I fired up my FritzX and had it look at the original game. Strangely enough, Fritz does play 2. . d6 against 2 d4. I was a bit puzzled at this at first, but when I went into the opening database this is Fritz's "book" response to 2 d4. But when I put in the two variations (2 . . d6 and 2. . Nxe4) and had Fritz do a positional analysis of each, it clearly likes Nxe4 better, feeling that black has more than equalized at that point. Fritz misses the opportunity, not because 2 d4 is strong, but because of a limitation in its opening book. As others have pointed out, in actually master's play 2 d4 scores very poorly. I am all for challenging conventional wisdom, but on the whole the evidence shows this is a very suspect gambit line.
A New Gambit in the Alekhine Defense

Here's my analysis of the original game. The cliff notes version is that I feel black had a superior position pretty much from move 2-3 onwards, and at no point was white genuinely threatening. I also pointed out a number of (relatively) poor moves by black.
This brings up two points. 1: If black can play that many sub-par moves and still reach a winning endgame then there's something fundementally wrong with white's plan.
2: I hate to accuse anyone of lying/cheating... but did the Fritz really play black here? I don't own a copy of the Fritz, but I fed the position on move 32 into the engines I do have (I had Houdini, Pro Deo and Stockfish loaded into arena at the time). All 3 of them spotted 32... Qd5 in under two seconds (I timed it!). I had noticed the move myself when trying to analyze black's 31 ... Rd1. I find it highly implausible that a GM level program could miss a 4 move deep tactic.
The problem is with the score of the game; I can't get my computer to print out the games from the computer. All I get is the blank scoresheet. I write the moves by hand on the score sheet, thus there is an error in notation somewhere. You are absolutely right that Fritz wouldn't miss that tactical shot, so there is something wrong with the score of the game towards the end. Sorry about the confusion.
While I thought that White has compensation for the pawn after 1.e4 Nf6 2.d4 d6 3.Nc4 Nxe4, what is amazing to me is that Fritz 10 not only says that White has compensation for the pawn, but demonstrates with plausible moves (I analyzed this position at about one move every 5-10 seconds, or at a level, where it would not miss any obvious tactical shots i.e. adequate for your typical class player) that White can even
transition to a favorable endgame. What is remarkable is just how good Fritz with White pokes and prods the position making little threats. With Black it sees the obvious threats but suffers from the horizon effect. It should jettison the extra pawn and steer for the draw. But just like a human professional Fritz stubbornly clings to the extra pawn only to lose it anyway with an inferior endgame.

The problem is with the score of the game; I can't get my computer to print out the games from the computer. [snip] You are absolutely right that Fritz wouldn't miss that tactical shot, so there is something wrong with the score of the game towards the end.
So your notation was messed up for at least 16 moves, and yet every single one of them was somehow still legal? That's got to be one of the most awful chess-related lies I've heard since someone tried to convince me he played the Sicilian... in a position where there were still pawns on e2 and c7!
Certainly in a game against other players rather than computers, I would never gambit the pawn, unless I was just playing for fun. I don't think that there is adequate compensation for the pawn (My intuition tells me there has to be some good way to give back the pawn and emerge with a plus)
I thought you were mad but I had stockfish play a 5min game against itself from your starting position, and not only does it think the initial position is equal but it drew in under 20 moves! If it weren't for the problem of 2.Nxe4 your opening seems perfectly playable.
Here is another gambit in the Alekhine Defense. Fritz 10 thought that White had a big edge, but I had to settle for the draw (50 move rule). For some reason White's 5.c4 was not in Fritz's opening book. This is something of a mystery because it is an obvious continuation. Then Fritz made a huge tactical mistake. It is amazing to me that Fritz would step into this obvious tactical shot.
This is the game Moody-Bullock, Postal 1994.