Black's last chance was 9... Qd8. After 9... Qh5? Morphy or Capablanca would have played the natural move 10.Nxf4 Qg4 11.d4 .. or 10.d4 .. first with an easy win (the black Queen is trapped). The gaudy 10.Nf6+? .. throws away most of his advantage, but is given an ! in Donaldson's book - I suppose that's what meant with "deeply annotated by Alekhine".

There was no reason for Euwe to resign yet, although I agree white has still better chances in the final position. After 12... c6 13.Qe8+ Kc7 14.Bxf7 ... my earlier suggestion 14... Nxf6 is met by 15.Qe7 Qf5 16.Re5 .. or 15... Qh6 16.Re6 .. and in the line posted by <who> white can improve with (after 14... Qf5 15.Bxg8 Qxf6) 16.Re5 .. first, attacking Ba5 and facilitating a later d4 .. (ultimately winning pawn f4 or h7).

I heard Euwe was drunk! ;->

"I presume this is the <(1)> shortest decisive game <(2)> between players who (subsequently or previously) played a match for the world championship <(3)> and/or both held the title?"

Condition <1> is easy to understand, now that forfeits and defaults have been excluded.

In Condition <2>, "players who ... played a match for the world championship" is ambigous; it doesn't necessarily mean they played each other. However, <Domdaniel> has since stated that that such was his intent.

Also, the clause "subsequently or previously" implies that the game to be found was <not> played in the WC match between the two players.

Condition <3> is also confusing. "Or" seems to be unnecessary, as no player has yet held the title without playing a match. (Karpov won the title by default, but played title matches later.)

Of course, "World Championship" needs to be defined. If it starts with Steinitz in 1886, then we are restricted to games among these players:

Steinitz--Lasker; Lasker--Capablanca; Capablanca--Alekhine; Alekhine--Euwe; Botvinnik-Smyslov; Botvinnik-Tal; Botvinnik-Petrosian; Petrosian-Spassky; Spassky-Fischer; Karpov-Kasparov; Kasparov-Kramnik

(I don't know whether to count all the recent "champions" who won their titles in tournament play.)

If the world championship goes back to 1851, Andersson--Morphy and Anderssoh--Steinitz could be added. Beyond that--in both directions--soneone else can decide.

Finally, the original statement made no restriction against "rapidplay" events. Should this be restricted to "serious" tournament/match play?

So my example is legal, as would be a (serious) game between Steinitz and Kasparov, but not a game between Steinitz and Short, because Short, although he played a match for the world championship, never held the title.

Anand played and lost a match in 1995, but some people claim he later somehow won the title (without defeating the incumbent champion in a match). Would that make a miniature between Anand and Steinitz valid?

Alekhine vs Euwe, 1921
Euwe vs Fischer, 1957
Tal vs Petrosian, 1962

If you want to include Morphy and Anderssen, then the Morphy vs Anderssen, 1858 game which you pointed out earlier would be a clear winner at 17 moves.