The fun killer:
After taking on e4 Black can simply move the knight back to f6 to avoid this variation. As far as I can tell, there is little better for White than simply playing Nf3 and going for a regular BDG after this.
However, this move order does prevent the O'Kelly Defense and the Lembergher Defense, which are two of the biggest problems White faces, so it is a victory in some ways. If it can be shown that White has something better against 6...g6 by playing Bc4 or Bg5 before Nf3 then this could be a major theoretical breakthrough for the whole Blackmar complex--as much as Diemer's 3.Nc3 was. If not, then oh well, it's a surprise weapon to add to the collection.
1.d4 Nf6 2.f3 d5 3.e4 dxe4 4.fxe4!?
I have been working on a new gambit that as far as I can tell has never been explored or named. A few people may have fallen into it mistakenly in the database but I see the potential for something good here.
For now I wish to call it the Lemur Gambit, or the Blackmar-Lemur Gambit.
With 4.fxe4 Nxe4 5.Nc3 Nxc3, borrowing an idea from the Stafford Gambit, we reach a position where I believe White has sacrificed their pawn under more favorable circumstances. Note - a knight that normally defends the Black king is traded with a queenside knight that often does not participate in the attack. White's d-pawn which is often the subject of counterplay for Black has been shored up. And while we have the same initiative as the Stafford, we have an extra f-file to work with which makes it extremely hard to defend.
An important declination. 4...e5 is critical. I misplayed my first attempt at it. 5.Bc4 was an idea from the Caro-Kann Fantasy that made sense at the time but 5.Nc3 was much better.
I'll have more to add on this. I have some important variations to cover. For posterity I'm just putting it all here.