k The Milner-Barry Gambit is one of the sharpest white weapons against the French Defense.
15
k Over the years there have been found many suitable ways for black to face the Gambit.
15
k One of these ways is shown in LectureBot's 'Refuting the Milner-Barry Gambit in the French Defense', written by Kabal.
15
k I recommend to anyone who plays the french, viewing that lecture. You will take by surprise many Milner-Barry fans with that secret weapon.
15
k Let's revisit the proposed variation:
8
e2e4
e7e6
d2d4
d7d5
e4e5
k The Advance Variation of the French.
8
c7c5
c2c3
b8c6
g1f3
d8b6
k All the fight is about the d4 square.
8
f1d3
k This move introduces the Milner-Barry Gambit.
8
k d3 is the Bishop's dream-square, but the drawback is that d4 is now insufficiently protected, so white will have to sacrifice his d4 pawn.
15
c5d4
c3d4
k Black must now avoid the well known trap: Nxd4 Nxd4 Qxd4 Bb5+, winning the Queen.
12
c8d7
b1c3
c6d4
f3d4
b6d4
e1g1
k This is the crucial position of the Milner-Barry.
8
k Black's most secure move is now a6, preventing ideas of Nc3-b5.
8
k However, accepting the second pawn sacrifice has become increasingly popular.
10
d4e5
f1e1
e5b8
c3d5
f8d6
d1g4
e8f8
c1d2
h7h5
k This is black's new idea, which is explained in detail in LectureBot's 'Refuting the Milner-Barry Gambit in the French Defense' by Kabal. Again, I recommend that lecture to anyone who plays the French as black.
10
k White's Queen must now go to the ugly square h3, as otherwise he loses his Nd5. Black will develop his knight with Nh6 or Nf6, with the idea of going to g4. Practice proved that black has the advantage.
15
k Having played the Milner-Barry for many years, the author (knackie) made up his mind how white could react against this new threat to his favourite weapon.
15
back 8
k The result of this search can be seen right at the start of the variation:
12
f1e1
e5b8
k And now, instead of playing the usual Nxd5, I propose:
d1f3
k White intends to take on d5 with his Queen. He also threatens to play Bf4 in some variations.
8
k This might seem illogical at first sight; white needs two moves to take on d5 instead of one.
10
k The main variation shown before proved, however, that the position of white's knight at d5 is very loose. White will have to lose some tempi for saving that knight anyway.
15
k Let's now look at some concrete analysis:
10
k What could be black's reply to Qf3?
10
k One idea would be defending the pawn at d5. This can be achieved with Nf6 or Bc6.
12
k Let's look at Nf6 first:
8
g8f6
k Bg5 is not good here: 
Bg5 
Be7 
Bxf6 
Bxf6 
k White takes on d5 but doesn't achieve anything.
18
back 4
c3d5
f6d5
f3d5
k Compared to the theory line, white doesn't need to worry about his loose Nd5, meaning the white Queen can move freely around.
15
f8d6
d5g5
k Black's best here is probably Kf8, similar to the theory line, but the exchange of knights certainly favours white.
15
k If black plays g7-g6 here, he will lose inevitably. e.g:
g7g6 
Qf6 
Rf8 
Bh6 
Rg8 
Bb5
k With the threat of Rxe6.
20
back 6
k Let's now see how play can proceed if black castles:
10
e8g8
g5h4
f7f5
k Black has little choice if he wants to avoid mate.
8
d3c4
k This strong diagonal along with the d-file will prove more than enough compensation for the pawn.
10
g8h8
k White cannot take on e6: Bxe6?? Re8!
10
c1g5
e6e5
a1d1
b8c7
e1e3
k Intending Rh3.
8
f5f4
c4d3
k Black is in trouble.
10
d7f5
d3f5
f8f5
e3h3
g7g6
h3d3
k Black will be overwhelmed on the dark squares.
10
back 24
1
k Let's now look what happens if black protects d5 by means of Bc6:
10
d7c6
c1f4
f8d6
k It is difficult to say which is white's best response here, but the following is very interesting:
12
f3g3
k Strange move, isn't it?
8
d6f4
g3g7
k The position is rather unclear, but white should certainly be no worse.
15
k Let's now look what else black can try after Qf3.
8
back 6
k Instead of protecting d5, black could proceed with Bd6 as in the main variation.
15
f8d6
f3d5
k It seems black can now safely play Nf6 followed by 0-0, but...
10
g8f6
d5g5
k This attacks g7, so black has little choice.
8
e8g8
g5h4
k This position happened many times in knackie's games (blitz, rapid, and even tournament games).
10
k The result? Black NEVER survived.
8
k White's idea is very simple: Bc1-g5 followed by BxNf6 and Qxh7 mate.
15
k While white's plan is easy to foresee, black has no easy way of preventing it.
15
k I recommend you try this position against your computer.
10
k Let's look at two sample variations: h7-h6 and Be5 (played by Fritz)
10
h7h6
k Black tries to prevent Bg5.
8
k The sacrifice Bxh6 now only leads to a roughly equal position:
8
c1h6
d6h2
k Everything else loses.
8
g1h1
g7h6
h4h6
b8f4
k After this strong defense, black gets an equal game.
8
h6h2
f4h2
h1h2
k The game is equal.
10
back 9
k Instead of Bxh6 white should play:
8
e1e3
k Threatening Rh3 followed by Bxh6. A game against Fritz5 continued:
8
f8d8
e3h3
e6e5
h3g3
k Black is now helpless against Bxh6 or Qxh6.
8
k So h7-h6 was obviously bad.
6
back 6
k Instead of h6 Fritz5 usually tries:
8
d6e5
k A very clever move; after Bg5 h6! white would be forced into the equal variation Bxh6 Bxh2! we've already seen.
15
k Again, white should prefer:
8
e1e3
k White simply threatens Rh3 followed by Bxh7+, while h7-h6 would again be met by Rh3.
15
g7g6
k Analysis showed this is probably the best reply.
10
d3e2
k A very strange move; white wants to be able to take on h5 (after Nh5 or h7-h5) in some variations.
12
d7c6
e3h3
h7h5
c1g5
k Let's just follow a sample analysis game against Fritz5.
10
f6d5
c3e4
k After Bxb2, white answers Bxh5!
10
k After f7-f5 white would answer Nc5! Bxb2 Bxh5! You can check these variations with your computer.
15
k The game now goes crazy. Fritz played:
15
f7f6
e2h5
f6g5
e4g5
d5f4
h5g6
f4h3
h4h3
e5h2
g1h1
g8g7
h3h7
g7f6
h7h6
f6e7
h6g7
e7d6
a1d1
c6d5
g5e4
d6c6
d1c1
k and black was mated on the Queenside.
12
back 38
k Further analysis is of course needed, but if you play the Milner-Barry gambit and you find the 'Refutation of the Milner-Barry' hard to meet, Qf3 should provide a serious alternative with the surprise factor being on your side.
20
k If you want extensive analysis in Chessbase or PGN format, just send a mail (or a message) to knackie. I'll send you the analysis by mail-attachment.
20
k That's all folks! The author now sits and waits for the refutation of the refutation of the refutation of the Milner Barry ... :-)