French Advance: Milner-Barry Gambit

I have problems to get compensation for the pawn in this line. I usually go for this:

I am already two pawns down in the final position and I feel under pressure, because I have to get some kind of initiative going. Technically I am up 3-4 tempi since I have already castled, activated my rook, have one more minor developed than black and it's White to move.

If you don’t want to be pawns down just play 6.a3 and if you want a one pawn gambit try 9.Nbd2 and look at the games of FM Melekhina

I just have problems with the line and want some suggestions.

And FizzyBand gave the suggestion.  6.a3!

The Milner-Barry Gambit is dubious at best!  Also, even stronger than 11...Qb8 is not taking on e5.  Black should play 10...a6 with a significant advantage!  Of course, he needs to know the theory, but if he does, White is playing for a draw at best!

Give up on that garbage line, and start playing 6.a3 against 5...Qb6 and 6.Be2 against 5...Bd7.  If you have any questions about those lines, I'll be glad to help you out.  If you are going to insist on playing a highly dubious gambit, can't help ya!  All advice I have in the MBG is for Black!

Nobody's afraid of 9. Nxd4. 9. Nbd2 however provides some compensation

I didn't mean to criticize anything, I was just pointing out what I was asking for. I am also not strictly insisting on playing the Milner-Barry Gambit, any suggestions are welcome.

And it may not be the answer you want to hear, but it's the truth, and I refuse to sugar coat it as the French Advance is basically my bread and butter from both sides.  I know the French Advance like the back of my hand, and the truth is, there is nothing good for White in the Milner-Barry.  White used to get good results in the 80s and 90s with it, but it's been figured out, and the ball has been in White's court for the last 15+ years, and that ball is made of lead - it ain't getting hit back to Black's court.  The Milner-Barry, simply put, is no good these days!

Whats the purpose of 6.a3 against 5...Qb6? It's also played against 5...Bd7 as I can see.

What becomes apparent is that the only square blacks DSB has is e7.

6. a3 is a good bit more accurate vs 5...Qb6 than against 5...Bd7. 5...Bd7 6. a3 f6 and black's fine, actually probably doing pretty well since a3 isn't as useful a plan if they haven't committed to Qb6

I still didn't get the purpose of 6.a3 from that though. To take away a square from the DSB? Is it about that?

White intends b4. After which black will have to move the c-pawn one way or another (typically capturing on d4), releasing tension. Furthermore it gains white a lot of space, in some lines allows Bb2 to defend d4, etc

Grandmasters still play it and they do fine ~ 9. Nbd2 is more in fashion right now though.

I recently had an IM play 9.Nbd2 against me. Seems that you are right.

It all comes down to d4.

Against 5...Bd7, which does nothing to attack d4, it gives White the time he needs to castle and consolidate.  After 6.Be2! Ne7 7.O-O cxd4 (7...Nf5 8.dxc5! with advantage) 8.cxd4 Nf5 and it's too late for Black.  9.Nc3! Qb6 10.Na4! with advantage.

However, after 5...Qb6, Black is a move faster and now 6.Be2 is a problem for White.  After 6...cxd4 7.cxd4 Nh6 (or 7...Nge7) 8.O-O Nf5 and White has no good way to guard d4.

Therefore, White instead tries 8.Nc3, but after 8...Nf5 9.Na4 Qa5+! (This check is critical and not available when ...Bd7 by Blakc and O-O by White are thrown in) 10.Bd2 Bb4 11..Bc3 b5! 12.a3 Bxc3 13.Nxc3 b4 14.axb4 Qxb4 with a slight advantage for Black.  His pieces are better than White's.

Therefore, since 6.Be2 is too slow if Black doesn't give him the time, White should expand on the Queenside instead.  After 5...Qb6 6.a3, 6...a5 is dubious and White should take a Milner-Barry Gambit approach since ...a6 is no longer possible and b5 is weak.  6...Bd7 is also useless here.  Black should either play 6...c4, stopping the queenside expansion and try to take advantage of the weak b3-square, or else play the more active 6...Nh6 (I have played both these moves as Black) 7.b4 cxd4 8.cxd4 Nf5 9.Bb2 (stronger than 9.Be3, which isn't refuted, but opens up more issues for White, and he also has to watch out for blunders in this line - after 9...Bd7, 10.Nc3?? is a blunder, for instance.  Now 10...Nxe3 11.fxe3 Nxb4!! 12.axb4 Bxb4 -+) and now Black has a critical decision to make.  9...Bd7 or 9...Be7.

The point is that if it is White's move, he can now develop the Bishop actively on d3 because Black can't take 3 times on d4 due to a fatal discovery at the end, losing his Queen.

The move 9...Bd7 stops this, and so White must play 10.g4! or else Black gets the advantage with a subsequent 10...h5.  The Knight is driven away from d4 via 11...Nh6 or 11...Nfe7.

The move 9...Be7 does nothing to stop 10.Bd3, and is therefore what White should play here.  If 10.g4?! here, Black can play 10...Nh4 to trade a set of pieces, a move that is not possible after 9...Bd7 as h4 is not guarded.

Personally, I play 9...Bd7 as Black, but there is nothing wrong with 9...Be7.

I think in this line black is better to return the queen to b8. It's actually not recommended to take the e5 pawn.

Nb5 is a very central idea to this gambit. Now here black is almost forced to play Bxb5, and white should have a good game:

But again, MatthewFreitag, 10...a6! is a problem.  So the argument of 11...Qb8 vs 11...Qd6 is irrelevant.  If someone forced me to play 10...Qxe5, then yes, I'd resort to 11...Qb8, but a move sooner, 10...a6 is so much stronger, and is the only line I'd ever play any more as Black.

Back about 15 years ago (before the huge outbreak of ideas for White), I used to play 10...Qxe5 and 11...Qb8, giving the d-pawn back, but new ideas for White came in that line, I seem to recall around the 17-20 range somewhere - would have to go back and check as it's been 13 or so years and I would never enter it from either side since then.  White I'd play a different 6th move (what that move is depends on Black's 5th), and as Black, 10...a6 rather than 10...Qxe5.

True, a6 almost seems to refute the gambit.

I would play 11.Qf3, where after 11....Qxe5 you can at least get some decent counterplay with 12.Bf4.

Frankly speaking, this gambit is just bad for white.

All black has to do is play 10. ... a6, prevent white's counterplay (which isn't hard to do) and then white has to fight to even stay alive. If you want to play the Advanced variation (which I don't recommend; the variations with 3. N.c3 are the real test), at least play with B.e2 to have a decent game.