What' s the head-to-head if you exclude the games from Kramnik's teenage years?"

To the best of my knowledge, this variation (7...h6) is dead, at least for the time being. No one has so far been able to find a way for Black to untangle and bring his king to safety - which is what he needs to do to make his extra piece count.

True, White is not able to blast his way through immediately, but the sacrifice gives him a tremendous advantage in space, activity and co-ordination. Long-term considerations like these are completely "Martian" to computers, since they only "understand" material count and short-term tactics. That's why Fritz, left to its own devices, will not play 8.Nxe6 unless it has an opening book that tells it to do so.

I guess this is why Kasparov decided to try this line in the decisive game against Deep Blue back in 1997. Kasparov obviously thought that the computer would only be "booked up" in the lines he normally plays. As we all know, Garry was in for a rather nasty surprise.

After this famous game, 7...h6 has virtually disappeared from international tournaments. Last year the strong Russian IM Denis Evseev decided to tempt fate by wheeling this risky line out against a weaker opponent and was duly rewarded by having his name added to the casualty list:

Kalod (2440)-Evseev (2539)

Pardubice (6), 1999

1.e4 c6 2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nd7 5.Ng5 Ngf6 6.Bd3 e6 7.N1f3 h6?

8.Nxe6! fxe6

Deep Blue-Kasparov went:

8...Qe7? ("After this Black is just lost." -GM John Nunn) 9.0-0 fxe6 10.Bg6+ Kd8 11.Bf4 b5 12.a4 Bb7 13.Re1 Nd5 14.Bg3 Kc8 15.axb5 cxb5 16.Qd3 Bc6 17.Bf5! exf5 (17...Nb4 18.Qc3 Kb7 19.Rxe6 Qd8 20.d5! Bxd5 21.Re8 and White wins the piece back with a decisive plus.) 18.Rxe7 Bxe7 19.c4 and Black resigned.

9.Bg6+ Ke7 10.0-0 Qc7 11.Re1 Kd8

White has only one pawn for the piece, but his lead in development, combined with the superior co-ordination of his forces, gives him a huge, if not decisive advantage. Black needs oceans of time to get his guys out, and White usually comes crashing through in the centre before this happens.

12.c4 Bb4 13.Bd2 Bxd2 14.Qxd2 Nf8 15.Bc2 Bd7 16.b4 Be8 17.Bb3 Bh5 18.Ne5 N8d7

19.f4!

White is not afraid of further exchanges, since Nxe5 will allow him to organise a devastating attack along the d-file.

19...Re8 20.h3 Re7 21.Kh1 a5 22.a3 axb4 23.axb4 Qb8

The grim alternative was to hand the a-file over to White with Rxa1.

24.Rac1 Ra3 25.Qb2 Ra6 26.Qd2

White virtually controls the whole board. Breaks on b5 and d5 are in the air, and he also has the option of grabbing even more space with g2-g4. Black tries to relieve the pressure by exchanging a pair of knights, but this only makes matters worse.

26...Nxe5 27.dxe5+ Nd7

28.f5! Bf7

28...exf5 29.e6 and White wins his piece back with his attack still intact.

29.c5 Ke8 30.Qd6!

After this it's clear that Black will have to shed material.

30...Ra8 31.fxe6 Bxe6 32.Bxe6 Nf8 33.Bf5 Qc7 34.Bg6+ Nxg6 35.Qxg6+ Kf8 36.Qh7 Rae8 37.Rcd1 Rf7 38.e6

and Black resigned, since mate is unavoidable. 1-0

So, White does not need to force matters in this variation. He can simply build up slowly with sensible moves. This is the kind of position that, in my opinion, does not lend itself to heavy analysis, so the evaluation has to be based on practical results and positional considerations. And these clearly indicate that Black should stick to the normal 7...Bd6.

With regards to Shirov vs Kramnik:

No, Shirov is not much older than Kramnik. There's only three years between them. Yes, including the games they played while Kramnik was still a junior might not be completely fair, but let's not forget that Kramnik's rating cleared the 2700-mark when he was only 18 (in 1993). At present he's 2767, so one would have to say that he was pretty well developed in his late teens.

But more importantly, most of their encounters took place after Kramnik turned 20. As a junior, Kramnik faced off against Shirov only eight times (in the period of 1992-1994), and Shirov "only" managed +1 in these encounters. Since then they've played 42 games, with Shirov scoring +4.