The comments I have added throughout are meant simply as a supplement to the annotations. The notes are sufficient to tell the story, but the comments are added for a bit more detail if you want it.

However, this move is a good lead in to the Colle System (planning quick development by Nf3; Bd3; c3; Nd2; 0-0; Re1; and then break by e4) or the Stonewall Attack (planning f4 and c3, developing a Knight at f3 and Bishop at d3). Both are good lines for developing players, setting them up for the middle game.

4.f4 - the Stonewall - is still available.

4.f4 is the incongruously named Stonewall Attack. The beauty of the system is that the strategy is straightforward and clear: attack on the K-side. The downside is the opening's strategic limitations. However, the Stonewall can also played as a Defence (with the black pieces), and as such has been seen in World Championship matches.

Black could have played a Stonewall Defence, here, with 4...f5! That said, this is a very good square for the Bishop.

White looks to be playing a Colle System. The problem of the c1-Bishop is deferred whilst the rest of White's game is developed apace (Bd3; 0-0; Re1 and then, depending on circumstances, play, or prepare, the e4 break.

Not if white is planning Qh5+, and if g6, Nxg6.

5...f6 falls between two stools it seems to me. Black could still have played 5...f5 (Stonewall defence), following up immediately with 6.Nf6 and 7.0-0, to prevent white's exploiting the weakened light squares. What damalfi has pointed out is what weaken colour complexes can mean - especially the h5-e8 diagonal (remember Fool's Mate?). The pawn standing at f6 makes the defence of this diagonal a bit more problematic. ...

Again, damalfi is on the money. Note that 7.Qh5ch Ke7?? 8.Qe7 would have been checkmate. Black's best defence would have been 7...Kd7 when White can choose between damalfi's line, or something like this: 8.Ng6!? Qe8 (pinning the Knight) 9.Ne5ch (it was only a 'half-pin') ...Bxe5 10.Qxe8ch Kxe8 11.dxe5 Nd7 12.f4.

? This makes for the previous move to be a mistake. Qh5+. If black goes Ke7, Ng6+ (edit: Qf7#, thank you Ion!). If Kd7, Qf7+, Qe7, white exchanges queens and retreats the knight - but black has lost the castling rights. Blackhas more space, though, but it is not a definite edge. (edit: in Ion's line, after 9. Ne5+, fxN seems to stop white.)

This knight is costing White a lot of time! Black should now tie together his K-side pawns by 8...f5 (or 8.h5, but I prefer the former), protecting the attacked g-pawn.

Qxg4, and if e5, Qh5+...

From a strategic point of view, this pawn move is fine, but winning the pawn by 9.Qxg4 was even better.

I must admit I didn't see the g4 pawn could be captured

I rather like what White has played here (leaving aside the much better Qxg4, of course). In fact it would have been even better played a move ago, instead of b3. The problem of the c1-Bishop will be solved once White plays e3-e4, which this move prepares.

This way white keeps the possibility to play e4 at any moment, and besides, presses on the f column (f6).

A very interesting decision.

0-0? Black castles directly into white's attack, with his pawns advanced...

White's best here, I think, was 12.e4, and if 12...dxe4 13.Nxe4. White would then have been well placed to attack on the k-side. A good alternative would have been 12.Bd3, but I prefer the pawn break at once.

Drives the King back to the g-file. Black ought to have tried ...Kh8, but probably felt that would have been too far away from his defenders. But now White has been given an extra move with which to build a telling attack.

Fine move, and to the point.

White can't win the fight for f5, even with this pawn. But if black goes f5, now white can open another column against the black king. Then, Ke2 in order to bring in the other rook asap. Any other plan is very slow, giving black tempoes to reorganize and defend.

This would have been a good response to 14...f5, as well. But also well worth considering was 14.e4, which opens up the diagonals for both bishops. Note that Black's Q-side remains 'in barracks' and will take a while to mobilize.

? White has just not enough firepower to do this.

15...b5 was of no real help to Black's K-side defence, nor by way of counter-attack. Instead, 15...Nd7 was worth a look, possibly with a view of shoring up the K-side defences (...Nf8), or perhaps contemplating a counter-action in the centre by ...c5 or ...e5. ############### This bishop check is a disastrous mistake, for it loses more than Bishop for pawn (-2). This will put White in -6 country. Instead a patient build-up was called for: 16.e4, say, or maybe 16.g5 f5 (to keep the g-file closed) 17.g6 Nxg6 18.Nxg6 hxg6 19.Rg1! pinning the g-pawn, and winning it and the g-file as a whole. A finesse like that is way better than taking the pawn at once by 19.Qxg6ch Rg7 20.Qh5, when Black owns the g-file.

The queen had no other squares to go to.

The Queen had no choice, being trapped with all its retreats cut off. So White sells it for what he can get for it. Such a pity, having a fine game in prospect.

?? White lacks of patience. (edit: of course, the WQ was doomed. I overlooked it).

True is that black's king is wide open and far form his pieces. But believe me, a good defender would have survived the storm.