Revisiting the Scotch Gambit: the Two Knights Defence with d4

I'm currently writing a Lichess study on the Scotch Gambit/Two Knights Defence position that arises after 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.d4 exd4 (or 3.d4 exd4 4.Bc4 Nf6).

I think this is especially worth revisiting because there have been quite a few new developments since I last examined the line in some depth, mainly in the 5.0-0 lines:

(a) 5.0-0 Nxe4 6.Nc3?!, the Nahkmanson Gambit.  

Until recently I had rejected this out of hand, but I was first introduced to one of the key points behind this a couple of years ago at my local chess club (6...dxc3 7.Bxf7+, with the idea of Qd5+, and then usually Re1 followed by recapturing on e4).  It's not very sound, largely because Black can settle for a one-pawn advantage after 6...Nxc3 or 6...Nd6, so I can't recommend it as a serious tournament weapon.  But it has been gaining some popularity at online blitz in particular, where it offers White plenty of hacking chances, especially if Black accepts the piece sacrifice.  Thus, in the study that is currently work in progress I have devoted a chapter to this line.

(b) 5.0-0 Nxe4 6.Re1 d5 7.Bxd5 Qxd5 8.Nc3 Qd7!?.

I don't think I have previously covered that line in much detail, but it has emerged as a serious alternative to the main lines (8...Qa5, 8...Qh5).  It aims to improve over the 8...Qd8 line (Black probably isn't fully equalising after 8...Qd8 9.Rxe4+).  Graham Burgess gave it a good look in his updated version of Chris Baker's A Startling Chess Opening Repertoire.  After 8...Qd7 9.Nxe4 White should be no worse, as White can put the knights onto active posts so that they aren't inferior to Black's knight and bishop, but it is trickier to generate winning chances.

The Max Lange line with 5.0-0 Bc5 6.e5 d5 7.exf6 dxc4 8.fxg7 is now looking roughly equal with best play, although I'd still rather have White in that line, and it's still looking better than the old main line 8.Re1+ Be6 9.Ng5 Qd5.

Via a search of the ChessBase database, I have also found a new try for White in the main line of the Canal Variation (5.0-0 Nxe4 6.Re1 d5 7.Nc3 dxc3 8.Bxd5 Be6 9.Bxe4 Qxd1 10.Rxd1 cxb2 11.Bxb2 f6 12.h4) which complicates matters and so could be a good practical try for rapid games and especially online blitz, though I still think Black's better with best play in that variation and that 7.Nc3 is therefore not as reliable as 7.Bxd5.

My fresh investigations into the 5.e5 lines are only just starting, so I'll have to see if anything new turns up in those.