Years ago I would try to slip into a Blackmar-Diemer after 1.d4 Nf6 with 2.f3, looking for an obliging 2...d5 3.e4 dxe4 4.Nc3. But Black has too many good replies, such as 3...c5, so these days I almost never play this line. However, a couple of decades ago in a rated OTB game I was surprised with another third move by Black. (From the August 1988 issue (Nr 32) of BDG World).
Purser, Tom - Watson, Lavelle
Ramstein, Germany, 1978
1.d4 Nf6 2.f3 e6 3.e4 Nxe4
4.fxe4 Qh4+ (Anyone who plays an early f3 ought to be alert to this idea.) 5.Ke2 Qxe4+ 6.Be3 b6 7.Nd2 Ba6+ 8.Kf2 Qh4+ 9.g3 Qf6+
10.Qf3 Bb7 11.Qxf6 gxf6 12.Bg2
And I stumbled on to win in 47 moves. I thought this was a unique opening (had you seen it before?). Imagine my surprise when months later I discovered a game in Diemer's Blackmar Gemeinde which was an exact duplicate for the first nine moves. In fact Diemer had even put a name on the opening: "Dr. Willy Linder Gambit." Perhaps "willy nilly" would be better.
Diemer,EJ - De Wolf
Gent, 1956
1.d4 Nf6 2.f3 e6 3.e4 Nxe4 4.fxe4 Qh4+ 5.Ke2 Qxe4+ 6.Be3 b6 7.Nd2 Ba6+ 8.Kf2 Qh4+ 9.g3 Qf6+
To this point, identical with Purser-Watson, above. 10.Ngf3 Bb7 11.Bd3 h6 12.h4 d5 13.Qe2 Rg8
(White was threatening 14.Bg5) 14.Rad1 Bd6 15.Kg2 Qe7 16.Bg1
("Under all circumstances e5 must be prevented, and pressure kept on f7."-- Diemer) 16...c5 17.Bb5+ Kd8 18.c3 a6 19.Ba4 b5 20.Bc2 Nd7 21.Re1 Kc7 22.a4 Bc6 23.dxc5 Nxc5 24.Nd4 bxa4 25.Nxc6 Kxc6 26.c4 Qf6 27.Nf3 Nb3 28.Bxb3 axb3 29.cxd5+ exd5 30.Nd4+ Kb7 31.Qd3 Qg6 32.Qxb3+ Kc8 33.Qc3+ Bc7 34.Re7 Ra7 35.Bf2 Qf6 36.Rhe1 Rb7 37.Qc5 Rd8 38.Nc6 Bb6 39.Na7+ Kb8 40.Rxb7+ Kxb7 41.Re7+ 1-0
[Black resigned, for as Diemer put it, "he only had a choice between different mates: 41.Re7+ Ka8 42.Qxb6 Qxe7 43.Nc6 Qe4+ 44.Kh2! " This game is from the tourney in which Diemer finished second behind O'Kelly, losing to the grandmaster in the final round.]
Play through and download these two games plus a few more in this variation here.