Nimzo Indian, Saemisch Variation

Returning to Smyslov's Sparklers, one of the striking differences between Smyslov's discussions of his games and Kasparov's discussions is the treatment of the opening. Where Smyslov usually gives a brief discussion of one or two of the opening moves, Kasparov often gives deep analysis. The diagrammed position is a good example. It shows the position after 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 c5 5.Bd3 O-O 6.a3 Bxc3+ 7.bxc3 Nc6 8.Ne2 b6 9.O-O Ba6 10.e4 Ne8.

In fact, this position isn't found in the next sparkler, Geller - Smyslov, Amsterdam 1956. It is instead from Geller - Smyslov, Zurich 1953. Kasparov annotates both games and points out that Smyslov's success in the two Candidates Tournaments, where he twice qualified for title matches with Botvinnik, was partly due to defeating Geller, one of his main rivals in the events, by identical 2-0 scores.

The two key games with Geller as White both used the Saemisch Variation in the Nimzo Indian (4.a3 is the earliest that it can be played). Referring again to the diagram, the Amsterdam 1956 game differed from the Zurich 1953 game after 10 moves in that it had a White Knight on g3 and White hadn't castled. Smyslov writes of 4.a3:

Grandmaster Geller's favorite continuation. After the exchange on c3 White gets the two Bishops and a strong center. However, the doubled Pawns on the c-file provide Black with sufficient counterplay on the Queenside.

Kasparov writes:

The Saemisch Variation is the most direct attempt to cast doubts on the Nimzo-Indian Defense. By provoking the exchange on c3, White strengthens his center and plans an attack on the Kingside, exploiting the strength of his two Bishops. The drawback to the plan is the chronic weakness of the doubled c-Pawns, which is especially perceptible in a protracted positional struggle.

To understand the position correctly, it is important to note that White has a possible Kingside attack.

Candidates Tournament
Zurich 1953
Smyslov, Vasily

Geller, Efim
(After 10...Nf6-e8)
[FEN "r2qnrk1/p2p1ppp/bpn1p3/2p5/2PPP3/P1PB4/4NPPP/R1BQ1RK1 w - - 0 11"]

Of 10...Ne8, which is not an obvious move to play, Smyslov declines to discuss the positional underpinnings. Kasparov writes:

Both sides are aiming to exploit their trumps: White is laying the basis for an attack, whereas Black is focusing his attention on the c4-Pawn. The Knight retreat, which was introduced by Capablanca in 1929, enables the unpleasant pin Bg5 to be avoided and is an important part of Black's plan: he is ready to meet the dangerous advance f2-f4 with ...f7-f5, keeping the position closed and restricting the activity of the White Bishops.

The Capablanca game, Johner - Capablanca, Carlsbad 1929, also reached the diagram. The players continued 11.Be3 d6. Smyslov decided against playing ...d6, saving the square for the Knight to attack the embattled c-Pawn by ...Nd6.

Kasparov also mentions two other games. The first is the earlier game Botvinnik - Reshevsky, Hague/Moscow 1948, which reached a position similar to the diagram, but one move earlier: White hasn't yet castled O-O and Black hasn't yet played ...Ba6. That game, which Kasparov annotated fully in the Predecessors IV chapter on Reshevsky, continued 10.Be3 d6.

The second game is the later Yusupov - Karpov, Linares 1993, which reached the diagram. It continued 11.f4 f5 12.Ng3 g6 13.Be3 cxd4!, 'effectively putting an end to the history of the variation', according to Kasparov.

I had never understood the nature of White's compensation for the doubled Pawns. Kasparov's explanation is a good introduction to the theory of the Saemisch Variation.