Matemax wrote on 12/24/07 at 11:25:38:
Nice to see, that there is good old stuff about this variation.
Of course there is "good old stuff" on this variation! It is absolutely fundamental to the effectiveness of the Leningrad and was investigated by the pioneers such as Zak in the 1940s, building on prior analysis by Alekhine himself. Without satisfying themselves first that there was at least one way for White to gain an advantage against the queen sac, there was little point in Zak et al investigating 4 Bg5 further.
Zak thought that after 4 Bg5 h6 5 Bh4 c5 6 d5 Nxd5?! 7 Bxd8 Nxc3 8 Qb3! Ne4+ 9 Kd1 Nxf2+ the move 10 Kc1 was best, since it avoided a later check by ...Nd4. Now if 10...Kxd8 the old book line goes 11 Qf3 Nxh1 12 Qxf7 Nc6 13 Qxg7 Re8 14 Qxh6 which he evauated as clearly better for White, since he reckoned that the Nh1 could not escape. This all needs checking with modern computer analysis, of course, and there are undoubtedly some discoveries to be made, but care is required, since computer evaluation of such positions (with a knight in the corner which may or not manage to escape) is notoriously untrustworthy unless analysed to an absolutely clear conclusion.
The old theory regarded 10...Nxh1 as a better try for Black when both 11 Qf3 (Alekhine) and 11 a3 (Zak) were analysed as better for White.
It is possible that the above is irrelevant if White can actually get away with 10 Kc2, which is obviously the move he wants to play if he can, rather than have the Ra1 blocked in by Kc1. Zak feared 10 Kc2 Nxh1 11 Bc7 Nc6 12 e3 Nf2 with tricky compensation for Black, mainly based on the judgement that White's queen can no longer run riot on the kingside. I wonder what Zak feared after the natural looking altenative 12 Nf3. If then 12...Nf2 13 a3 looks a bit awkward for Black, with the idea 13...Ba5 14 Bxa5 Nxa5 15 Qc3.
After 10 Kc2 the other move 10...Kxd8 seems to fail to 11 Qg3 or 11 Qf3).
Memo: check whether the inclusion or not of 4...h6 5 Bh4 makes any substantial difference in the key lines.
The queen sac line is very interesting but I fear that the best modern analysis (i.e. by human + Fritz) will eventually find more resources for White than for Black.