I liked to know more about the ideas behind 6.Nf3 a6!? for Black.  During my brief use of the Caro-Kann I had planned to play 6.Nf3 Be6 (and 6.Bg5 Be6 but maybe 6...dxc4 is actually the best way to go) but I certainly thought White was at least +/= with best play.  Then again I really only looked at 6.Nf3 Be6 7.c5 g6.

This is good news for the adherents of the Scandinavian Marshall Gambit too. One question:
why is x...g6 good against the Pseudo Panov, but not against the Panov proper? I always thought, White could simply transpose after 1.e4 c6 2.c4 d5 by playing d2-d4 a bit later, eg 3.exd5 cxd5 4.Nc3 Nf6 5.cxd5 Nxd5 6.Nf3 Nc6 7.d4.

I´m playing the Caro-Kann now for 15 years and I´ve always thought that the panov attack should be countered by the active and natural 5...Nc6!

The nimzo-like 5...e6 6.Nc6 Bb4 may be the best choice on grandmaster level, but in the transposition to the nimzoindian 7.Bd3 0-0 8.0-0 dxc4 9.Bxc4 b6 (Karpov system) Black usually exchanges his bishop on c3. And as most C-K-players I do not matter exchanging by queens bishop for a knight but I really love my kings bishop (on e7 or d6)! - The calm 5...e6 6.Nf3 Be7 is too passive though not lost as many books suggest. - And there´s also the problem of 5...e6 6.c5 gaining more and more supporters recently. Summa summarum I don´t like 5...e6
Systems with g6 are good against 2.c4, but in the original panov move order white is simple better if he knows when to return the sacrified d-pawn with d5-d6.

Back to 5...Nc6. In my opinion 6.Bg5 dxc4! (Be6 is an alternative) is not problem at all for black neither after 7.d5 Ne5 8.Qd4 h6! nor after 7.Bxc4 Qxd4! when after a short firework it is white who is struggling for equality. 
The problem is 6.Nf3! when 6...e6 and 6...g6 are good for white. After 6...Bg4 7.cxd5 Nxd5 8.Qb3 Bxf3 9.gxf3 black has the choice between 9...Nb6 with a slight push for white and 9...e6 which is the theoretical recommendation and should lead to a drawish endgame.

Please excuse me for repeating these wellknown facts (anyone disagrees?!), now to the point.
As I don´t like the mentioned endgame as black I´ve started to play some very rare sidelines. These are 
(a) 6.Nf3 Be6 7.c5 g6
(b) 6.Nf3 Be6 7.c5 Bg4
(c) 6.Nf3 a6 
There are not much games and nothing like theory. Some tryouts from Miles and Speelman suggest that white may gain a slide advantage though the black position should be perfectly playable. Much more it´s very easy for white to get an inferior position by playing natural moves. The black side is much easier to play but somehow white might be better if he knows how to handle the position!? In internet blitz games I have a very good score by varying these variations. And computers are soon out of book!
7.c5 seems to be white´s best choice against 6...Be6 while it seems to be a mistake against 6...a6 regarding 7...Bg4 . Then 6...a6 7.Bg5 might be a better try as 7.cxd5 Nxd5 8.Bc4 e6 is likely to transpose to a QGA with no problems for black. 

Does anyone else have experience with these variations or knows theory about it ?
Is there already a refutation ?? Of course I would not like to give up my pet lines but better now than after the next tournament game.

tracke