Peligroso Patzer: <Jason Frost: *** 43...e6 (in the actual game) is just a bad blunder, missing that after 43...e6?? 44. dxe6 Qxd3+ 45. Qxd3 Rxd3 46 e7! the pawn cannot be stopped from promoting. *** >
The variations are somewhat more complicated than the above comment suggests, but <43. ... e6?> was definitely a blunder (as was much-discussed on the tournament page: Biel International Chess Festival (2009)).
The following lines show that after <44. dxe6!>, Black’s position was going to be difficult or outright lost in all variations:
<44. ... Kg7>
(I) Worse would be: 44...Rxd3 45.Qc8+
(Also winning for White would be: 45.e7 Bc7 46.Qxc7 Qe4 47.Ka2 Qa8 48.Qe5 Qe8 49.b3 a4 50.bxa4 , but 45. Qc8+ is the most accurate winning move.)
(If here 45...Kh7 then 46.exf7 (but note that in this line pushing the e-pawn to the 7th does not work: 46.e7?? Rb3 and BLACK wins!).)
46.e7 Rb3 47.Qf8+ Kh7 48.Qh6+ Kg8 49.e8=Q (or R)#;
(II) Also losing outright (after 44. dxe6!) would be: 44...Qxd3+ 45.Qxd3 Rxd3 46.e7 (the line given by <Jason Frost>).
< 45.e7 Re8 46.Qd5 Rxe7 47.Re3 Qh2 48.Bd2 f6 49.gxf6+ Kxf6? >[Better would be: 49...Bxf6 50.Rxe7+ Bxe7 51.Bxa5±] <50.Qf3+ Kg7> (There is more resistance in 50...Qf4, but Black is still losing the exchange-down ending after: 51.Qxf4+ Bxf4 52.Rxe7 Bxd2 .) <51.Bc3 Qg1+ 52.Ka2 Qg5 53.Rxe5 Rxe5 54.Qd5 Kf6 55.Qd6+> 1–0