French: Exchange Variation," but after 6.Bxc4 switches to labelling it "D20: Queen's Gambit Accepted: 3.e3 and 3.e4"

*****

*****

*****

*****

*****

The diagrammed position dates back in Mega24 to the second Pierre de Saint-Amant - Howard Staunton match of 1843, appearing in game 19 after Staunton accepted Saint-Amant's Queen's Gambit. The Frenchman won in 79 moves, but went on to lose the match +6=4-11.

6...Nf6 7.0-0

Stockfish16 and Komodo14.1 suggest 7.Qe2+, which has occurred at grandmaster level, although to me it seems strange to go for an exchange of queens after contracting an isolated queen's pawn. The engines continue 7...Qe7 8.Qxe7+ Bxe7 9.0-0, claiming an edge for White, an evaluation supported by White scoring 64% with the line in Mega24.

7...0-0 8.Bg5

This position goes back to the first Louis-Charles de La Bourdonnais - Alexander McDonnell match of 1834

*****

*****

*****

*****

*****

8...Nc6

McDonnell preferred 8...h6 9.Bh4 g5!? 10.Bg3 Bg4, but lost in 35 moves.

9.Nc3 Bg4 10.Nd5!? Be7 11.Nxe7+ Qxe7

White has won the bishop-pair, but the d pawn is weak.

How should White continue?

*****

*****

*****

*****

*****

12.Re1

Hitting the black queen is reasonably popular, but the strongest players have tended to prefer 12.Bd5!? or 12.h3!? After the latter, which is the commonest continuation in Mega24, 12...Bxf3 12.Qxf3 Nxd4 looks tempting. White then has 13.Qxb7, when the engines give what appears to be a novelty, 13...Qc5!? After the forced 14.Bxf6, the engines reckon 14...Qxc4 15.b3!? Ne2+ 16.Kh1 leaves White slightly better, wherever the black queen goes.

12...Qd6

This is the only move in Mega24, and it gives complete equality, according to Stockfish16, which reckons 12...Qb4 does the same. Komodo14.1 prefers the text.

13.Rc1!?

This may be a novelty. The engines reckon it and the known 13.d5 maintain complete equality, albeit in a sharp position in which one slip is likely to be fatal.

13...Rae8

Black must avoid 13...Bxf3 14.Qxf3 Nxd4 15.Qxb7 Rfb8? as White has 16.Bxf7+ Kxf7 17.Rxc7+, eg 17...Kf8 18.Rf7+ Kg8 19.Rxg7+ Kh8 20.Qf7 with a winning position (Stockfish16), or at least the upper hand (Komodo14.1). However Black can improve with 15...Ng4 16.g3 Rab8 17.Qe4 Nf6 18.Qd3 Rxb2, although the engines reckon White has enough for a pawn.

14.Rxe8!? Rxe8

What should White play?

*****

*****

*****

*****

*****

15.Bb5?

Moving the same piece twice! OK, we are in a middlegame, rather than the opening, at least in my opinion, but the position is sharp. I should have preferred the complications of 15.Qb3, when the engines reckon best play runs 15...Bxf3!? 16.Bxf7+ Kf8 17.Bxe8 Bd5 with ...Nxe8 to come, either before or after ...Be6. That is equal, according to the engines, which marginally prefer 15.h3!? Bxf3 16.Qxf3 Qxd4, after which they reckon White has enough for a pawn, eg 17.Bxf6 Qxf6 18.Qxf6 gxf6 19.Bd5 Re2 20.Bxc6 bxc6 21.Rxc6 Rxb2 22.a4 Ra2 23.Ra6.
15...Bxf3 16.Qxf3?

White is lost after this, although 16.gxf3 leaves White with a deeply unpleasant position.

16...Nxd4 17.Qxb7

Black to play and win

*****

*****

*****

*****

*****

17...Ne2+?

I feared 17...Ng4, which is indeed strong as long as after 18.g3 Black finds 18...Qb4! The engines reckon even stronger is the immediate 17...Qb4!, after which White can only 'sensibly' defend against back-rank mate by giving up the light-square bishop. Other moves, eg pushing the h pawn one or two squares, give equality.

After the text White wins with a simple forcing sequence;

18.Bxe2 Rxe2 19.Bxf6 gxf6 20.Qc8+ 1-0

*Müller is the commonest family name in Germany.

Lessons From Coulsdon III

Spanton (1965 ECF/1747 Fide) - Tom C O'Neill (1510 ECF/1416 Fide)

Coulsdon Late Summer U2000 Round 4

Queen's Gambit Accepted

1.d4 Nf6 2.Nf3 d5 3.c4 dxc4 4.e3 e6 5.Bxc4 Bb4+!?

This looks odd in that Black is apparently offering to trade off  a white bishop hemmed in by white central pawns. However Gunsberg tried it in his 1890-1 world championship match with Steinitz. The main moves are 5...c5 and 5...a6.

6.Bd2

Despite what I stated earlier about an exchange of bishops, Steinitz against Gunsberg preferred 6.Nc3, which has also been played by Capablanca, Botvinnik and Fine. If then 6...Ne4, White has 7.Qc2, so Black usually castles, but that reaches a position in ChessBase's 2021 Mega database that is normally reached with Black to move.

6...Bxd2+ 7.Qxd2

More popular in Mega21 is 7.Nbxd2!?

7...0-0

Again ...Ne4 can be met by Qc2.

8.Nc3

How should Black proceed?

*****

*****

*****

*****

*****

8...c6?!

This is almost certainly too passive, leaving White a free hand in the centre.

Generally speaking, Black should be looking to challenge White's centre by preparing ...e5 or, more commonly, ...c5. The analysis engines Stockfish14 and Komodo12.1.1 suggest 8...b6, while the most-popular move in Mega21, albeit from a small sample, is 8...Nc6.

9.0-0 b5!? 10.Bb3!?

The engines prefer 10.Bd3.

10...Na6 11.a4!?

The engines prefer 11.Rfd1.

11...b4

Forcing this was the point of my last move, but the text is why the engines were not keen on 11.a4!? They reckon that after  11...b4 White is only slightly better, whereas I was hopeful that long-term Black would have trouble with queenside weaknesses.

12.Na2 Rb8

The engines suggest 12...c5 13.Bc4 Bb7, but still prefer White.

13.Rac1 Bb7 14.Ne5 Rc8 15.Rfd1 Ne4?!

The knight will be something of a lone raider. The engines like 15...c5 16.Bc4 Qd6 with what they reckon is a slight edge to White.

16.Qe1 Qg5?!

The engines still prefer ...c5, but give White the upper hand.

17.f3 Nec5??

Black had to drop the knight back to the third rank, but is losing a pawn, eg 17...Nd6 18.Nxb4 Nxb4 19.f4 and 20.Qxb4.

18.f4 Qe7 19.dxc5 Nxc5?? 20.Qxb4 (1-0, 31 moves)