http://tws27.blogspot.com/2014/08/the-cochrane-gambit-vs-petroff.html
Cochrane Gambit as Black.

I think the remedy involves g6 & Kg7 but this is murder OTB in any event. When you can look up lines on computers, the advantage entirely dries up.

When untangling the initial knot you get in, try to keep the lookout for the helpful ..d5 clearence sacrifice. It's especially useful for dealing with the common Bc4 in this line.

i love the cochrane gambit!! its the last thing every petroff player is looking to deal with..and my favorite part about playing it is the look on their face after the immediate sac on f7. I personally like to grab my N and motion as if im moving back to f3 as per main line,,, and before putting it down stomp it down on f7 ;)
Yes, sometimes I lack class.

I dunno man. In all my experiance, and from what I've read, as long as black doesn't fall asleep at the wheel the best white can hope for is equality. And once the inital mess clears, it's really only black having the fun in the position.

yeah, thats assuming black is an engine. when a human is sweating on the other side and isn't rated 2300+, its another story ;)

I play the Petrov and I have a good record against the Cochrane Gambit. It's supposed to be close to sound for White, which means Black's material advantage is long-lasting and important. Some advice for Black:
Remember your goals of piece placement: ...Be7, ...Rf8, ...Kg8, usually in that order.
Learn a little ways into the main variations, especially an early ...c5 followed by ...Nc6, which seems to be intended to return the pawn for positional equality or even as a countergambit.
Meet Bc4+ with ...d5 if you can, or ...Be6 if necessary, and don't worry too much about your king getting centralized at e6 after Bxe6+ ...Kxe6.
After the standard Bc4+ ...Be6 Bxe6+ ...Kxe6 exchange, when you get a free move, start retreating your king to f7, headed for g8.
Here are some examples of some of the main variations that show my suggestions in operation...

I have only recently heard about this gambit, but I keep getting wildly varying remarks about how sound it is. How sound is the cochrane gambit compared to the king's gambit, danish gambit, or even the halloween gambit? So far, without fail, I have see. people do nothing but laugh at exchange sacrifices in the opening. But sometimes when I see this gambit people seem to take it really seriously like it's a legitimate opening that GMs use.

How sound is the cochrane gambit compared to the king's gambit, danish gambit, or even the halloween gambit?
Below is what Schiller said about it in 1998.
Note in my notes about the moves in my above posted book lines that the percentage of wins by Black (in my notation it looks like "b 80%.") tends to be high, once Black gets beyond the 5th move safely/correctly. Until then, White's win rates are higher, probably only because Black wasn't expecting the gambit, so the losses by unprepared Black players got averaged in with the wins from proper continuations. The statistics lead me to believe that Black has a slight advantage in this gambit, with proper play by Black.
----------
(p. 376)
RUSSIAN GAME
COCHRANE GAMBIT
rnbq1b1r/ppp2kpp/3p1n2/8/4P3/8/PPPP1PPP/RNBQKB1R w - - 0 1
1.e4 e5
2.Nf3 Nf6
3.Nxe5 d6
4.Nxf7 Kxf7
The Cochrane Gambit is an ancient opening which has found a
new following in the 1990s. White sacrifices a whole piece to expose
the enemy king. Anyone interested in this opening should certainly
pick up W. John Lutes' thorough study of the opening. The sacrifice
was worked out by John Cochrane, who used it in a number of games
against Mohescunder Bnnerjee [sic] in Indai [sic] in the late 1840s. The opening
remained in obscurity for some time, because it was generally felt that
a piece was just too much to gambit in the opening.
In recent years it has been reconsidered, and a lively debate has
taken place both on the chessboards of the world and on the Internet.
White can follow the sacrifice with three distinct plans. 5.Bc4+ is the
most obvious, but taking the center with 5.d4 and developing with
5.Nc3 also have a following. Here is some recent theory on each
of the lines, but it is only a taste of the vast menu of attacking options
offered by the Cochrane Gambit.
EXAMPLES AND DISCUSSION
5.d4.
5.Nc3 is not popular now. 5...Qe7 (5...c6; 6.d4 Qe8! and play transposes to
5.d4.) 6.d4 g6; 7.Bd3 Bg7; 8.Bg5 h6; 9.Bh4 g5; 10.Bg3 Rf8; 11.O-O Kg8 and
White does not have enough for the piece, Svenn - Maiorov, Hallsberg 1980.
5.Bc4+ d5!? (5...Be6; 6.Bxe6+; 7.d4 Kf7; 8.Nc3 Be7; 9.Qe2 Rf8; 10.Be3
Kg8 and Black had achieved a safe position, though White has some compen-
sation for the piece in the form of two pawns and a strong center. Gossip -
Birmingham Chess Club, Postal 1860.
(p. 377)
5...Ke8; 6.O-O c5; 7.h3 Qc7; 8.f4 Nc6; 9.Nc3 a6; 10.a4 Qe7; 11.Nd5 Qd8;
12.d4 cxd4; 13.e5 Nxd5; 14.Bxd4 dxe5; 15.Bxc6+ bxc6; 16.Qh5+ Kd7; 17.fxe5
Kc7; 18.Rf7+ Kb8; 19.e6 Bd6; 20.Bg5 Qb6; 21.a5 Qc5; 22.b4 Qe5; 23.Bf4 Qxe6;
24.Qc5 Qxf7; 25.Bxd6+ and Black resigned in Cochrane - Mohishunder,
Calcutta 1848.)
Schiller, Eric. 1998. Unorthodox Chess Openings. Cooper Station, New York: Cardoza Publishing.
----------

I have nevere lost a cochrane gambit as white in classical play ;). I've admittedly only played 3 games, and never an opponent above 2000, but still I've had a blast with it and have played it differently each time. 2 of the games I played the questionable 5.bc4?! to which both were my quickest wins. It's a bit of a gamble I suppose if black is well prepared to play the cochrane. and yes, pfren is correct- the most testing response is the early ..c5.
I've studied the positions a bit, but i've been lucky enough to never actually see that move order in OTB play.
The main point is, if you are the type of player that likes to be in control of play (not so much on the defensive) it is a great gambit to play. It is not the soundest response, but it comes down to personal style as always.

There's one thing I don't quite understand about this attack: Why does White play Bc4+ in nearly every game? Is the idea merely to keep the initiative by constantly attacking, or is the idea to lure Black's king into the center via ...Be6 Bxe6+ ...Kxe6 with the hope that it can be more easily mated there? Or both ideas together? It looks like White rarely gets any advantage from that exchange, so it makes me wonder why White even tries.
As for the 5. d4 c5 6. dxc5 Nc6 attack, I was curious and I found only one game in the database I'm using where White played an early ...cxd6, and this is what happened:
I have been using Petrov's Defence as Black for a while now, and I (usually) like playing it. However, I absolutely cannot stand facing the Cochrane Gambit:
It's so annoying! How do you deal with this? I really don't want to give up the Petrov because of this variation!