Lessons From Mariánské Lázně VII

EDMAR Mednis in his book Practical Opening Tips (Cadogan 1997) spends a chapter discussing the merits of pawn captures towards the centre, as opposed to pawn captures away from the centre.

When Black, for example, plays ...Bxb3 it is nearly always better to recapture with axb3 rather than cxb3.

Generally speaking, the nearer a pawn is to the centre, the stronger it is.

This particularly applies to a rook's pawn, which, when 'promoted' to a knight's pawn, doubles the number of squares it controls.

Of course, in the great scheme of things, we are talking small margins - a better pawn-structure is unlikely to make up for the loss of a piece, although it might make up for the loss of a pawn.

A prominent exception to the desirability of capturing towards the centre comes in the Spanish, after White captures a knight on c6.

This commonly happens in the Exchange Variation (1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Bxc6), but also occurs in delayed forms of the Exchange, and in some other lines.

My round-seven game at Mariánské Lázně, when I had white against Poland's Michał Nisztuk (1696), featured one such line.

It started as a Berlin Defence to the Spanish, ie 1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6.

I continued with the relatively unusual 4.Qe2!?, and after 4...Bc5!?, which is the most popular reply in ChessBase's 2026 Mega database, I played the rare, but engine-approved, 5.Bxc6!?, which occurs in 56 out of 788 games in Mega26, ie 7.1% of the time.

Position after White's capture on c6

*****

*****

*****

*****

*****

The game saw 5...dxc6?!, but after 6.Nxe5 Qd4 7.Nd3 it was clear something had gone wrong for Black, who, with best play, is not getting the pawn back.

It turns out the diagram is one of those exceptional cases in the Spanish when a capture on c6 should be answered with the positionally desirable option of capturing towards the centre.

After 5...bxc6 the engines still approve of 6.Nxe5, but then 6...Qe7 7.Nd3 can be met by 7...Ba6, when Stockfish17.1 and Dragon1 reckon Black has at least a slight edge.
LESSON: general rules about which way to recapture are useful guidelines but, as American author and international master John Watson would be sure to point out, they should not be relied on - calculation is necessary.

Lessons From Mariánské Lázně VI

IN round six I was on the black side of an Accelerated Dragon against Fide master Jan Bartoš (2124).

After 1.e4 c5 2.Nf3 g6 3.d4 cxd4 4.Nxd4 Nc6 5.Nc3 Bg7 6.Be3 Nf6 the commonest continuation is 7.Bc4, but when Garry Kasparov reached the position in a 2022 rapid game, he preferred 7.Nxc6, and this is what my opponent played.

After 7...bxc6, 4,173 games out of 4,567 (that is 91.4%) in ChessBase's 2026 Mega database have the continuation 8.e5, but my game saw 8.Bc4!?

Position after 8.Bc4?!, a move new to me

*****

*****

*****

*****

*****

It was perhaps naive of me to think an FM may not have properly considered Black's possibility of relying on the well-known central fork trick to capture on e4.

However, it seems he had not, as after 8...Nxe4?! 9.Nxe4 d5 JB failed to find, or perhaps failed to properly evaluate, 10.Bd4!, instead playing the 'normal' 10.Bd3, when the position is equal, according to Stockfish17.1 and Dragon1, although I still went on to lose.

10.Bd4!. however, causes Black problems, eg 10...0-0 11.Bxg7 Kxg7 12.Qd4+ f6 13.0-0-0, when White has a huge lead in development, and Black has still not regained the sacrificed piece.

After 13...Qb6 14.Rhe1 Qxd4 15.Rxd4 e5! 16.Rd2 dxc4 Black has got the piece back and, temporarily, won a pawn into the bargain, but 17.Rd6 shows who is in charge.

Instead of castling, Black could start with 10...Bxd4, and after 11.Qxd4 0-0 12.0-0-0 the engines reckon 12...Qc7 is marginally better than 12...Qb6. Black will again get the piece back, but White has the upper hand, according to the engines.

Capturing on e4 in the diagram position is tempting, but as always one cannot just play by analogy - the specifics of a position have to be properly considered.

LESSON: 'similar' is not 'the same'.

Summing Up Kidlington

MY score of +1=1-2 in the U2000 (I also took a halfpoint bye) lost 29.4 ECF elo.

Kidlington U2000 Round Five

Spanton (1929) - Jeffrey Bryant (1745)

French Exchange

1.e4 e6 2.d4 d5 3.exd5 exd5 4.Nf3 Nc6 5.Bd3!?

The main move in ChessBase's 2026 Mega database is 5.Bb5!?

5...Nb4 6.Be2 Bd6

If 6...Bf5, then 7.Na3, with c3 to come.

7.a3 Nc6 8.c4!? dxc4 9.Bxc4 Nge7?

Sensible is 9...Nf6.

How should White proceed?

*****

*****

*****

*****

*****

10.Ng5

Crude, but effective.

10...0-0 11.Qh5

Black has no good way to defend both h7 and f7.

11...Bf5

Much worse is 11..h6?

12.Nxf7 Rxf7

Slightly less ruinous is the engines' 12...Bb4+!? 13.axb4 Qxd4, when White needs to find 14.Qe2! to maintain a large advantage.

13.Qxf7+ Kh8

White is the exchange and a pawn up, but is uncastled and well behind in development - indeed White has only one move that keeps a winning advantage

*****

*****

*****

*****

*****

14.Bg5

Winning is 14.Bh6!!, when 14...gxh6?? allows 15.Qf6#, while 14...Qf8 15.Qxf8+ Rxf8 16.Be3 sees White consolidate his material advantage.

14...Bg6 15.Bxe7 Bxf7

The engines prefer 15...Nxe7!?, claiming Black has full compensation for being down the exchange and a pawn.

16.Bxd8 Bxc4 17.Nd2 Bd3 18.Bxg5 Nxd4

Black has won back the pawn, and now 19.Nf3 Nc2+ 20.Kd2 Nxa1 21.Kxd3 Nb3 wins back the exchange, after which the engines reckon 22.Re1 gives White a slight edge.

19.0-0-0?? Ne2#

Kidlington U2000 Round Four

FACED a junior (born 2016).

Aviraj Bhaduri (1707) - Spanton (1929)

Sicilian Hyper-Accelerated Dragon

1.e4 c5 2.Nf3 g6 3.c3 d5!?

The main line in ChessBase's 2026 Mega database runs 3...Bg7 4.d4 cxd4 5.cxd4 d5 6.e5, with at least a slight edge for White, according to Stockfish17.1 and Dragon1.

4.exd5 Qxd5 5.Bb5+!?

This trails 5.Be2, 5.Na3 and especially 5.d4 in popularity.

5...Bd7 6.Bxd7+ Nxd7 7.d4 Ngf6 8.0-0 Bg7 9.Be3 cxd4 10.Qxd4 0-0

Now both sides have castled, how would you assess the position?

*****

*****

*****

*****

*****

Black has the only pawn on the two centre files, which could be a middlegame asset, while White has the possible endgame asset of a farside pawn-majority. The engines reckon the position is equal. 

11.Re1!?

They much prefer 11.Rd1 or 11.Na3.

11...Rfe8

The engines reckon 11...e5 12.Qxd5 Nxd5 gives Black the upper hand.

12.Na3 a6 13.Nc4 b5 14.Nb6 Qxd4 15.Bxd4 Nxb6 16.Bxb6 Nd5 17.Bd4 e6 18.Bxg7 Kxg7

How would you assess this late-middlegame position?

*****

*****

*****

*****

*****

Black has more space on the queenside, but the only other major difference between the two sides is the rival pawn-majorities. The engines agree the game is equal.

19.g3 Rec8 20.Rec1 Rc7 21.Rc2 Rac8 22.Rac1 Kf6

Not 22...Nb4 23.Rd2 Nxa2? 24.Ra1 Nxc3 25.bxc3 Rxc3 as, after 26.Kg2, White is well-placed to blockade, and ultimately destroy, Black's connected passers.

23.a3 h5 24.Kf1 g5!? 25.Ke2 g4!? 26.Nd2 Ke7 27.Kd3 f5 28.c4 bxc4+ 29.Rxc4 Rxc4 30.Rxc4 Rxc4 31.Nxc4

How would you assess this knight-and-pawn ending?

*****

*****

*****

*****

*****

White's queenside majority looks dangerous, but the engines reckon the position is completely equal.

31...f4?

But this is the wrong plan. Instead 31...Nf6, eyeing the e4 square, keeps White in check.

32.Ke4 f3 33.Ne3?

White is winning after 33.Ke5, according to the engines.

How should Black proceed?

*****

*****

*****

*****

*****

33...Nb6

The engines prefer a move I dismissed without bothering to analyse, namely 33...Nxe3!? I assumed 34.Kxe3 gave White a winning pawn ending. However, after 34...e5!? 35.Ke4 Ke6 36.b4 Kd6 37.a4 Ke6 38.b5 axb5 39.axb5 (39.a5?? loses to 39...Kd6) Kd6 40.b6 Kc6 41.Kxe5 Kxb7 42.Kf5 Kc5 43.Kg5 Kd4 44.Kxh5 Kd3 45.Kxg4 Ke2 46.h4 Kxf2 47.h5 Kg2 48.h6 f2 49.h7 f1=Q 50.h8=Q Qf3+, the game is drawn.

34.b3 Nd7!?

The text is for a long time strongly disliked by Dragon1, but the move seems to be fine. However, the engines are sure Black maintains complete equality with 34...Kd6, one point being 35.Kf4? loses to 35...Nd5+, eg 36.Nxd5 Kxd5! 37.Kg5 e5 38.Kxh5 e4 39.Kxg4 Kd4 40.h4 Kd3 41.h5 e3 etc.

35.b4 Nb6 36.Nd1!?

The engines for quite some time marginally prefer this over 36.Ke5, but come to view the two moves as equal.

36...Nc4 37.a4 Nb6

Going after the h2 pawn with 37...Nd2+? does not work, eg 38.Kf4 Nf1? 39.Kg5 Nxh2 40.Ne3!? Kd6 41.Kxh5. Slightly better is 38...Kf6, but 39.Ke3 Nf1 40.Kd4, with b5 to come, wins for White.

How should White proceed?

*****

*****

*****

*****

*****

38.Nc3

Interesting is 38.a5!? Nd5 39.b5 axb5 40.a6 Nc7 41.a7, but both 41...b4 and 41...Kd6 seem to hold.

The engines suggest 38.Kf4!? Kf6 (38...Nxa4? 39.Kg5 Nb6 40.Nc3! is very good for White, and if 40...Nd5 41.Nxd5+ exd5, White wins with 42.Kxh5 d4 43.Kxg4 etc), and now 39.Nc3, when 39...e5+ 40.Ke4 Ke6 transposes to the final position in the game, except with White, rather than Black, to move. The engines reckon White has to continue 41.a5!?, when Dragon1's 41...Nd7?! loses, according to Stcokfish17.1, but 41...Nc4 42.b5!? axb5 43.a6 b4!? 44.a7 Nb6 45.Na4 Na8 seems to hold.

38...Kd6 39.Kd4

39.Kf4 Nd5+ 40.Nxd5 exd5! is completely equal, according to the engines.

39...e5+ 40.Kd3 Ke6 41.Ke4 ½–½

Lessons From Mariánské Lázně V

AT every move it is vital to perform a simple blunder-check.

Mine takes the form of a mantra, Captures-Checks-Hangers-Threats, which is adapted from the teachings of first world correspondence champion Cecil Purdy.

In other words, I quickly examine every possible capture and every possible check, followed by looking for hangers, ie pieces and pawns that are not protected by other pieces are pawns, and finally I search for threats, eg moves that attack a queen or other valuable pieces,

More often than not, this takes less time to carry out than it takes to type (especially if, like me, you type with one finger on each hand).

Unfortunately, in the heat of battle, it is easy to forget to blunder-check.

That is what happened in the following position from my round-five game, where I had white against Russia's Sergei V Ivanov (2029).

Black has just avoided the loss of the exchange by playing 20...Rf6-f5

*****

*****

*****

*****

*****

White has a promising sacrifice in 21.Nxg7!?, which equalises, according to Dragon1, although Stockfish17.1 rates the move less highly.

Instead I played 21.g4?, almost without thought, as I had expected Black's last move, and had pre-planned my reply.

But if I had carried out my mantra, Captures-Checks-Hangers-Threats, I would surely have recognised the danger of my opponent replying 21...Rxh5!

After 22.gxh5 Qxh3 Black has a pawn, the bishop-pair and a strong attack - much more than enough for an exchange, according to Stockfish17.1 and Dragon1.

LESSON: forget to blunder-check at your peril!

Kidlington U2000 Round Three

AFTER taking a halfpoint bye in round two, I was upfloated.

Spanton (1929) - Matthew Stewart (1694)

Sicilian Bb5(+)

1.e4 c5 2.Nf3 d6 3.Bb5+ Bd7 4.Bxd7+ Qxd7 5.0-0 Nc6 6.c3 Nf6

Stockfish17.1 marginally prefers 6...d5!?, when the main line in ChessBase's 2026 Mega database runs 7.exd5 Qxd5 8.d4 e6 9.Be3 cxd4 10.Nxd4 Nf6, with what Stockfish17.1 and Dragon1 agree is an equal position.

After 6...Nf6, how should White proceed?

*****

*****

*****

*****

*****

7.Qe2

Most popular in Mega26 is 7.Re1, while the engines like the gambit move 7.d4!?, eg 7...Nxe4 8.d5 Ne5 9.Re1 Nxf3+ 10.Qxf3 Nf6, after which they reckon White has at least enough for a pawn.

7...g6

The engines prefer 7...e6 or 7...e5!?

8.d4 cxd4 9.cxd4 Bg7 10.Nc3 0-0

Now both sides have castled, how would you assess the position?

*****

*****

*****

*****

*****

White's classical centre gives a slight edge, according to the engines.

11.Rd1 e6

The engines suggest 11...d5!? 12.e5 Ne8.

12.Bf4!?

How should Black respond?

*****

*****

*****

*****

*****

12...Nh5!?

This comes to be Stockfish17.1's top choice, at least for a while, but the engines come to more-or-less settle on 12...h6!? or 12...Rad8, eg 12...h6!? 13.Nb5 a6!? 14.Nxd6!? Nh5 15.Qe3 g5 16.Be5 f6 17.Bg3 Nxg3 18.Nc4 (threatening a fork with Nb6) Nxe4!? 19.Qe4 Rad8, a long and complicated line that they reckon ends in equality.

But note that 12...Nxd4?? fails to 13.Rxd4.

13.Be3 a6?!

The engines suggest 13...Nf6, but then White has got in the move Be3 for free.

14.e5

Threatening to win a piece with 15.g4.

14...f5 15.exf6?!

This throws away most of White's advantage. The engines give 15.Na4 Rad8 16.Bg5 Rb8 17.Nb6 Qc7 18.Nc4, claiming White has a positionally won game.

15...Nxf6 16.d5 Nb4?

Better is 16...exd5 17.Nxd5 Nxd5 18.Rxd5 Ne7, although Black's isolani means White has an edge.

17.dxe6 Qxe6 18.Ng5 Qe7 19.Qc4+ 1-0

MS resigned almost instantly - was he right to do so?

*****

*****

*****

*****

*****

Black has 19...d5, when best play, according to the engines, runs 20.Bc5!? dxc4 21.Bxe7 (forking rook and knight, as well as capturing the queen) Nd3 22.Bxf8 Bxf8!?, after which Black is the exchange down, but not without chances, and, at least below elite level, I feel most people would play on.

Kidlington U2000 Round One

Philip Neatherway (1811) - Spanton (1929)

Spanish Four Knights

1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.Nc3 Bb4 5.d3 0-0!?

The main move is 5...d6.

How should White proceed?

*****

*****

*****

*****

*****

6.0-0

This is Stockfish17.1's top choice. Dragon1 prefers 6.Bxc6 bxc6 7.Nxe5, although it reckons 7...d5 8.0-0 Qe8 gives equality.

6...d6 7.Bg5 Bxc3

This is overwhelmingly the commonest move, but the engines marginally prefer 7...Ne7!? Keeping symmetry with 7...Bg4?! has been tried, but 8.Bxf6!? gxf6 9.Bxc6!? bxc6 gives White at least a slight edge, according to the engines.

8.bxc3 Qe7 9.Re1 h6 10.Be3!?

This may be a novelty. Known moves are 10.Bxf6, 10.Bd2, 10.Bc1!? and especially 10.Bh4.

10...Bg4 11.h3 Bh5 12.Qd2?

White's pieces are not well-placed to protect a shattered kingside.

12...Bxf3 13.gxf3 Kh7 14.Kh2 Nh5 15.f4?

The engines much prefer 15.d4.

15...exf4 16.Bxf4 Ne5 17.Qe3?!

The engines reckon White should offer the exchange with 17.Kh1!?, or protect the f3 square with 17.Kg2.

17...f5

Even stronger is 17...Qf6.

18.d4

Black to play and win

*****

*****

*****

*****

*****

18...Nxf4?

Winning is 18...Ng4! 19.hxg4 Qh4+, eg 20.Qh3 Qxf2+ 21.Qg2 Qh4+!? 22.Qh3 Qxh3+ 23.Kxh3 fxg4+ 24.Kxg4 Rxf4+ 25.Kh3 (Kxh5?? g6#) Rf3+ 26.Kh2 (26.Kg4?? Raf8 etc) Rxc3, netting two pawns, or 20.Kg1 Qxg4+ 21.Kf1 Nxf4, when Black is a pawn up, with a continuing attack. The engines reckon 18...a6!? is also better than the text.

19.Qxf4 Ng6

Not 19...fxe4?? 20.Qxe4+ and 21.dxe5.

20.Qg3?

The game is equal after 20.exf5 Nxf4 21.Rxe7, according to the engines.

Black to play and win

*****

*****

*****

*****

*****

20...f4?

Winning is 20...fxe4, eg 21.Bd3 d5 22.c4 Qf6!? with a large advantage.

21.Qg4 Nh4!? 22.e5 d5?

It was time to go for king safety with 22...Qg5, although 23.Bd3+ Kh8 24.Qxg5 Nf3+! 25.Kh1 hxg5 leaves White with at least a slight edge, according to the engines.

23.Bd3+ Kh8 24.Rg1 Rae8?

Too slow, but the engines' 24...c5 25.Qh5 c4 26.Bg6 is also very good for White.

25.Qh5 f3

Black's game is past saving, eg 25...Rf7 26.e6, or 25...Qd8 26.Rxg7! Nf3+ 27.Kh1!

26.Rg4 Ng2 27.Qg6 Kg8 28.Qh7+ Kf7 29.Rxg7+ 1-0

Lessons From Mariánské Lázně IV

CERTAIN words in chess have a technical meaning separate from their usage by the general public.

An obvious example of this is in the common pub-quiz question: "How many pieces are there on the board at the start of a game of chess?"

Most people, assuming they had at least a passing acquaintance with the game, would answer 32.

But an experienced chess player would probably pause, knowing that in chess the word pieces has a technical meaning that includes rooks, knights, bishops, kings and queens, but excludes pawns.

Similarly, 'good' and 'bad' has a technical meaning when applied to bishops.

A bad bishop, in this sense of the word, is one on the same colour complex as many of its own side's pawns, especially centre pawns, while a good bishop is one unobstructed in this way.

The importance of whether a bishop is technically good or bad becomes paramount in endings.

In the middlegame, however, what counts for more is whether a bishop is active or passive.

All of which is a long-winded introduction to my round-four game against Germany's Achim Heller (2038), in which the following position was reached after seven moves.

I have just played 7...Bf8-e7

*****

*****

*****

*****

*****

Black's dark-square bishop is technically good - Black's centre pawns are on light squares - but passive.

Black's light-square bishop is technically bad, but is active as it helps control the e4 square, reducing the chance of White making the pawn-break e4, and being well-placed to support the black king's knight should it ever go to e4.

White's light-square bishop is both good and active, while White's dark-square bishop is unmoved and technically bad.

It is also currently passive, but it can be made active by developing it to b2, and, indeed, 8.Bb2 is overwhelmingly the commonest move in the position (grandmasters have also played 8.Nc3 and 8.cxd5) as on b2 it helps prevent the pawn-break ...e5, and can support a white knight landing on the e5 square.

My opponent came up with what is probably a novelty in 8.Ba3!?

This is not an out-and-out mistake, but I think it is significant that after 8...0-0 Dragon1 suggests playing 9.Bb2!?, and Stockfish17.1 fluctuates between that move and the game's 9.Bxe7.

LESSON: whether a bishop is good or bad can be of decisive importance in an ending, but in the middlegame (and the opening) what usually counts is whether a bishop is active or passive.

Quick Weekender

ARRIVED in Kidlington, five miles north of Oxford, after making a late decision to play in a five-round weekender that starts tomorrow.

It is being played at a venue, the Leonardo Royal Hotel, that was apparently recommended by the 4NCL, and is a short walk from Oxford Parkway train station.

There are 179 entries, divided into four sections, with a time control of 90 minutes and a 15-second increment.

I have entered the U2000, where I am seeded fifth. However, I have booked a halfpoint bye for Saturday afternoon, so by the evening will be at least a halfpoint behind the leaders ...