| keypusher: I took a look at Tal's notes to the game, aided by Stockfish. <any thoughts as to why Tal ignored the knight on f4? is it b/c of Qf6 and Bg4 sac after? thoughts? thx> Yes, exactly, although taking the knight was playable. Tal: <After the text-move (12....Nf4), the variation 13.Nxf4 exf4 14.Bxf4 Qf6 15.Qd2 Bxg4 16.Bg2 was quite acceptable, with White stunting Black's counterplay.> In this variation, Black could play 14....Bxf4 immediately, after which 15.Bg2 loses to ...Rxe4+, but 15.Be2 is playable. After 28....Ne8:
click for larger viewTal: <After the text-move both players were short of time. The realisation of White's material superiority is by no means easy, since the king's bishop is quite passive. I wasted the remaining minutes trying to find a satisfactory continuation after 29.Bxc5 dxc5 30.Rxe5+ Kd6. In reply to 31.Rf5, one possibility is 31...Ng6 and if 31.Rxe8 Bxe8 32.g5 then 32....Ke5(!!) 33.gxf6 Rh1 which is probably even better for Black. Having verified this and being in time-trouble, I opted for a tranquil continuation.> The engine doubts Tal's evaluation of the final position, but 32....Ke5 is quite a move, and definitely the best Black has. Nevertheless, 29.Bxc5 was indeed the best, with the continuation 29....dxc5 30.Rxe5+ Kd6 31.Rg5! (threatening 32.e5+) Ng6 32.Qe3 Kc7 33.e5 Rb6 34.e6 Bxe6 35.Rxg6 fxg6 36.Qe5+ and wins. < Pawn and Two: After move 38, Tal notes: <Black apparently considers the position hopeless and thus shortens his resistance. A particularly interesting ending would have materialized after 38...Rg4 39.Bxf5 Rxg7 40.hxg7 Nf6 41.Bh7 Ng8! 42.Bxg8 Kf6 43.Bh7 Kxg7 44.Bf5 Kf6 45.Bc8 b6.>
Tal concludes by saying, <"...one thing is certain, it would be most difficult to find a way of winning, if one could be found at all.">Now with the aid of a computer we know that White is indeed winning in this variation indicated by Tal. Can anyone provide the winning line after 45.Bc8 b6?> Tal's suggestion was <gradually forcing the advance of the e- and f-pawns to the fourth rank, when the bishop would pin them both down from the front while the king skirted them via h4 and finally took them. The plan is workable in principle, but the pity is that at a suitable moment Black plays ...a6 and ...b5. If White wanted to frustrate this with a4 his bishop would be even more tightly hemmed in and it would be even more difficult for it to return. For that reason White would probably have to permit ...b5 which would lead to a reduction in material. A detailed analysis of such an ending would take too much space, but one thing is certain, it would be most difficult to find a way of winning, if one could be found at all.> I don't want to express a definitive opinion about this ending, but SF (i) has no compunction about playing a3-a4 (ii) has no difficulty bringing the bishop back. When I play the ending out (with the engine really making the decisions for both sides) the Black e- and f- pawns eventually do advance, and White wins them. If I override the engine and hold them back, the White king eventually penetrates down the h-file. So, I think the ending really was won for White. But I'd be curious to see what a stronger player thought. | 
Kibitzing
