Bobby Fischer v The Sicilian (part four)

Sicilian Accelerated & Hyper-Accelerated Dragon

Fischer treated these two variations in the same way, so generally they merged into one.

Specifically he usually rejected the chance to set up a Maróczy Bind and instead continued with his normal anti-Sicilian strategy of developing the white light-square bishop to c4.

The starting tabiya of these two merged dragons comes after move-orders that typically run 1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 g6 5.Nc3 Bg7 6.Be3 Nf6 7.Bc4 and 1.e4 c5 2.Nf3 g6 3.d4 cxd4 4.Nxd4 Nc6 5.Nc3 Bg7 6.Be3 Nf6 7.Bc4.

Starting tabiya of the merged Accelerated & Hyper-Accelerated Dragon

Black attacks the bishop with 7...Na5?!

Fischer - Norbert Leopoldi

Western Open (Bay City, Michigan) 1963

8.Bb3!?

Two years earlier Fischer had preferred 8.Be2, which has also been played by Keres, Stein, Nakamura and Anand. The text has been played by Suetin and Carlsen.

It may be significant that none of these great names chose the line favoured by my main analysis engines, Stockfish12 and Komodo11.01, namely 8.Bxf7+!? Kxf7 9.e5, one point being that retreating the attacked knight by 9...Ne8? or 9...Ng8 loses to 10.Ne6! A computer rapid game between Spaghetti Chess and Komodo1 in 2008 continued 9...Nc4 10.exf6 Bxf6 11.Qf3 Kg7 12.0-0-0 d6, reaching a position my modern engines reckon favours White, although the game was drawn in 36 moves.

8...Nxb3 9.axb3

As we have seen earlier in this series, Fischer prefers to capture on b3 with a pawn rather than withdrawing his aggressively placed king's knight.

9...d6 10.f3 Bd7 11.g4!?

Later games, albeit between weaker players, featured 11.Qd2.

11...a6

Leopoldi is understandably reluctant to castle into Fisher's kingside attack.

12.h4 h6

My engines strongly dislike 12...0-0?!

13.Qd2 Rc8 14.h5 e5

The engines prefer 14...g5, but like White.

15.Nde2 gxh5 16.gxh5 Be6

The engines give 16...Bf8, but basically they cannot suggest anything good for Black.

17.0-0-0 Rc6 18.Kb1 b5 19.Nd5 a5 20.Rhg1!

A pawn sacrifice that was hardly needed to confirm White's advantage, but it is the best move in the position, according to my engines.

20...Nxh5 21.Nec3 b4 22.Nb5 Bf8 23.Na7 Ra6 24.Qd3

The black queen's rook is trapped.

24...Bc8 25.Nxc8 Qxc8 26.Qxa6!?

A simple-enough combination - once you have seen it! - but by now there were several ways to win.

26...Qxa6 27.Nc7+ Kd7 28.Nxa6 Kc6 29.Rd5 1-0

Black commits to castling short with 7...0-0

This is the most-popular move in the position in ChessBase's 2020 Mega database, and is generally regarded as the sharpest.

Fischer replied 8.f3!? the first time he reached this position, but subsequently always played 8.Bb3, which is considered by most authorities to be most accurate.

A)

Fischer - Samuel Reshevsky

US Championship (New York) 1958

8...Na5?

While ...Na5 is regarded as dubious at move seven, it is a definite mistake one move later.

9.e5 Ne8

Or 9...Nab3 10.exf6 Nxa1 11.fxg7 Nxc2 12.Qxc2 with a large development advantage for White in Georgy Bastrikov - Leonid Shamkovich, Russian U18 Championship (Sochi) 1958 (1-0, 43 moves).

10.Bxf7+! Kxf7 11.Ne6

This is also the answer to 10...Rxf7.

11...dxe6 12.Qxd8 (1-0, 42 moves).

B)

Fischer - Reshevsky

Match, Game 4 (Los Angeles) 1958

8...Ng4!? 9.Qxg4 Nxd4 10.Qd1 Nxb3 11.axb3 b6!?

This is easily the most-popular move, but Fischer's reply makes it look strange. The main alternative is 11...a6, but the engines reckon White's lead in development is enough for an edge after 12.Qd2, eg 12...d6 13.Bh6 Bxh6 14.Qxh6 with a kingside attack.

12.Qd5

Also strong is the more-popular 12.Bd4.

12...Bxc3+ 13.bxc3 Qc7! 14.0-0-0

Not 14.Qxa8? Qxc3+ 15.Kd1 (15.Ke2 Ba6+ etc) Qxa1+ 16.Bc1 e6, when Black is a pawn up and has the safer king.

14...Qxc3 15.Bd4 Qc6 16.Qe5 f6 17.Qxe7 Bb7 18.f3

White is much better, according to the engines, but Reshevsky saved the game in an opposite-coloured bishops ending.

C)

8...a5

This, the main choice of dedicated Accelerated Dragon fans, was not faced by Fischer.

The main line in Mega20 runs 9.0-0 a4!? 10.Nxa4 Nxe4 11.Nb5 Ra6 12.c4 d6 13.Qe2 Nf6 14.h3!? with the better position for White, according to my engines.

D)

8...d6

This is actually Black's main move in Mega20, but again was not faced by Fischer.

The main line runs 9.f3 Bd7 10.Qd2 Rc8 11.0-0-0, reaching an opposite-sides castling position that the engines reckon is good for White. However, there are more 7,600 games with this position in Mega20 and we are very much still in theory.

Black develops the queen with 7...Qa5

This arguably more-positional approach has been recommended by grandmaster Nigel Davies in a ChessBase DVD.

Fischer - Friðrik Ólafsson

Bled (Yugoslavia) 1961

8.0-0 d6!? 9.Nb3!? Qc7 10.Be2 0-0

This position is normally reached via the move-order 8...0-0 9.Nb3!? Qc7 10.Be2 d6. A major difference is that in the latter move-order White more often plays 9.Bb3. The point of Ólafsson's move-order may have been that in it 9.Bb3 can be met by 9...Ng4!?, which the engines reckon gives equality.

11.f4 a5 12.a4 Nb4!?

This is the most-common move in Mega20. Black has latent pressure against c2 and, in some lines, may get in ...d5.

13.Rf2!?

More popular today is 13.Bf3 but Fischer presumably judged that he did not yet need to impede ...d5.

13...e5 14.Bf3

This would probably have also been the answer to 13...Rd8, one point being that it allows White to play Rd2.

14...Bd7 15.Rd2 Rfd8! 16.Kh1

16.Rxd6?! is met by 16...Ba4 or 16....Ne8 17.Rd2 exf4, both of which give Black equality, according to the engines.

16...Bc6 17.Qg1!

Fischer decides the backward d pawn is defendable from immediate direct attack and instead latches on to the weakness at b6.

17...Nd7 18.f5 b6 19.Rad1 Nc5 20.Nb5 Qe7 21.Nxd6 Nxc2 22.Nxc5 Nxe3?!

Better, according to the engines, is 22...bxc5 23.Rxc2 Rxd6, but they still prefer White.

23.Qxe3 bxc5 24.Be2!

The engines point out that 24.b3 can be met by 24...Bf8 25.Qxc5 Bxa4! 26.bxa4 Ra6 when Black will regain his piece with a reasonable game.

24...Bxa4

Black has to accept the pawn sac as, for example, 24...Bf8? now runs into 25.Bc4 Be8 26.fxg6 hxg6 27.Qg3 with a big attack.

25.b3 Be8 26.Bc4 a4 27.Bd5! Rxd6!?

Komodo11.01 agrees this exchange sac is Black's best chance, but Stockfish12 gives 27...Ra7 28.fxg6 hxg6 29.Nxe8 Rxe8, albeit much preferring White.

28.Bxa8 Rd4 29.fxg6 hxg6 30.bxa4 Bxa4 31.Ra1 Qf8?

White is only slightly better after the largely forcing-line 31...Qd8 32.Bd5 Bc6 33.Rxd4 exd4 34.Bxf7+ Kxf7 35.Qb3+ Ke7 36.Ra7+ Kd6 37.Rxg7 Bxe4. Indeed Stockfish12, but not Komodo11.01, reckons Black's bishop and pawn give full compensation for the exchange despite the exposed position of the black king.

32.Bd5! Bh6 33.Rad4!

Fischer gives up his queen, having seen the attack on f7 means White is winning.

33...Bxe3 34.Rdxa4 Qh6 35.Rf1 Bf4 36.g3 Qh3 37.Raa1 Bxg3?

Black loses trivially after this, but is lost anyway.

38.Ra8+ 1-0

Chess Evolution: QGD Exchange (part two)

QUEENSIDE castling by White in the Exchange Variation of the Queen's Gambit Declined was not a popular option in the early days of 1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.cxd5.
It quickly became established that the lines to be feared by Black were those in which White went for a Minority Attack.

One of the earliest and most successful exponents of this was former child-prodigy Samuel Reshevsky.

Reshevksy - George Treysman

US Biennial (New York) 1936

1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.cxd5 exd5 5.Bg5

It took a little while for this to become the standard move - 5.Nf3 also being popular - but today it is almost automatic. The bishop puts pressure on Black, practically ensuring that ...c6 will become necessary at some point. Meanwhile there seems no way for Black to take advantage, as can happen in other openings, of White unprotecting b2.

5...Nbd7!?

Today Black usually plays 5...Be7, but ...Nbd7 nearly always follows at some point, even though it blocks the black light-square bishop. One reason for playing ...Nbd7 is to set a trap that has claimed 527 victims in ChessBase's 2020 Mega database.

White to play and lose

*****

*****

*****

*****

*****

6.e3

The losing line runs 6.Nxd5? Nxd5 (45 players in Mega20 failed to play this) 7.Bxd8 Bb4+ 8.Qd2 Bxd2+ 9.Kxd2 Kxd8, after which Black is up a knight for a pawn, as in J Foster (no grade) - Spanton (151 BCF), Highbury (London) Rapidplay 1989 (0-1, 28 moves) and many other games.

6...Be7 7.Qc2 c6 8.Bd3 0-0 9.Nf3 Re8
More than 10,000 games in Mega20 reach this standard position by various move-orders. White scores +4,658=4,013-2,134, which works out at a very healthy 62%.

10.h3!?

This move first appears in Mega20 in a 1928 win of Sämisch's against Berthold Koch. Its ideas include taking the g4 square away from Black's pieces, particularly the king's knight and queen's bishop, and leaving open the option of castling long.

10...Nf8

This opens a diagonal for Black's light-square bishop and unties the king's knight from protecting h7.

11.Bf4!?

This move, which is the most popular in the position (at least in Mega20), seems to have been first played by Reshevsky in a 1934 win against Mario Monticelli. The arguably more-natural 11.0-0 occurred in Sämisch - Koch, Berliner SG 100th Anniversary (Berlin) 1928. That game continued 11...Ne4 (a standard way for Black to seek simplifying exchanges) 12.Bf4 Nxc3 13.bxc3, when the analysis engines Stockfish12 and Komodo11.01 give White a slight edge (1-0, 51 moves).

11...Bd6 12.Bxd6 Qxd6

Black has simplified to a certain extent, but in doing so he has given up his good bishop.

13.0-0

Overwhelmingly more popular in Mega20 is 13.0-0-0!? (229 examples compared with 75 for the text). Reshevsky preferred short castling all seven times he reached the position after 12...Qxd6.

13...Be6

Other moves faced by Reshevsky were ...g6, ...Qe7, ...Ng6 and ...b6. Overall he scored +5=2-0. For what it is worth, the text is Stockfish12's choice, while Komodo11.01 prefers 13...Qe7.

14.Rfb1!?

Making it starkly clear a Minority Attack is on the cards. White's plan is to advance his a and b pawns, seeking pawn-exchanges that will leave Black with a weak pawn or pawns.

14...Re7?!

Treysman has played standard moves up to this point, but the text, makes it difficult to coordinate the black pieces. Black often seeks kingside counterplay in the QGD Exchange, but here there does not seem a promising way of achieving that. Stockfish12 suggests 14...a5, which is normal in such positions. The idea is that when the white b pawn advances to b4, after a preparatory a3, Black plays ...axb4 so as not to later end up with an isolated a pawn, as happens in the game.

15.b4 Rc7 16.Qd2!? Qe7 17.Qc2!?

This was almost certainly not a tacit draw-offer by Reshevsky. I suspect he had come to the conclusion that at move 16 he could he could have successfully played a4, or even an immediate b5.

17...Ne8

The engines prefer this to repeating with 17...Qd6.

18.a4 Nd6

If Black can engineer an exchange of light-square bishops, this knight could take up a useful outpost at c4 or e4, but that will not be easy.

19.Rc1!?

Reshevsky decides, and the engines agree, that the king's rook is best-placed on the c file. This is why, going back to the position after 13...Be6, it is more usual for White to play 14.Rab1, with Rfc1 to follow. Karpov has played the immediate 14.Rfc1.

It should be remembered the Minority Attack was still in its infancy in 1936, so allowances must be made. Even so, the overall impression of the game, at least to me, is strikingly modern.

19...Rac8 20.Qb2 Ng6 21.b5

Reshevsky's dithering with his queen has given Black time to prepare his defences, but I think it fair to say White's position is easier to play.
21...Qf6?!

The engines prefer 21...c5!?, when 22.dxc5 Rxc5 leaves Black with one weak pawn (d5).

22.bxc6 bxc6

Possibly better is 22...Rxc6!?, although then d5 is weak and White has pressure against the black a and b pawns.

23.Ba6 Rd8 24.Nb5! Nxb5 25.axb5 Rb8?!

The engines give 25...c5 26.b6! Rb8 27.bxa7! Rxb2 28.a8=Q+ Nf8 29.Rxc5 Rxc5 30.dxc5 Rc2 31.Rb1, when 31...Rxc5 restores material equality, but White has the upper hand after 32.Rb8.

26.Qa3 Bf5 27.Rc5 Qd6 28.Qa5 Be6?!

Best, according to the engines, is 28...Ne7, but Black remains under intense pressure.

29.Rac1 Rb6 30.bxc6 Rbxc6 31.Bb7 Rxc5 32.dxc5 Qe7 33.c6 Bc8 34.Bxc8 Rxc8 35.Qxd5

Not only has Black lost a pawn, but he has a weak a pawn and is hardly in position to cope with White's passer.

35...Qe6 36.Qc5 Rc7 37.Nd4 Qe7 38.Nb5 1-0

Chess Evolution: The French Defence (part 11)

AS international chess returned to normal, and indeed increased, after World War One, there was a big question mark over the Winawer: 1.e4 e6 2.d4 d5 3.Nc3 Bb4.
The statistics bear this out. Of the games with 1.e4 e6 2.d4 d5 3.Nc3 from 1919-27 in ChessBase's 2020 Mega database, 358 continued with the Classical: 3...Nf6 compared with just 105 for 3...Bb4.
The main reason for the relative unpopularity of 3...Bb4 was 4.exd5, which was played in 67 of the 105 games.
It was thought that switching into a line of the Exchange Variation: 1.e4 e6 2.d4 d5 3.exd5 was detrimental to Black because in that variation the black dark-square bishop was believed to be best-placed on d6.
Then came game one of the Capablanca - Alekhine world championship, in which Capablanca played 4.exd5 against Alekhine's Winawer and lost.
From 1928-39, when World War Two brought another cessation to most international chess, 3...Nf6 featured in 631 Mega20 games compared with 262 for 3...Bb4.
In other words the percentage of Winawer games, compared with Classical, rose from 22.7% to 29.3%.
That was significant, but even more dramatic was the eclipse of 4.exd5 as the preferred answer to the Winawer.
It was played 63.8% of the time from 1919-27 but just 16.8% from 1928-39.
The new kid on the block - actually, not so new but suddenly much more popular - was 4.e5, which accounted for 55% of the games from 1928-39.
The main champion of the Winawer became Mikhail Botvinnik. He played it throughout the 1930s, continued (when he could) during World War Two and kept faith with it in the 1948 tournament to decide who should replace the deceased Alekhine as world champion.
Samuel Reshevsky - Mikhail Botvinnik
1948 World Championship Round 19
French Winawer
1.d4 e6 2.e4 d5 3.Nc3 Bb4 4.e5 c5 5.a3
This is now firmly established as the main move but for a long time it played second fiddle to 5.Bd2.
5...Bxc3+!?
Botvinnik won with 5...Ba5 against the same opponent two years earlier at Moscow in a match between the Soviet Union and the US.
6.bxc3 Qc7 7.Qg4!?
Reshevsky was noted for his positional play, while also excelling at tactics. At the time this game was played, the main move was the non-forcing 7.Nf3, but here Reshevsky aims to take advantage of Black's vulnerability on the kingside dark squares.
7...f5!?
A major alternative is 7...Ne7, transposing to a position normally reached via the move-order 6...Ne7 7.Qg4 Qc7.
8.Qg3!?
The first game in Mega20 to feature this move, which is now standard.
8...cxd4 9.cxd4 Ne7 10.Bd2
Not 10.Qxg7?? Rg8 and 11...Qc3+ etc.
10...0-0 11.Bd3
Defending c2 and hoping to later lever open the b1-h7 diagonal.
In many lines of the Winawer with Qg4, Black sacrifices one or two kingside pawns in return for dynamic piece-play on the queenside and in the centre. Here material is level but positionally unbalanced - White has more space and kingside attacking chances, while Black has pressure down the half-open c file and generally the better of it on the queenside.
White also has the bishop-pair, although the blocked centre means that is not an advantage at present. Nevertheless Black has a bad bishop, so Botvinnik sets about exchanging it.
11...b6 12.Ne2 Ba6 13.Nf4 Qd7 14.Bxa6!?
The point of voluntarily exchanging is that the black queen's knight has few immediate prospects on a6.
14...Nxa6 15.Qd3 Nb8 16.h4 Nbc6 17.Rh3 Rac8 18.Rg3 Kh8
Kasparov (2795) - Short (2690), Novgorod (USSR) 1997, saw 18...Rf7 19.h5 Nd8 20.c3* with an edge for White, according to the analysis engines Stockfish11 and Komodo11.01 (1-0, 31 moves).
19.h5 Rf7 20.h6 g6 21.Rc1!?

A rule-of-thumb from Cecil Purdy, the first world correspondence champion, was to never put a rook behind an unmoved pawn, even if you intend to move the pawn fairly quickly. Such rules are often broken, correctly, but here the engines much prefer 21.a4 or 21.Kf1!?

21...Rff8!? 22.Nd2

White's kingside pressure has not resulted in a breakthrough, but he still has an edge, according to the engines. However, Black now has time to try to work up an initiative on the queenside.

22...Nb8!?

Opening a line for the queen's rook, but the engines are unimpressed, preferring 22...Nd8 with the idea of proceeding with ...Nf7!?

23.Kf1 Rc4

A rook in the middle of the board in a crowded middlegame can easily get into trouble, but it will be hard for the white knight to harass it.

24.Kg1 Nbc6!? 25.Bg5 Ng8 26.Re1 Qf7 27.c3 Na5

Botvinnik decides the knight is a better occupant of the c4 outpost.

28.Nf4 Rc6

Defending e6 as well as clearing the c4 square and possibly preparing to add pressure down the c file with ...Rfc8.

29.Bf6+?

A serious misjudgment. The engines reckon White keeps an edge - "winning," according to Stockfish11, "slight," according to Komodo11 - with a move such as 29.a4.

29...Nxf6 30.exf6

How should Black proceed?

30...Nc4

Did Reshevsky miss this simple zwischenzug? White is massively better after 30...Qxf6? 31.Rge3, but 30...Nc4 scuppers Reshevsky's plans.

31.Qb1 Qxf6

The net result of 29.Bf6+? is that White has lost a pawn and exchanged an active bishop for an inactive knight.

32.a4 g5

Botvinnik seizes the initiative on the kingside while the white pieces are uncoordinated.

33.Nd3 f4 34.Rh3 g4 35.Rh1 Rc7 36.Qd1 Qg6 37.Rh4 f3 38.g3

Reshevsky has been able to close lines on the kingside, but his forces remain uncoordinated. His one plus is pressure against the weak e pawn.

38...Rcf7?

Abandoning the c file lets White off the hook. The engines reckon Black keeps an edge with 38...Rfc8 or 38...Rf6.

39.Nf4?

Walking into a simple and very strong exchange sac. The position seems equal after 39.Ne5 Nxe5 40.Rxe5, White's giving up the c file meaning he cannot reply 40...Rxc3.

39...Rxf4! 40.gxf4 Rxf4

Black has two pawns and a powerful attack for the exchange.

41.Qb1

This is best, according to the engines, but White is already busted.

41...Rf5 42.Qd3

The engines give 42.Re4 or 42.Kh2 as best but in each case a simple answer is 42...Nd2.

42...g3 43.Qf1 gxf2+ 44.Kxf2 Rg5 45.Qh3 Rg2+ 46.Kf3 Nd2+ 47.Ke3 Rg3+ 0-1
*The move-count was one less in Kasparov-Short, but the game reached the same position as in Reshevsky-Botvinnik.