Garry Kasparov's Forgotten Weapon Against The Queen's Gambit (part 15)

Karpov (2700) - Kasparov (2710)

World Championship (Moscow) 1984 Game 9

QGD Tarrasch

1.d4 d5 2.c4 e6 3.Nf3 c5 4.cxd4 exd5 5.g3 Nf6 6.Bg2 Be7 7.0-0 0-0 8.Nc3 Nc6 9.Bg5 cxd4 10.Nxd4 h6 11.Be3 Re8 12.Qb3 Na5 13.Qc2 Bg4 14.Nf5 Rc8 15.Bd4!?

Karpov is the first to deviate from game seven of the match, in which he gained the bishop-pair with 15.Nxe7+.

15...Bc5 16.Bxc5

Probably not 16.Bxf6?! Qxf6 17.Ne3 (definitely not 17.Nh4? Bxe2) Bxe3 with an edge for Black.

16...Rxc5 17.Ne3

Position after 17.Ne3

*****

*****

*****

*****

*****

17...Be6

17...d4!? is the thrust both players must look out for in the Tarrasch. The analysis engines Stockfish12 and Komodo11.01 continue 18.Rad1 Nc4 19.Nxg4 Nxg4 20.Nd5 (they also like 20.Qb3) Rxd5 21.Qxc4 Rxd7 22.Qa4 with what they reckon is a slight edge for White.

18.Rad1

18.b4 Rc8 19.bxa5 d4 is fine for Black.

18...Qc8 19.Qa4 Rd8 20.Rd3 a6 21.Rfd1

The engines reckon White can safely win the d pawn with 21.Ncxd5 Nxd6 22.Bxd5 Bxd5 23.Qd1!, which suggests Black has already gone wrong (perhaps at move 17).

21...Nc4 22.Nxc4 Rxc4 23.Qa5 Rc5 24.Qb6 Rd7 25.Rd4 Qc7 26.Qxc7 Rdxc7

We have reached a queenless middlegame that seems slightly better for White, who is more active.

27.h3 h5 28.a3 g6 29.e3 Kg7 30.Kh2 Rc4 31.Bf3

This will allow the king to protect f2.

31...b5 32.Kg2 R7c5 33.Rxc4

Not 33.Nxd5?? Nxd5 34.Bxd5 Rxd4, when Black emerges up a bishop for a pawn.

33...Rxc4

33...dxc4?! 34.Rd6 a5 (34...Bc8? 35.Rxf6!) 35.Rb6.

34.Rd4 Kf8

White is better after 34...Rxd4 35.exd4, but by how much? Komodo11.01 reckons the edge is small; Stockfish12 at first gives White the upper hand, but gradually comes nearer Komodo11.01's evaluation.

35.Be2

Karpov evidently thinks it is worth spending a tempo to encourage an exchange of rooks.

35...Rxd4 36.exd4

The ending is better for White because his bishop has targets while Black's is a purely defensive piece, apart from hitting h3, which White can easily remedy.

36...Ke7 37.Na2 Bc8 38.Nb4

Black's queenside pawns cannot move, so there is no need to fix them with 38.b4?!, which is probably a mistake anyway as it turns c4 into a possible outpost square for the black knight which, after 38...Ne4, is also a menace to the a3 pawn.

38...Kd6 39.f3 Ng8 40.h4 Nh6 41.Kf2 Nf5 42.Nc2 f6 43.Bd3 g5

43...Ne7!? intends 44...Bf5 but White can stop this with 44.Ne3 as 44...Bf5? 45.Nxf5+ Nxf5 46.Bxf5 gxf5 47.Ke3 is a winning pawn-ending for White. So after 44.Ne3 Black has to play something like 44...Ke6 and await developments. The engines prefer the text.

44.Bxf5 Bxf5 45.Ne3

The knight is slightly better than the bishop in this ending because the bishop has no real targets, apart from the f3 pawn, and Black has to be wary of White infiltrating on the dark squares.

45...Bb1 46.b4!?

Fixing Black's queenside pawns, but ruling out a queenside invasion by the white king.

46...gxh4?

Black should temporise with 46...Ke6 or a similar non-committing move.

47.Ng2!

Komodo11.01 very slightly prefers 47.gxh4, but it is far from certain White can break through after that.

47...hxg3+ 48.Kxg3 Ke6

Black cannot hold the h pawn, so this is not a pawn-losing blunder.

49.Nf4+ Kf5 50.Nxh5 Ke6 51.Nf4+ Kd6 52.Kg4 Bc2 53.Kh5 Bd1 54.Kg6!

There is no need to save the f3 pawn, eg 54...Bxf3 55.Kxf6 Bg4 56.Nd3 Bh3 57.Nc5 Bc8, after which the bishop cannot move, so 58.Kf7 Kc6 59.Ke7 Kc7 60.Ne6+ Kc6 61.Kf6 Kd6 62.Nc5 Kc6 63.Ke5 etc. This is a long line, but both players would have realised Black eventually arrives in zugzwang after 54...Bxf3.

54...Ke7! 55.Nxd5+?!

This natural-looking capture may be premature. The engines prefer 55.Nh5!?, eg 55...Bxf3 56.Nxf6 Be4+ 57.Kg5 Bc2 58.Ng4 Be4 59.Ne5 Kd6 60.Kf6 Bg2 61.Nf7+ Kd7 62.Ke5 etc.

55...Ke6 56.Nc7+ Kd7?

Better is 56...Kd6, keeping the king closer to the d pawn, eg 57.Nxa6 Bxf3 58.Kxf6 Kd5 59.Nc7+ Kxd4 is a database draw (despite Komodo11.01 giving White the upper hand). White should therefore try 57.Ne8+ Ke7 58.Nxf6 Bxf3 59.Kf5, but after 59...Kd6 it is hard to see how White makes progress (Stockfish12 agrees, but Komodo11.01 again gives White the upper hand).

57.Nxa6 Bxf3 58.Kxf6 Kd6

Black is a tempo down on the first line in the previous note.

59.Kf5 Kd5 60.Kf4 Bh1 61.Ke3

White's king is in time to save the d pawn.

61...Kc4 62.Nc5 Bc6 63.Nd3 Bg2 64.Ne5+ Kc3 65.Ng6 Kc4 66.Ne7 Bb7?!

The bishop is more active if it stays on the kingside with 66...Bh1! Then if, as in the game, 67.Nf5, Black can play the desirable 67...Kd5 as White does not have 68.Kd3??, which loses to 68...Be4+. It therefore seems Black has a draw - at least the engines and I have not been able to find a win for White.

67.Nf5 Bg2

Here 67...Kd5 is simply met by 68.Kd3, after which the black king will eventually be forced back, eg 68...Ke6 69.Ng7+ Kd5 70.Ne8 Bc6 71.Nf6 Ke6 72.Ne4 Kd5 73.Nc3+ Kd6 74.d5! Bxd5 75.Nxb5+.

68.Nd6+ Kb3 69.Nxb5 Ka4 70.Nd6 1-0

Garry Kasparov's Forgotten Weapon Against The Queen's Gambit (part 14)

KASPAROV challenged Anatoly Karpov for the world title in Moscow in 1984.

The champion won game three on the white side of a Sicilian and game six on the black side of a Queen's Indian.

Then came game seven:

Karpov (2700) - Kasparov (2710)

QGD Tarrasch

1.d4 d5 2.c4 e6 3.Nf3!?

Karpov avoids 3.Nc3. This was not normal for him. In his career with the white pieces, after the moves 1.d4 d5 2.c4 e6, he played 3.Nc3 twice as often as 3.Nf3, and scored a much higher percentage with it (80% for 3.Nc3, 67% for 3.Nf3, according to ChessBase's 2021 Mega database). I have speculated earlier in this series on possible reasons for choosing 3.Nf3, but only Karpov knows for sure why he played this way.

3...c5 4.cxd5 exd5 5.g3 Nf6 6.Bg2 Be7 7.0-0 0-0 8.Nc3 Nc6 9.Bg5 cxd4 10.Nxd4 h6 11.Be3 Re8 12.Qb3!?

This seems to have been first played by Viktor Korchnoi in a 1970 win over Borislav Ivkov. The queen adds immediate pressure against d5, frees d1 for a rook and pressurises b7.

12...Na5 13.Qc2

Position after 13.Qc2

This position was also reached in the Beliavsky-Kasparov quarter-final match, except there the queen went to c2 on move 12, meaning the black queen's knight was still at c6. Is the knight at a5 an improvement for White or for Black? I guess it depends on the coming moves. If the knight finds useful employment at c4, Black has gained by the manoeuvre. But if the knight has to fall back to c6, Black has lost time.

13...Bg4

The immediate 13...Nc4 has been played by grandmasters, but seems well met by 14.Bf4.

14.Nf5

Previous games had mostly seen 14.h3, until Lajos Portisch played the text earlier in 1984, beating Michael Wilder at New York and drawing with Murray Chandler at Amsterdam. 14.Nf5 is now the main move.

14...Rc8!?

A novelty. Wilder and Chandler preferred 14...Bb4, which is more popular in Mega21.

15.Nxe7+

15.Bxa7?! b6 is problematic for White.

15...Rxe7 16.Rad1 Qe8

16...Be6?! is the type of defensive retreat Black wants to avoid.

17.h3 Bh5

It is too late for 16...Be6? as now 17.Bxa7 works, so Kasparov keeps the light-square bishop pressurising e2.

18.Bxd5!?

This may be too greedy. My main analysis engines Stockfish12 and Komodo11.01 reckon White should increase the tension with 18.Bd4, eg 18...Ne4 19.Rfe1 Bg6 20.Qc1 Rd7, when Stockfish12 gives White the upper hand after 21.Qf4, but Komodo11.01 rates 21...Nc6 as an equaliser. For practical purposes, the position after move 17 should probably be rated as unclear.

18...Bg6 19.Qc1 Nxd5 20.Rxd5 Nc4 21.Bd4

Almost certainly not 21.Bxa7?! b6 22.b3 Ne3! 23.fxe3 Rxa7, when White is two pawns up but White's multiple weaknesses and Black's more-active pieces give the latter an edge, according to the engines.

21...Rec7 22.b3 Nb6 23.Re5 Qd7 24.Qe3 f6 25.Rc5 Rxc5 26.Bxc5 Qxh3

Black has regained his pawn, and the game is equal, according to the engines.

27.Rd1 h5

The engines prefer 27...Be8!?, which both covers the d7 entry point and plans redeployment at c6.

28.Rd4 Nd7!? 29.Bd6

The pawn-grab 29.Bxa7 is probably playable but looks risky after 29...Ne5.

29...Bf7 30.Nd5 Bxd5 31.Rxd5 a6 32.Bf4 Nf8 33.Qd3 Qg4?!

Consistent with previous play are the engines' suggestions of 33...h4 and 33...Ne6.

34.f3

It seems Karpov could have gained an advantage with the engines' 34.Rd6!, the point being to threaten 34.Qd5+, forking king and b7 pawn. The engines reckon best play then goes 34...Qg6 35.Qf3 Qf7, when White has an initiative but there does not seem to be anything immediately decisive.

34...Qg6 35.Kf2 Rc2?!

Kasparov has the slightly more-exposed king, so logical here is 35...Qxd3 36.Rxd3 Rc2 with a likely draw, but he pushes for more and quickly runs into trouble.

36.Qe3!

Threatening to win with 37.Rd8. There is only one defence.

36...Rc8 37.Qe7 b5?

Black had to give up the h pawn by 37...Qf7 38.Qxf7+ Kxf7 39.Rxh5, after which the engines reckon 39...g5 40.Bd6 Ne6 gives drawing chances.

38.Rd8 Rxd8 39.Qxd8 Qf7 40.Bd6 g5 41.Qa8 Kg7 42.Qxa6 1-0

White will win the b5 pawn as well, rendering Black's game hopeless.

Champion Of Champions (part six)

HERE is the tournament bracket after five matches.

Round of 16               Quarter-Finals
1. Steinitz
v----------------------------Carlsen
16. Carlsen (+102.5)

8.Tal (+8.5)
v----------------------------Tal
9.Petrosian

6. Botvinnik (+66)
v----------------------------Botvinnik
11. Fischer

13. Kasparov
v----------------------------Alekhine
4. Alekhine (+21.5)

3. Capablanca
v----------------------------Kramnik
14. Kramnik (+12)

5. Euwe
v--------------------------
12. Karpov

7. Smyslov
v--------------------------
10. Spassky

15. Anand
v--------------------------
2. Lasker

Round of 16
Match Six: Euwe v Karpov
Game One
Max Euwe - Anatoly Karpov
Nimzo-Indian
1.d4 Nf6 2.c4 e6 3.Nc3
The first time in this tournament White has allowed a Nimzo-Indian.
3...Bb4 4.Qc2 0-0 5.Bg5!?

Karpov has not faced 5.Bg5!?

Stockfish11 and Komodo11.01 are unimpressed with Euwe's opening play, evaluating the position as dead-equal, giving the Dutch world champion an uphill task in game two.

Game Two
Anatoly Karpov - Max Euwe
QGD Semi-Tarrasch
1.d4 Nf6 2.c4 e6 3.Nf3
No Nimzo-Indian this time.
3...d5 4.Nc3 c5!?
Aggressive play by Euwe. He preferred this over 4...c6 by 17 games to 16 in ChessBase's 2020 Mega database.
5.cxd5 Nxd5 6.e4 Nxc3 7.bxc3 cxd4 8.cxd4 Bb4+
Euwe reached the position after 8.cxd4 twice as Black, playing 8...Nc6 in a 1932 draw against Flohr and the text in a 1937 draw against Alekhine. The text is chosen on tiebreak by having been played more recently.
9.Bd2 Bxd2+ 10.Qxd2 0-0 11.Rc1

As far as we can go

The engines give Karpov an averaged advantage of +11.5, meaning he wins the match by the same score.
Here is the updated tournament bracket.

Round of 16               Quarter-Finals
1. Steinitz
v----------------------------Carlsen
16. Carlsen (+102.5)

8.Tal (+8.5)
v----------------------------Tal
9.Petrosian

6. Botvinnik (+66)
v----------------------------Botvinnik
11. Fischer

13. Kasparov
v----------------------------Alekhine
4. Alekhine (+21.5)

3. Capablanca
v----------------------------Kramnik
14. Kramnik (+12)

5. Euwe
v----------------------------Karpov
12. Karpov (+11.5)

7. Smyslov
v--------------------------
10. Spassky

15. Anand
v--------------------------
2. Lasker

Champion Repertoire (part 12)

ANATOLY Karpov was world champion from 1975 to 1985, and arguably from 1993 to 1999 via the separate Fide world title.
A repertoire based on his games emphasises positional mainlines in hardcore mainline openings.
White
Open 1.d4 (although Karpov preferred 1.e4 in his younger days, he has almost twice as many games with 1.d4 in ChessBase's 2020 Mega database), aiming for a set-up with Nf3 against attempts to play the Nimzo-Indian, ie 1...Nf6 2.c4 e6 3.Nf3.
If Black goes for a Queen's Indian: 3...b6, counter-fianchetto with 4.g3, meeting the mainline 4...Ba6 with 5.b3 Bb4+ 6.Bd2 Be7 7.Bg2 c6 8.Bc3. Against 4...Bb7 5.Bg2 Be7 play the slightly unusual 6.Nc3, meeting 6...Ne4 with 7.Bd2 and meeting 6...0-0 7.0-0 Ne4 with 8.Qc2 (Karpov also played 8.Bd2 and 8.Nxe4, but scored 20 percentage points more with 8.Qc2). Against 4...Bb4+ play 5.Bd2.
If Black makes it a Queen's Gambit Declined with 3...d5, play 4.Nc3, meeting 4...Be7 with 5.Bg5, generally avoiding playing the Exchange Variation. Against 4...c6 play 5.e3, meeting the main reply 5...Nbd7 with 6.Qc2. Against 4...dxc4 play 5.Qa4+. Against 4...Bb4 play 5.Bg5 h6 6.Bxf6 Qxf6, reaching a position from which Karpov had most success, albeit from a small sample-size, with the slightly unusual 7.Qb3.
Against the Bogo-Indian: 3...Bb4+ play 4.Bd2, meeting 4...Qe7 and 4...a5 with 5.g3 and meeting 4...Bxd2+ 5.Qxd2 d5/0-0 with 6.g3. Against 4...c5 play 5.Bxb4 cxb4 6.g3.
Against the Benoni: 2...c5 play 3.d5, meeting 3...e6 with 4.Nc3 exd5 5.cxd5 d6 6.e4 g6 7.Nf3 Bg7 8.h3 0-0 9.Bd3. Against the Benko Gambit: 3...b5 play 4.cxb5 a6 5.bxa6, meeting 5...g6 with 6.Nc3 and continuing 7.e4 against 6...Bxa6 and 6...Bg7. Against 5...Bxa6 play 6.Nc3, transposing to a line already considered after 6...g6 7.e4.
Against 2...g6 play 3.Nc3, meeting the King's Indian: 3...Bg7 4.e4 d6 with 5.f3 and meeting the Grünfeld: 3...d5 with 4.cxd5 Nxd5 5.e4 Nxc3 6.bxc3 Bg7 7.Be3 c5 8.Qd2.
Against the Budapest Gambit: 2..e5 play 3.dxe5 Ng4 4.Bf4 Nc6 5.Nf3 Bb4+ 6.Nbd2 Qe7 7.e3.
Against the Slav: 1...d5 2.c4 c6 play 3.Nf3 Nf6 4.e3. Against 2...e6 play 3.Nc3, transposing to lines already considered after 3...Nf6 4.Nf3 and 3...Be7 4.Nf3 Nf6 5.Bg5.
Against the Tarrasch: 1...d5 2.c4 e6 3.Nc3 c5 play 4.cxd5 exd5 5.Nf3 Nc6 6.g3.
Against the Queen's Gambit Accepted: 2...dxc4 play 3.e4 e5 4.Nf3 exd4 5.Bxc4. If 3...Nf6 play 4.e5 Nd5 5.Bxc4 Nb6 6.Bd3.
Against the Chigorin: 2...Nc6 play 3.Nf3, and against the Albin Countergambit: 2...e5 play 3.dxe5 d4 4.Nf3 Nc6 5.a3.
Against 1...e6 Karpov mainly switched to a French with 2.e4, but scored much better rating-wise, and with a higher percentage score, by playing 2.c4, which will nearly always transpose to lines already considered. If Black makes it a delayed Dutch with 2...f5, play a fianchetto set-up with g3, which was also Karpov's favourite answer to 1...f5.
Against 1...d6 and 1...g6 Karpov liked to head for a Classical Pirc or Classical Modern with 2.e4, 3.Nc3 and 4.Nf3.
Black v 1.e4
Play 1...e5 with the aim of meeting the mainline Spanish: 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 d6 8.c3 0-0 9.h3 with 9...Bb7 10.d4 Re8. Meet 11.Nbd2 with 11...Bf8 12.a4 h6 13.Bc2 exd4 14.cxd4 Nb4 15.Bb1 c5. Today 12.d5 is probably most popular - Karpov liked meeting it with 12...Nb8 13.Nf1 Nbd7. Meet 11.Ng5 with 11...Rf8. If White continues with the main move 12.Nf3, and a draw by repetition is not satisfactory, avoid 12...Re8 and instead play 12...Nd7, when the main line runs 13.Nbd2 exd4 14.cxd4. The only time Karpov reached this position he continued with the unusual 14...Nb4!?, which for quite some time is Stockfish11's top choice (the normal move is 14...Bf6).
White has many alternatives to following the main line of the Spanish, an important one, at least at club level, being the Mordern Exchange Variation: 4.Bxc6 dxc6 5.0-0, against which Karpov scored 80% by heading for a queenless middlegame with 5...f6 6.d4 exd4 7.Nxd4 c5, meeting both 8.Nb3 and 8.Ne2 with 8...Qxd1.
Also important at club level is 5.Qe2, against which play 5...b5 6.Bb3 Bc5, and the closely related 5.0-0 Be7 6.Qe2, against which Karpov played 6...b5 7.Bb3, usually continuing 7...0-0 8.c3 d6 or 7...d6 8.c3 0-0.
Against the Deferred Exchange: 5.0-0 Be7 6.Bxc6 play 6...dxc6 7.d3 Qd6!? but meet 7.Nc3 with 7...Nd7.
Against the Italian Game: 3.Bc4 play 3...Bc5, meeting the traditional mainline 4.c3 Nf6 5.d4 with 5...exd4 6.cxd4 Bb4+. If 7.Bd2 play 7...Bxd2+ 8.Nbxd2 d5 9.exd5 Nxd5. If 7.Nc3 play 7...Nxe4 8.0-0 Bxc3 9.d5 Ne5!? If 5.d3 play 5...d6 6.0-0 0-0. Meet 4.d3 and 4.0-0 with 4...Nf6, both lines usually transposing to those with 4.c3 and 5.d3. There is only one game with Karpov facing the Evans Gambit: 4.b4 - he took the pawn with 4...Bxb4 and, after 5.c3, retreated the bishop to a5.
Against the Scotch: 3.d4 exd4 4.Nxd4 Karpov mostly played 4...Nf6.
Against the Scotch Four Knights: 3.Nc3 Nf6 4.d4 exd4 5.Nxd4 he mostly played 5...Bb4, and against the Spanish Four Kinghts: 4.Bb5 he mostly played 4...Bb4.
Against the Vienna: 2.Nc3 he liked 2...Nf6, meeting 3.f4 d5 4.fxe5 Nxe4 5.Nf3 with 5...Be7 and meeting 3.g3 with 3...d5.
Against the Bishop's Opening: 2.Bc4 he narrowly preferred 2...Nc6 over 2...Nf6.
Karpov usually accepted the King's Gambit: 2.f4, but after 2...exf4, with which he scored 67%, he varied his response the three times he faced 3.Nf3 and the two times he faced 3.Bc4. There is not enough to give a repertoire recommendation but, for what it is worth, Karpov won - against a 2634 -the only time he declined the gambit with 2...d5.
Black v 1.d4 and Others
Aim for a Nimzo-Indian: 1...Nf6 2.c4 e6 3.Nc3 Bb4, meeting 4.e3 with 4...c5, meeting 4.Qc2 with 4...0-0, meeting 4.Nf3 with 4...b6, meeting 4.f3 with 4...d5, meeting 4.Bg5 h6 5.Bh4 with 5...c5, meeting 4.a3 with 4...Bxc3+ 5.bxc3 c5 and meeting 4.Qb3 with 4...c5.
Against 3.Nf3 play a Queen's Indian: 3...b6, meeting 4.g3 with 4...Ba6, meeting 4.a3 with 4...Bb7, meeting 4.Nc3 with 4...Bb4 (transposing to the Nimzo-Indian line 4.Nf3 b6) and meeting 4.e3, 4.Bf4 and 4.Bg5 with 4...Bb7.
Against 3.g3 play 3...d5 with transposition to a mainline Catalan after 4.Bg2 Be7 5.Nf3 0-0 6.0-0 or 4.Nf3 Be7 5.Bg2 0-0 6.0-0. Karpov liked to continue 6...dxc4, meeting 7.Qc2 and 7.Qa4 with 7...a6 and meeting 7.Ne5 and 7.Nc3 with 7...Nc6.
Against 2.Nf3 Karpov normally played 2...e6 with obvious transpositional possibilities. If 3.Bg5 he marginally more often played 3...h6 over 3...c5, scoring almost equally well with both moves. If 3.Bf4 he did best with 3...Be7, varying on whether he followed up with ...c5, ...b6 or ...d5.
Against the Trompowsky: 2.Bg5 Karpov scored best with 2...d5.
Against the English: 1.c4 Karpov scored best with 1...e5, usually meeting 2.Nc3 with 2...Nf6 3.g3 Bb4 4.Bg2 0-0. But his most common reply to 2.g3 was 2...g6, happily meeting 3.d4 with 3...d6!? 4.dxe5 dxe5 5.Qxd8+ Kxd8 and meeting 3.Bg2 Bg7 4.Nc3 with 4...Ne7.
Against 1.Nf3 he generally replied 1...Nf6 and 2...e6, but if White went for a King's Indian Attack set-up with 2.g3 he preferred 2...d5 3.Bg2 c6 4.0-0, scoring best by continuing 4...Bg4.
In two games against 1.b3 he played 1...d5.

Here is one of Karpov's earliest games with 1.d4.
Karpov - Boris Spassky
Candidates Semi-Final (Leningrad) 1974, Game 11
Queen's Gambit Declined
1.d4 Nf6 2.c4 e6 3.Nf3 d5
Karpov wrote in My Best Games (RHM Press 1978): "An interesting psychological moment. White has not allowed the Nimzo-Indian and Spassky, for his part, refrained from playing the quiet Queen's Indian Defence. Maybe he did not think his opponent was ready for every opening beginning with d4, for in serious play I had opened with the queen's pawn only five times in my life."
4.Nc3 Be7 5.Bg5 h6 6.Bh4 0-0 7.e3 b6

Black has committed to a queenside fianchetto

This is known as the Tartakower Variation, which is easily most popular plan after 7.e3.
8.Be2!?
The main line goes 8.cxd5 Nxd5 9.Bxe7 Qxe7 10.Nxd5 exd5 11.Rc1, which Fischer used to beat Spassky in their 1972 world championship match.
8...Bb7 9.Bxf6!?
Giving up the bishop-pair may look strange, but Karpov has a concrete plan in mind. First he ensures Black cannot meet cxd5 with ...Nxd5, which keeps the long light-square diagonal open for the black queen's bishop.
9...Bxf6 10.cxd5
Now Black has an unpleasant choice between 10...Bxd5!?, after which 11.Nxd5 deprives him of his bishop-pair and leaves queenside light-square weaknesses (this is probably playable but is very unpopular), and ...
10...exd5
... after which the long light-square diagonal is blocked and, as Matthew Sadler points out in Queen's Gambit Declined (Everyman Chess 2000), Black has kingside light-square weaknesses (f5 is no longer defended by a pawn and h7 is no longer defended by the king's knight). In addition the black c pawn is backward and Black does not look ready to play ...c5 any time soon.
11.0-0
More popular today is 11.b4 but Carlsen and Kramnik have preferred the text.
11...Qd6!?
There are six more popular moves than this in Mega20, but that does not mean the text is necessarily bad.
12.Rc1
The game Korchnoi - Antonio Medina García, Palma de Mallorca 1972, went 12.Qc2 Nc6 13.Rc1 a6 14.a3 Rfe8 15.Rfd1 with an unclear position (1-0, 35 moves).
12...a6
Karpov: "13.Nb5 must be prevented."
13.a3 Nd7 14.b4 b5 15.Ne1
This is an inaccuracy, according to Karpov. He gave "15.Nd2!, followed by Nb3 and possibly Na5," the idea being to prevent ...a5.
15...c6!?
Karpov awarded this move an exclamation mark, but it is not liked by Stockfish11 or Komodo11.01, who reckon White can now get an initiative with 16.e4!? dxe4 17.Nxe4.
16.Nd3 Nb6?!
Karpov gives this a question mark, saying 16...a5 equalises "easily (by) opening the a file and practically eliminating all his weaknesses."
17.a4 Bd8!?
Karpov says Spassky hoped to significantly weaken the castled position of the white king "but this (manoeuvre) will definitely destroy the interaction among his own pieces." Karpov reckoned best is 17...Rad8.
18.Nc5 Bc8 19.a5 Bc7 20.g3 Nc4 21.e4 Bh3 22.Re1 dxe4 23.N3xe4 Qg6 24.Bh5!
Karpov: "After 24.Bxc4? bxc4 25.Rxc4 f5! and 26...f4, Black gets a very strong attack."
24...Qh7
Karpov says 24...Qf5 is bad because after 25.Rc3 "Black has no good defence to g4."
25.Qf3 f5?
Karpov reckoned that after 25...Qf5 White has a "considerable advantage" but not one that is simple to convert.
26.Nc3 g6 27.Qxc6 gxh5 28.Nd5
The black king is suddenly clearly more endangered than its white counterpart.
28...f4 29.Re7 Qf5 30.Rxc7 Rae8 31.Qxh6 Rf7 32.Rxf7 Kxf7 33.Qxf4 Re2 34.Qc7+ Kf8 35.Nf4 1-0

My Best Ever Tournament (part seven)

Laurent Jacot (2250) - Spanton (2096)
Lausanne Round 7
Sicilian Closed
1.e4 c5 2.Nc3 g6 3.g3 Bg7 4.Bg2 Nc6 5.d3 d6 6.f4
The first time I had reached this position with either colour. I see from my original notes that I strongly considered setting up a Reversed Botvinnik formation with 6...e5. LJ told me in the postmortem he had only just taken up the Closed Sicilian, and was only prepared for a Reversed Botvinnik!
6...e6 7.Nf3 Nge7 8.0-0 0-0 9.Be3 Nd4 10.Bf2!?
The main idea of this move, which has been played by Boris Spassky and Michael Adams, is to enable White to capture on d4 with his king's knight. Most popular is 10.e5.
10...Rb8
Anatoly Karpov has pre-empted White's idea by playing 10...Nxf3+!?
11.Nxd4 cxd4 12.Ne2 Nc6 13.c3 dxc3 14.Nxc3!? Nd4!?
This apparent-novelty is liked by Stockfish10 and Komodo10.
15.Ne2 ½–½
LJ explained that he offered the draw because my results showed, in his view, I was much stronger than my rating, and he was playing an opening with which he was unfamiliar.

My Best Ever Tournament (part two)

Spanton (2096) - Vladimir Tukmakov (2581)
Lausanne 1999 Round 2
Caro-Kann Panov-Botvinnik
Vladimir Tukmakov is a Ukrainian grandmaster whose best results, according to Wikipedia, were in the days of the Soviet Union when he finished second to Bobby Fischer at Buenos Aires 1970, second to Anatoly Karpov at Madrid 1972 and first ahead of Tigran V Petrosian at Vilnius 1978.
1.e4 c6 2.c4 d5 3.cxd5 cxd5 4.exd5 Nf6 5.Nc3 Nxd5 6.d4 e6 7.Nf3 Bb4 8.Bd2
By a slightly unusual move-order, we have reached one of the main variations of the Panov-Botvinnik. I had just bought Jacob Aagaard's Easy Guide To The Panov-Botvinnik Attack - it had been published by Cadogan in the previous November - and had started playing the opening to learn how to use an IQP.
8...Nc6 9.Bd3 Nf6 10.0-0
Aagaard only covers 10.a3, which turns out for quite some time to be Stockfish10's top choice, narrowly ahead of the mainline text.
10...0-0
White seems to get more than enough compensation for a pawn after 10...Nxd4?! 11.Nxd4 Qxd4 12.Be3.
11.Qe2!?
However this move is labelled a "reckless pawn sacrifice" in Karpov & Mikhail Podgaets' 2006 Batsford book, Caro-Kann Defence: Panov Attack. Normal is 11.Bg5.
11...Be7
Critical is 11...Nxd4!? 12.Nxd4 Qxd4, when Larry Christiansen (2563) - Yasser Seirawan (2647), USA Championship 2000 (Seattle), quickly finished 13.Rfd1 Qg4 14.Qxg4 Nxg4 15.Be2 Nf6 16.Bf3 Bxc3 17.Bxc3 Nd5 18.Bxd5 exd5 19.Rxd5 Bxe6 ½–½.
12.Rad1
Here the main move is 12.Ne4, but the text scores a better percentage in ChessBase's 2020 Mega database.
12...Nb4
Again Black can try grabbing the pawn. After 12...Nxd4 13.Nxd4 Qxd4, Stockfish10 and Komodo10 suggest 14.Bg5, which has been seen in three high-level games, with all three possible results having occurred.
13.Bb1 b6 14.Ne4!?
14.Rfe1 was played in Mihai Suba (2500) - Petrosian (2605), Interzonal 1982 (Las Palmas), which continued 14...Bb7 15.Ne5 Rc8 16.Qe3, transposing to a position that had been reached by Petrosian as White in 1967 against Miguel Najdorf. The latter played ...Nfd5 and lost; Petrosian, against Suba, played ...g6 and drew.
14...Ba6?!
This natural-looking move may be a mistake. My move 14.Ne4!? only appears once in Mega20, in the game Petra Papp (2256) - Pal Kiss (2362), Hungarian team championship 2011, and was met by 14...Nbd5 15.Ne5 Bb7 (½–½, 26 moves).
15.Nxf6+ gxf6

How should White proceed?

*****
*****
*****
*****
*****
16.Bxh7+?
In playing my 14th move, I had missed that 16.Qe4 is met by 16...f5. I thought the bishop sacrifice gave me my best chance of saving the game, but the engines point out 16.Qe3, when 16...Bxf1? leaves White on top after 17.Bxb4 Bc4 (17...Bxb4?? 18.Qh6) 18.Bxe7 Qxe7 19.Nd2, eg 19...Bd5 20.Qg3+ Kh8 21.Qh4. Best for Black, according to the engines, is 16...Nd5, when 17.Qh6 f5 18.Rfe1 is slightly better for White, they reckon.
16...Kxh7 17.Qe4+ f5 18.Qf4 Rh8
Best, according to the engines.
19.Ne5 Kg8 20.Rfe1 Nd5?!
The engines much prefer 20...Kf8.
21.Qg3+ Kf8 22.Ng6!
My notes show Crafty - I do not know which version - did not like this move, but it is definitely White's best shot, according to modern engines.
22...fxg6 23.Qxg6 Qd6?
I recall being worried about 23...Bc8, which is indeed better than the text. But best, according to Stockfish10 and Komodo10, is 23...f4, blocking White's bishop from the kingside. They continue 24.Rxe6 Bc8 25.Re5, and now 25...Qd6!? or 25...Qe8!? to get queens off at the cost of returning some material, eg 25...Qd6!? 26.Qxd6 Bxd6 27.Rxd5 Ke7, when Black's extra bishop is worth much more than White's extra three pawns, according to Stockfish10, although Komodo10 is not so sure.
24.Bh6+ Rxh6 25.Qxh6+ Kf7 26.Qh7+ Kf6 27.Qh6+ Kf7 28.Qh7+ Kf6 29.Qh6+ ½–½
Tukmakov could have played on with 25...Ke8, but Stockfish10 and Komodo10 reckon 26.Rxe6 Qf4! (this seems an only-move) 27.Qh8+ Kf7! 28.Rxe7+! Nxe7 (it is important that the knight protects the black king from a check at a7) 29.Qxa8 Be2 is dead-level despite the unbalanced material (White has rook and three pawns for bishop and knight).