I find that the 3.Nc3 lines of the French Defence often lead to crazy and rich positions, although of course 3.Nc3 is one of the main lines, and many of the variations are quite theory-heavy. The various attempts to steer play into a sort of pseudo Blackmar-Diemer Gambit with 3.Be3 and 3.c4 are not convincing, although there is a subvariation of 3.Nc3, 3...Nf6 4.Bg5 dxe4 5.f3!?, which aims to transpose into a line of the Blackmar-Diemer (Euwe Defence) which probably gives White full compensation for the pawn.
The Alekhine-Chatard Attack is one of the soundest gambits that I've looked at so far; indeed French Defence guru John Watson considers that it is holding up well at high levels. I've been a fan of this gambit for many years, and recall having quite a few nice wins with it as a junior in the late-1990s.
I have only got around to covering 6...c5 and 6...0-0 so far, but am trialing out a new way of displaying the coverage (sort of like a ChessCafe.com article but with the games still presented as replayable java games via ChessBase).
The coverage is here:
http://tws27.weebly.com/alekhine-chatard-attack.html
My overall assessments seem to broadly agree with Watson's comments on 6...c5 and also 6...Nc6, which I cover briefly as a sideline, although of course Watson will have gone into far more detail.
I had a recent game as White in the Alekhine-Chatard Attack but unfortunately lost the game score. I managed to crash through on the h-file by putting rooks on h1 and h6 and a queen on h4, and breaking through on h7, and Black's counterattack ended up being a tempo too slow. (If I remember rightly, Black met 6.h4 with 6...c5 7.Bxe7 Kxe7, and later moved the king over to the kingside to guard h7).
Meanwhile I've recently received my copy of Smerdon's Scandinavian. It is refreshing to see a grandmaster frequently using and being enthusiastic about a line that is objectively of marginal soundness. I'll be looking at his Caro-Kann transposition lines (1.e4 d5 2.exd5 Nf6 3.c4 c6 4.d4, involving an early ...g6) with some interest since I didn't look at those when I last covered the line. Of course White should avoid 4.dxc6?! Nxc6 in that variation; I remember a few games when I tried that greedy variation as White at the local chess club just to see if it was really as bad as its reputation, and inevitably I got crushed every time.
Smerdon also recommends the Vienna Defence (1.e4 d5 2.d4 dxe4 3.Nc3 Nf6 4.f3 Bf5) against the Blackmar-Diemer Gambit, and is quite dismissive of White's chances. Personally I always thought 4...exf3 5.Nxf3 Bf5 was a more serious test of the gambit, but I'll be looking at his lines more closely shortly to see if he's found any major improvements for Black over what I know of.
