When Is a Game or a Position a Foregone Conclusion?

Continuing with Lasker's Moves that Matter, the diagram shows the continuation of Evans Gambit: Chigorin - Lasker, St Petersburg 1895. The first post has a link to the complete game on Chessgames.com.

There are two types of flaws in analysis that annoy me. One is assuming that the result of a game is a foregone conclusion before it is played. For example, when Chigorin played Lasker, the German master won because he was the better player. The other is assuming that a position leads to a foregone conclusion because it is obvious. For example, Lasker won from the diagrammed position, and there was nothing Chigorin could do to change the result.

St Petersburg 1895
Lasker, Emanuel

Chigorin, Mikhail
(After 11.Bb6-a7)
[FEN "r1bqk2r/b1p2ppp/p1pp1n2/P3p3/3PP3/2P2N2/5PPP/RNBQ1RK1 w kq - 0 12"]

The game continued 12.dxe5 Nxe4 13.Qe2. After White's 13th move, the game was lost. Would 12.Qa4 have drawn? What about 13.exd6? At what point was White's game irretrievably lost?