Hello,
If you play this "defense" in Blitz games, chances are you'll have a lot of successes, especially if you know the lines and your opponent wants to punish you for daring to play such moves against him. Often, chess is about ego..., not elo !
Actually, the "serious" players could play it as it is real chess. I mean the "serious player" is mainly reciting openings and variations and most of the time, you gotta wait the 2Oth move to really start playing. The advantage of the Fred is you start playing as of the very first move.
I personally call 1.e4 f5 2. exf5 Kf7 the "drunk man defense". I like chess engine games and I have built up a database dealing with that opening. The games are surprisingly very long when Hiarcs or Deep Shredder, for instance, are confronting each other. This is why I think the "serious players" (id over 2800 ELO) could give it a try. Here's a most recent exemple. Deep Rybka 4 (White side) is defeated :
Deep Rybka 4 - Stockfish 1.7.1 0-1
Drunk man defense tournament
5/2010
B00 - Irregular Opening
1.e4 f5 2.exf5 Kf7 3.Qh5+ g6 +2.82/18 4.fxg6+ Kg7 +2.86/19 5.gxh7 Rxh7 +3.11/20 6.Qe5+ Nf6 +3.23/18 7.Bd3 Rh6 +3.31/20 8.Qg3+ Kf7 +3.11/18 9.Nf3 Nc6 +3.27/16 10.h3 e6 +2.30/17 11.Nc3 Bd6 +2.02/19 12.Qg5 Qh8 +2.06/18 13.a3 b6 +1.33/17 14.Nb5 Bb7 +1.01/16 15.Bf1 Rh5 +0.80/15 16.Qe3 Ne5 +0.28/25 17.Nxd6+ cxd6 +0.20/14 18.Ng5+ Rxg5 +0.00/19 19.Qxg5 Rg8 +0.00/21 20.Qf4 Rxg2 +0.00/22 21.Be2 Rg8 -0.52/19 22.Rf1 Qxh3 -0.24/19 23.f3 Rg2 -2.58/18 24.d3 Ke7 -2.38/21 25.Qe3 Nh5 -4.04/18 26.Rf2 Qh4 -4.92/19 27.d4 Rg1+ -5.13/18 28.Kd2 Ng3 -5.45/19 29.dxe5 Nf5 -5.49/20 30.exd6+ Kf7 -5.49/20 31.Qe5 Qxf2 -5.69/19 32.Kc3 Re1 -5.73/19 33.Bg5 Rxa1 -5.57/19 34.Qf6+ Kg8 -5.93/18 35.Qd8+ Kh7 -5.73/20 36.Qxd7+ Kg6 -5.93/20 37.Qxe6+ Kxg5 -5.81/19 38.Qg8+ Kf4 -5.93/19 39.Qc4+ Ke5 -5.97/19 40.d7 Qe1+ -6.06/17 41.Kb3 Nd4+ -5.81/18 42.Qxd4+ Kxd4 -6.06/19 43.d8(Q)+ Ke3 -5.81/18 44.Bd3 Bxf3 -6.14/17 45.Qe7+ Kf2 -6.18/19 46.Qh4+ Kg2 -6.18/22 47.Qg5+ Qg3 -6.26/20 48.Qd2+ Kh3 -6.26/22 49.Qh6+ Qh4 -6.26/19 50.Qd2 Kg3 -6.58/19 51.Qc3 Qf4 -6.62/19 52.Qg7+ Kf2 -6.86/22 53.Kc3 Rd1 -7.03/20 54.Qe7 Rb1 -7.39/21 55.Qg7 a5 -7.79/19 56.Qe7 a4 -8.00/20 57.Qg7 Qe3 -8.52/24 58.Qd4 Qxd4+ -9.85/24 59.Kxd4 Rxb2 -10.14/26 60.Kc3 Ra2 -15.79/26 61.Bb5 Rxa3+ -21.17/23 62.Kb4 Ra1 -84.20/27 63.Bxa4 Rb1+ -89.14/30 64.Bb3 Ke3 -128.76/27 65.Kc3 Rxb3+ -93.67/29 66.cxb3 Be2 -#19/32 67.Kc2 Kd4 -#17/33 68.Kd2 Bd3 -#15/24 69.Ke1 Kc3 -#14/33 70.Kf2 Kxb3 -#13/31 71.Kg1 b5 -#12/18 72.Kf2 Kc4 -#12/15 73.Ke3 b4 -#10/14 74.Kf4 b3 -93.52/9 75.Ke5 b2 -#8/28 76.Kf4 b1(Q) -93.67/8 77.Ke5 Qb2+ -#6/29 78.Ke6 Qg7 -#6/7 79.Kd6 Bf5 -#3/100 80.Kc6 Qa7 -#2/100 81.Kd6 Qc5# -#1/100
As you can see, the "serious player" can play up to 81 moves and defeat the white. Be it deep rybka. But you gotta be a serious player !!!
Wow, if those games are real, it's funny how the computer evaluation goes from "totally winning for me" to "mate in 4 for my opponent." Seriously!! 
There may be some psychological point to this idea, 1e4 f5!? I just checked the stats in this sites Game Explorer. Amazingly, only two players replied 2.exf5! ( they both won). Most tried 2.d4, with an overall losing score. A lot also tried 2.Nf3!? This actually looks kind of interesting after:
A couple of comments:
2.Nf3 transposes to the Lisitin Gambit, which I agree is interesting. 3.Ng5 is more common, for example 3...Nf6 4.d3 with compensation. If you're going to play 5.Be2 then why not 4.Be2 right away? In your final position after 6...g6 7.Nxg6 hxg6 8.Bxg6+ Kd7 followed by ...Rg8 looks better for Black.