Well, let me deal with your proposal. 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 c6 O´Kelly defence. This is the line which also Eric Schiller is recommended on his homepage here:
http://www.chesscity.com/PDF/BDG_Okelly.pdf The best way according to my opinion is 5.Bc4! and now black has several possibilities:
A) 5..e3?! isn´t really a challange as after 6.Bxe3 white is better.
B) 5..exf3 6.Nxf3 transposes into the Ziegler defence.
C) 5..Bf5 is well playable, but white can fight for advantage after 6.fxe4!? and you can find some analyses on this line at ED web here:
http://www.mujweb.cz/www/rajmunde/ec_ena.htm The analyses are quite old and I know about some places where they does not work so they need to be checked carefully but a main idea is perhaps correct.
6.g4?! as alternative is not so good. The position is about even then. By heart 6..Bg6 7.g5 Nd5 8.fxe4 Nxc3 9.bxc3 and now e5!? if I remember well was recommended by IM Wahls in his Scandinavian defence. Also 9..Bxe4 followed by e6 is o.k. for black.
D) 5..Qa5 is recommended by Schiller. First, when he published his article, he simply has an advantage for black. Then Rajmund sent him his line with 6.Bd2!N and in his 2004 update he has following "6.Bd2!? Qb6 7.Qe2 Qxb2 8.Rb1 exf3 9.Nxf3 Qa3 gives black two pawns, but white has a huge lead in development and there is quite a bit of compensation" (actually, I think, black is lost). Then he stated that 7..Qxd4? 8.0-0-0 is "too risky for black". So he proposes 6..Qf5!? with "is perhaps best". Well, for me, his attitude is very unconvincing. He gives 5..Qa5 an exclamation mark but when dealing with subvariation starting with 6.Bd2, he proposes only lines which actually favour white.
So, for now, I consider 6.Bd2! to be refutation of his line.
That´s all to the O´Kelly system. As you see, the only real challange is 5..exf3 leading to the Alchemy Variant and 5..Bf5.
Patrik Schoupal