Showing posts with label Spanish Steinitz. Show all posts
Showing posts with label Spanish Steinitz. Show all posts

Wednesday, 15 July 2026

Lessons From Magdeburg II

IN round two I was White against Jürgen Gottschalk (1711), who chose to play the Steinitz Defence to the Spanish: 1.e4 e5 2.Nf3 Nc6 3.Bb5 d6.
The defence is certainly playable - indeed, there are are 21,760 examples of 3...d6 in ChessBase's 2026 Mega database.
But it is surely no coincidence that Black's miserable score of 32% is significantly below Black's score of more than 40% with most other third-move continuations, including ...a6, ...Nf6, ...Bc5, ...f5, ...Nge7 and ...g6.
After 4.d4 exd4 (Stockfish18 and Dragon1 reckon this is better than Steinitz's preferred 4...Bd7) 5.Nxd4 Bd7 6.0-0 Nf6 7.Nc3 Be7, a position reached 2,665 times in Mega26, White has at least a slight edge (Stockfish18) or even the upper hand (Dragon1).
Position after 7...Be7
Black was positionally lost just four moves later, according to the engines, and never recovered.
LESSON: if you choose a defence that statistics suggest is suspect, you must know what you are doing, as otherwise it is easy to quickly drift into a lost position.

Tuesday, 7 July 2026

Magdeburg Round Two

I WAS upfloated.

Spanton (2039) - Jürgen Gottschalk (1711)
Spanish Steinitz
1.e4 e5 2.Nf3 Nc6 3.Bb5 d6
The Steinitz Defence is generally out of fashion, but is still the third-most popular continuation in ChessBase's 2026 Mega database, behind 3...Nf6 and especially 3...a6.
4.d4 exd4
Steinitz almost always played 4...Bd7, but the text is preferred by Stockfish18 and Dragon1.
5.Nxd4 Bd7 6.0-0
More common is 6.Nc3, maintaining the option of castling long.
6...Nf6 7.Nc3 Be7 8.b3!?
The engines do not like this. They suggest 8.Nxc6, when Black usually replies 8...bxc6, after which they reckon 9.Bd3 0-0 10.f4 gives White the upper hand.
8...a6!?
They also dislike this, because of the possibility of White replying 9.Nxc6.
9.Be2!? Ne5!?
Black should probably castle.
10.f4!?
Apparently a novelty, and an improvement, according to the engines, over the known 10...a4.
After 10.f4!?, how should Black respond?
*****
*****
*****
*****
*****
10...Ng6
This is probably best, as 10...Nc6!? represents the loss of two tempi, while 10...Neg4? 11.h3 Nh6 is very good for White, eg 12.e5!? (this seems even better than 12.g4, when the engines' 12...d5!? just about keeps Black in the game) c5!? 13.Nf3 dxe5 14.fxe5 Nh5 15.Qe1, with an advantage worth about a minor piece, according to the engines.
11.Bb2 c5?!
The engines suggest castling or 11...b5.
After 11...c5?!, how should White proceed?
*****
*****
*****
*****
*****
12.Nf5!
White would have enough for a pawn, but no more, after 12.Nf3 Nxf4 13.e5 dxe5 14.Nxe5 Nxe2+ 15.Qxe2 Be6, according to the engines.
How should Black continue?
*****
*****
*****
*****
*****
12...0-0!?
The engines reckon this is best, as 12...Bxf5 13.exf5 Nh4 can be well-met by 14.Qd3 or 14.g4, but probably not 14.Bd3?!, when Black has 14...d5!, eg 15.g3 c4! but almost certainly not 14...Qd7?!, as then 15.Qe1! Nxf5 16.g4 Nd4 17.g5! is strong, one line running 17...Qg4+ 18.Kh1 Ng8 19.Nd5, with a large advantage for White.
13.Nxd6 Qc7!?
Possibly better is 13...Bc6, although 14.e5 is good for White, and 14.Nf5!? may be even better, eg 14...Nxe4 15.Nxe4 Bxe4 16.Nxg7.
14.e5 Ne8?
Definitely better is 14...Bxd6 15.exd6 Qb6, but White would be a pawn up, and with the better position.
15.Nd5 Qb8
This is Dragon1's top choice; Stockfish18 marginally prefers 15...Qd8.
16.Nb6
It is unusual, to say the least, for a player to get two knights to the sixth rank this early in a game
*****
*****
*****
*****
*****
16...Bc6!?
The engines agree this is best, but there is no way for Black to avoid serious material loss.
17.Nxe8 Rxe8 18.Nxa8 Qxa8 19.f5!? Nh4
Retreating the knight seems even worse.
20.f6 gxf6 21.exf6 b5!?
Muddying the waters. Objectively best, according to the engines, is 21...Bd6!?, although they reckon 22.Qxd6 Rxe2 23.Rf2 Rxf2 24.Kxf2 leaves White with an advantage worth more than a rook.
22.fxe7 Nxg2
Perhaps a better try is 22...Bxg2, or 22...Rxe7, but both can be successfully met by, among other moves, 23.Bh5.
The game finished:
23.Bh5 Rxe7 24.Qg4+ Kf8 25.Qg7+ Ke8 26.Bxf7+ Kd7 27.Rad1+ Kc7 28.Qg3+ Kb6 29.Rd6 Ne3 30.Rff6 Nf5 31.Rxc6+ Qxc6 32.Rxc6+ Kxc6 33.Qf3+ 1-0

Friday, 10 July 2020

Champion Repertoire (part three)

IT will surprise many people to learn that third world champion José Capablanca played 1.e4 more often than 1.d4.
But he scored a higher percentage with the latter and it came to dominate his opening choice with the white pieces.
After 1.d4 Capablanca scored highly against 1...d5 but even better against the modern favourite 1...Nf6, which makes him an interesting repertoire-model for today's club player.
White
After 1.d4 Nf6 2.c4  the Nimzo-Indian: 2...e6 3.Nc3 Bb4 was all the rage when Capablanca was at the peak of his career. He scored 76% (all percentages are from ChessBase's 2020 Mega database) with 4.Qc2. The main line then went 4...d5 5.cxd5 Qxd5 when Capablanca preferred 6.Nf3, meeting 6...c5 with 7.Bd2 Bxc3 8.Bxc3. Today 5...exd5 is more popular - Capablanca met it with 6.Bg5, which is still the main reply.
Capablanca scored heavily against the Grünfeld: 2...g6 3.Nc3 d5 without establishing a favourite between 4.Nf3, 4.Bf4 and 4.cxd5. Perhaps 4.Bf4 is the way to go as it is the least-common of these moves today but has been played by Giri, Anand, Aronian, Ivanchuk and many other top grandmasters.
Capablanca only faced the King's Indian: 3...Bg7 once, choosing the Sämisch: 4.e4 d6 5.f3.
After 1.d4 Nf6 2.c4 e6 3.Nc3 most of Capablanca's opponents transposed to the Queen's Gambit Declined with 3...d5. Most games continued 4.Bg5 Nbd7 5.e3 Be7 (Black's last two moves were quite often played in reverse-order) 6.Nf3 0-0 7.Rc1 c6 8.Bd3 dxc4 9.Bxc4 Nd5 10.Bxe7 Qxe7 which is still the main line today. Capablanca scored 71% after 11.0-0.
Against the Slav: 1.d4 d5 2.c4 c6 Capablanca scored 79% with 3.Nf3, rising to 85% when he met 3...Nf6 with 4.e3. He was usually happy going down the main line: 4...e6 5.Nc3 Nbd7 6.Bd3 dxc4 7.Bxc4 b5, scoring 67% in three games with 8.Bd3 but 50% in three games with less-popular 8.Be2.
He seems to have only faced 4...Bf5 twice, drawing with 5.Qb3 but winning with 5.cxd5 cxd5 6.Qb3.
He only faced the Queen's Gambit Accepted: 2...dxc4 four times, playing four different replies, drawing with 3.Nc3 and 3.e4 but winning with 3.Nf3 and 3.e3. 
Against the Dutch: 1...f5 Capablanca scored a sprightly 60% with the Staunton Gambit: 2.e4.
Against 1...e6 he nearly always went for a French with 2.d4 d5 3.Nc3, meeting the Classical: 3...Nf6 with 4.Bg5 Be7 5.e5 Nfd7 6.Bxe7 and after 6...Qxe7 he scored highly with both 7.Nb5 (Lasker's favourite) and 7.Qd2.
Again like Lasker he met the McCutcheon: 4...Bb4 with 5.exd5.
Against the Winawer: 3...Bb4 he kept the game open and scored well, albeit from a small sample, with 4.Bd3!?
Black v 1.e4
Capablanca overwhelmingly replied 1...e5, meeting the Spanish: 2.Nf3 Nc6 3.Bb5 with the Steinitz Deferred: 3...a6 4.Ba4 d6. He only faced today's favourite 5.c3 three times, but won both games in which he replied with the sharp Siesta Variation: 5...f5. More common in his day was 5.0-0 Bd7 6.c3, which he met with 6....g6.
Against the modern Deferred Exchange: 5.Bxc6+ bxc6 6.d4 he scored 87% with the main reply 6...f6.
Against the Vienna: 2.Nc3 Capablanca played the slightly unusual 2...Nc6, which has also been the choice of Carlsen, Caruana, Nakamura and Aronian. One of the points is to avoid the theory-laden sharp lines of 2...Nf6 3.f4. Capablanca's opponents nearly always replied 3.Nf3 after which he was happy to transpose into a Four Knights with 3...Nf6.
Against the King's Gambit: 2.f4 Capablanca scored 100% with 2...exf4, meeting 3.Nf3 with 3...g5.
Against the Italian Game: 2.Nf3 Nc6 3.Bc4 he usually replied 3...Bc5 but scored much better with the Two Knights: 3...Nf6, meeting the critical 4.Ng5 with 4...d5 5.exd5 Na5.
He successfully met the Scotch: 3.d4 exd4 4.Nxd4 with 4....Nf6, but had trouble with the Scotch Gambit: 4.Bc4, scoring just one draw from three outings with the very-sharp 4...Bc5 5.0-0 Nf6 6.e5 d5 7.exf6 dxc4 8.Re1+ Be6 9.Ng5 etc. Theoretically this is fine for Black but requires a lot of memorisation. An interesting alternative is 5...Nge7 which Capablanca won with the only other time he faced this line.
Black v 1.d4 and Others
Capablanca normally replied 1...Nf6 but after 2.c4 e6 3.Nc3 he usually switched to the Queen's Gambit Declined with 3...d5, whereupon many of his games went 4.Bg5 (no one played 4.cxd5) Be7 5.e3 0-0 6.Nf3 Nbd7 7.Rc1 c6 8.Bd3 dxc4 9.Bc4 Nd5 10.Bxe7 Qxe7. This is very mainline but not so relevant today, at least at club level, as most Whites now play 4.cxd5, for which Capablanca's games are no help. Similarly, after 3.Nf3 he most often played 3...d5.
Capablanca only faced the London: 2.Nf3 d5 3.Bf4 four times, scoring two-out-of-two with 3...c5, the idea being to quickly generate queenside pressure with ...Qb6.
Against 1.Nf3 he usually played 1...Nf6, meeting 2.c4 with 2...e6.
The English: 1.c4 was a rare bird in Capablanca's day but he usually replied 1...Nf6, meeting 2.Nf3 with 2...e6 and 2.Nc3 with 2...c5.

Capablanca famously lost his world title to Alexander Alekhine, but what is less well-known is that their career score favoured the Cuban by +9 =33 -7.
His match wins against Alekhine, and his win at Nottingham 1936 (their first encounter since Capablanca lost their 1927 match), appear in many anthologies.
Here is a lesser known game from earlier in their careers.
Alekhine - Capablanca
St Petersburg 1914
Spanish Steinitz
1.e4 e5 2.Nf3 Nc6 3.Bb5 d6 4.d4 exd4 5.Nxd4 Bd7 6.Nc3 Nf6 7.0-0 Be7
There are more than 1,600 games with this position in ChessBase's 2020 Mega database.
8.Nf5!?
A double-edged choice - White gets the bishop-pair at the cost of a damaged pawn-structure.
8...Bxf5 9.exf5 0-0 10.Re1
Twenty-one years later Alekhine played 10.Bf4 in a simul loss.
10...Nd7!?
This seems slow. Normal is 10...Re8.
11.Nd5 Bf6 12.c3
Both players presumably felt 12.Nxf6+ Qxf6 swops a well-placed knight for a bishop that looks active but after the text is blunted by White's queenside pawn-structure.
12...Nb6 13.Nxf6+
Alekhine changes his mind and goes for exchanges that will see him give up the bishop-pair but have the slightly better pawn-structure.
13...Qxf6 14.Bxc6 bxc6 15.Qf3 Rfe8 16.Be3
What, if anything, should Black do about his hanging c6 pawn?
*****
*****
*****
*****
*****
16...c5
Much better than 16...d5? which leaves the white bishop very active on the central dark squares. However 16...Nd5!? is playable although the analysis engines Stockfish11 and Komodo11.01 like White after preserving the bishop with 17.Bd4 or 17.Bd2.
17.Re2 Re5 18.Rae1 Rae8
Tarrasch pointed out that 18...Rxf5?? loses to 19.Bg5, and also gave 18...Qxf5 19.Qxf5 Rxf5 20.Bxc5!
19.Qb7?!
Decentralising the queen to go after pawns proves dangerous. White holds the balance with, for example, 19.Bf4 Rxe2 20.Rxe2 Rxe2 21.Qxe2 Kf8 22.g4 as 22...g6?? loses to 23.Bh6+.
19...Qxf5 20.Qxc7 Qe6 21.Qxa7 Nd5
White has, temporarily at least, won a pawn, and his extra pawn is passed on the a file, but it is a long way from threatening to queen.
22.Kf1?
The engines reckon a lesser evil is 22.g3 although they prefer Black after 22...h5 or 22...f5.
22...Nf4 23.Rd2 Nxg2
Also winning is Tarrasch's 23...Qg4 24.f3 Qe6 and the engines' 23...Qc4+ 24.Kg1 Nxg2.
24.Kxg2 Qg4+ 25.Kf1 Qh3+ 26.Ke2 Rxe3+ 27.fxe3 Qxe3+ 28.Kd1
Or 28.Kf1? Qf3+ 29.Rf2 Qh1#
28...Qxe1+ 29.Kc2 Qe4+ 30.Kb3?
This makes it easier for Black. As Tarrasch pointed out, there was more resistance in 30.Kc1, eg 30...Qf4 31.Kc2 g6 32.Qd7 although Black's extra pawn and safer king should tell in the long run.
30...Qc6
The engines point out a quicker win starting with 30...Ra8, eg 31.Rxd6 h6 32.Qd7 Rb8+ 33.Qc2 etc.
31.a4 d5
But now ...Ra8 is not so clear after 32.Rxd6!
32.a5 Qb5+ 33.Ka3 Rb8 34.Ka2 h6 35.a6?
This allows an immediate finish.
35...Qb3+ 0-1
White is mated after 36.Kb1 Re8 etc.