The Urusov Gambit most often arises from the Bishop's Opening: 1.e4 e5 2.Bc4 Nf6 3.d4 exd4 4.Nf3, but Scotch/Göring Gambit aficionados can also enter the line via 1.e4 e5 2.d4 exd4 3.Bc4 Nf6 4.Nf3, or 1.e4 e5 2.d4 exd4 3.Nf3 Nf6 4.Bc4.
At the Urusov Gambit section of my site I have chosen to give more of an "outline" type of coverage,
highlighting the key lines and deviations and suggesting what I believe
to be best play for both sides, while again offering some annotated illustrative games. For those after a thorough analysis I recommend that readers have a browse of Michael Goeller's Urusov Gambit site, while Danish Dynamite contains some good analysis of the accepted lines of the gambit.
Some brief comments on the declined variations:
1. Black's most reliable way to decline the gambit is with 4...Nc6 transposing to the Two Knights Defence, but 4...Bb4+ is also quite critical. Here I recommend that White boldly sacrifices two pawns with 5.c3 dxc3 6.0-0, since 6.bxc3 d5 is strong for Black and I don't trust the line 7.Qa4+ Qd7 8.Qxb4 dxc4 9.Ba3 Nc6 10.Qxc4 Qe6.
2. I am in strong agreement with Goeller that 4...d6 should be met by heading for a line of the Antoshin Variation of the Philidor Defence, with 5.0-0, followed by Re1 and Nxd4 in most cases. My investigations into this line have suggested that it is promising for White. 4...c5 should also be met by 5.0-0, and if 5...Nc6 then 6.Ng5 intending f2-f4, and 6.Re1 intending c2-c3, are both sufficient to give White a theoretical edge and attacking chances. However, 4...c5 is not as bad as it looks and should not be underestimated. Finally, 4...Bc5 should be met instead by 5.e5 intending 5...d5 6.exf6 dxc4 7.Qe2+ Be6 8.fxg7 Rg8 9.Bg5 as recommended at Goeller's site.
3. I don't think 4...d5 5.exd5 Bb4+ equalises. My main recommendation against this is 6.c3 Qe7+ 7.Be2 dxc3 8.Nxc3 rather than the more popular 8.bxc3. Estrin and Panov's recommendation against this, 8...0-0 9.0-0 c6, does not appear to equalise after 10.a3 forcing the b4-bishop away to an inferior square. White can also get an edge with 6.Kf1, which wins a pawn in most lines, but White's king ends up misplaced and holding onto the extra pawn at d5 can be tricky, so it is probably not as easy to play.
And onto the accepted variations:
1. After 4...Nxe4 5.Qxd4 Black has no good alternative to retreating the knight with 5...Nf6. Then, I am in agreement with Goeller and others that 6.Bg5?! is inaccurate as it fails to cover the d5-square and thus runs into 6...Nc6 intending 7.Qh4 d5.
2. 6.Nc3 Nc6 7.Qh4 Bb4 is a tricky sideline as the threat of shattering White's queenside pawn structure forces White to change plans and castle short, and this gives Black greater scope to play ...h7-h6, inviting a Bxh6 sacrifice, without suffering an immediate disaster. See D.Grobler-J.Antal, email 2011.
3. The main lines all seem to be holding up well for White. The main line of the Urusov Gambit Accepted is probably 6.Nc3 Be7 7.Bg5 Nc6 8.Qh4 d6 9.0-0-0 Be6 10.Rhe1! Bxc4 11.Qxc4 0-0.
A key idea in many lines is the Rxe7 exchange sacrifice which undermines the protection of the knight on f6. Carlos Torre (best known for beating Emmanuel Lasker with the Torre Attack, using 1.d4, 2.Nf3 and 3.Bg5) tried this sacrifice immediately at move 12 and won a fine attacking game with it, though it is probably better for White to defer the sacrifice until a better moment (e.g. after 12.h4 h6?! 13.Rxe7! and if 13...hxg5 then 14.hxg5!)