My favourite opening, as many of you will know  .

In this game, I had just started experimenting with the Stonewall Attack, or System, which tends to block up the centre but aims at a powerful attack on the kingside. This is supposed to be "a good weapon against players ECF graded 80-130". Sadly I don't know what the equivalent is in GK ratings! It is worth knowing that, at this stage, the Stonewall can easily transpose into a Colle System: 1.d4 2.e3 3.Bd3 4.Nf3 instead of 1.d4 2.Nf3 3.e3 4.Bd3, but the latter is more flexible, as it develops the knight earlier.

Typically Nbd2 is played before f4, to stop the black knight from posting at e4. Currently that square is threatened by the bishop, but black would love to trade his knight for the bishop, as this would effectively stop White's plan, which often including a bishop sacrifice on h7 or g6. This is another reason the stonewall is rarely seen in top-tier chess; with the knight at d2 development of the dsb is hard to come by.

6. c3 is a bit of a wasted move, especially with black's knight at c6. c3 is often played when black pushed his pawn to c5, to provide a square the bishop can retreat to and stay on the b1-h7 diagonal.

I often play c3 to prevent Black's queenside knight from annoying me and disrupting my formation with ...Nb4 on the queenside.

On the whole, I'm with fatcat2 on that decision. White would probably like to keep the bishop on the b1-h7 diagonal if he can.

Yes...and then the queenside knight comes active via b1-d2-f3.

It's worth knowing that, often, White plays 9.Qe1 before the attack starts.

In some lines, the DSB gets redeployed to e1 thence to h4 - rather long-winded, I agree. But there is also a potential , if circumstances permit, to bring the DSB out to a3, striking towards f8. As for White's Q, often it gets played to e1 and even h4, where it stands perfectly safe as the attack builds.

Of course, 11.h3 is not possible owing to 11...Nxe3.

Rather a strange move at first sight, but Black wants to keep his bishop active, rather than being boxed in after 11...e6. All the same, the move seems dubious, at it makes communication between Black's flanks very problematic... A possible continuation is 12.h3 Nh6 13.Ng5 Bf4 14.e4, freeing up White's Q-side for attack.

Better was the unlikely-looking 12...f6, taking up the challenge in the centre. Now with 13.Ng5 White could have got a strong attack: 12...c6? 13.Ng5! ... The threat is 14.Nxe6, then to pick up the loose g4-knight. If 13...Bc8 14.e6! Nh6 (14...Nf6 15.e5) 15.exf7+ Nxf7 16.Nxf7 Rxf7 17.Rxf7 Kxf7 18.Qf3+ Kg8... Now White has the backward e-pawn to work on: 19.Bg5 Bf6 20.Qf4 Bxg5 21.Qxg5 ... This situation that arises might not look like much, material being equal and the game much simplified, but White has in my view several advantages: the weakness of Black's e-pawn, the more compact pawn formation overall, a stronger K-side presence, and the more active pieces.

13.exd5 is OK, but in my view Ng5 was slightly stronger.

I wanted to put the queen behind the bishop into a powerful attack line.

Broadly speaking, fatcat2's idea is thematic and well-motivated. But maybe more accurate was a bolder execution of this plan, with the view to freeing up the a1-rook (bearing in mind it will still be somewhat inhibited in its movements by the pressure being applied to a2 by the d5-bishop). For instance: 15.Bf4 with the idea 16.Qd2, and, when conditions are right, maybe to bring the a1-rook to the centre or K-side.

I can't subscribe to the view that the capture at f6 is mandatory for White, playable though it is. I rather think I would play 16.Qe2, hoping for Black to exchange on e5, which would laeve Black with a porous K-side and a backward, isolated e-pawn. If 16.Qe2 f5, then 17.b3 would free the a1-rook for duty in the centre. One possibilty will be to play the rook to c1, with the c-pawn advance in mind. Finally, if 16.Qe2 Nf7 17.Bf4 consciously overprotects the strongpoint e5, cramping Black's position and making it difficult for Black to free up his game.

I agree with the annotator's view, here. After 17...Nf7, Black would have been on level terms with White.

True enough. Perhaps less obvious, but by no means hard to find, is the move 18.Qc2! to exploit the suddenly weakened light squares in Black's game. The h7-pawn would have been unsalvageable (18...f5? 19.Bxg5). Although I consider 18.Qc2 to be the better move, 18.h4!? is a very good "second best"!

Black was probably ill-advised to go in for this exchange, depriving him as it does of his most actively posted piece. The rook recapture was far better than the pawn recapture (gxf3?) which would have blocked in the f1-rook.

He had to keep the piece. If the opponent is attacking, it is likely he has the advantages in space and time (which is certainly true in this game here). Then the defender can not afford to cede an advantage in material as well. Black had to play 20...Nf7 and try to hang on. White's advantage would not have been so huge as to offer no hope of Black's holding out. But once the piece goes, Black's chances of saving the game become vanishingly small.