White need not have been in such a hurry, though. 30.Ne7+ Kf8 31.Nf5 preserves the material status quo, after which White can think about developing his less active rook by Rh1-h2-c2 etc., or to continue manoeuvring his knight to c4.

Not 31.Nxa7 Ne6 and White loses a piece. But 31.Rxf4 Rxc6 32.Rd1 looks drawish.

White could have tried 34.Re7 here, forking the c- and f-pawns e.g. 34.Re7 R5d7 35.Rfxf7 Rxe7 36.Rxe7 and White has the marginal edge.

A very strange decision. After 39.Kd3 Rh2 (say) 40.Ke4 White's pawn deficit doesn't mean a whole lot against the activity of his king. White ought to have at least a draw, and could even try for more.

With the rooks off and White to play, White wins owing to the extra tempo towards promotion (at its simplest: 43.a5 f5 44.a6 f4 45.a7 f3 46.a8=Q f2 47.Qg2 etc. For this reason, White can fancy his chances even with the rooks on. It is White to play...

White's 43rd is very well chosen. A possible strategy for Black might be to see if he can arrange the exchange of his rook and 2 pawns for White's Q-side pawns, at the same time leaving the WK stranded far over on the Q-side. This is a well-known drawing strategy in desperate situations. To achieve this, maybe Black has to bring his rook behind the passed pawns, or somehow at least to try and slow down White's pawns. White's last move keeps watch and ward over the b-pawnwhilst the a-pawn begins to edge forward.

Very strange: the rook was quite well placed where it stood. 44.a5 Rg5 45.b4 h4 46.a6 h3 47.a7 h2 48.a8=Q+ Kg7 49.Rb8 Rg1 50.Rh8 etc. Maybe he overlooked the check which gave time for the subsequent uncovering of the long diagonal. Nor will a cunning interpolation of the rook help: 47...Rg2+ (or earlier) 48.Ka3 h2 49.a8=Q+ Kg7 50.Rxf7+ Kxf7 51.Qxg2 etc.

After which White may be struggling to save the game! A possible line is 45.a5 Rg5 46.b4 h3 47.a6 h2 48.a7 Rg2+ 49.Kb3 Rg3+ 50.Kc4 h1=Q 51.Rd8+ Kg2 52.a8=Q Qc1+ ...

49.Ka3 looks too slow. Maybe 49.a7 Rh8 50 Rxh3 Ra8 51.b4 Rxa7+ 52.Kb3 g5 53.Rg3 f5 and White is still in the fight. Can't say I like his chances, though. *** Instead of 50...g3. 50...Rh8 or 50...f5 would have sealed the deal.

Maybe Black ought a few moves back to have gone for the quick lunge... 54...f4 or even 54...h2, putting immediate pressure on the K-side. White would then have to solve the problem of maintaining the h-file blockade and dealing simultaneously with the advancing f-pawn. It seems, though, that White has sufficient: 54...f4 55.Kb2 h2 56.Kc3 f3 57.Kd3 f2 58.Ke2 and wins.