yet the GK analysis dubs this a mistake... interesting

Not so interesting. It is considered inferior to the main lines (otherwise it would be a main line itself) but the stats favour White because it's one of those oddball moves that's most likely to be played by players like Freddy who know what they're doing and want to spice things up a little.

Qc2 is not good because Black has not committed to ...d5. Black can just play ...d6 and play for an Old Indian Defense in which the White Queen is not doing much of anything on c2.

Freddy is not exactly fond of playing main lines  

Esta jugada Qc2, la considero muy debil, ya que la teoria sugiere que despues de mover peones del centro, hay que mover los caballos, aqui el ngro podria hacer fianchetto al lado del rey y con Bf5 poder atacar la dama de Qc2, inclusive podia estudiarse profundamente esta variante de Qc2 en la apertura inglesa.

I noticed that GK's analysis recommends, instead of the kingside fianchetto, 4... Bd6. Proof that the engine doesn't really understand openings once it's out of book, or is there something to it?

I'm afraid Freddy is playing somewhat passively... no h4, no b4... castling on the same wing than you...

With the K's on the same side, a Pawn-storming attack would now open up White's OWN king to a counterstrike, thus rendering it useless as a tactic.

White's knight has a nice place in d5, yes... and probably, the white's attack would go through f4 somewhere in the future. Doubled pawns in the center are strong, but also block white's freedom, leaving the queen on the queenside. And, yes, e4 controls f5 - a sound move in this kind of position. In my opinion, white didn't get much from the opening.

hm... I'm not deep enough to say, but I see as interesting to recapture with the rook, thinking in fxe5 and open the game a little (before black defends d6 in order to kick it out with c6). Or maybe, blocking the center this way, white is thinking in f5. It is hard to see how would the white's dsb contribute to attack, now.

Hitting the semi-backward pawn at c7. It also discourages that pawn's attempt to strike at White's centre (20...c6? 21.dxc6ch). I believe these are the considerations that decided White's capture at move 19.

One thing those doubled pawns will do is to hamper communication between White's wings. It will be near impossible for the bishop at b2 to participate in any K-side action for a long time to come, by the look of things.

It doesn't seem that Qc4 and f5 coordinate each other. With the white queen in e2, now black would have to face Rh3 and Qh5 as a possible attack way, but now black has tempo to defend. And the attack on the c column just lacks the "unbalanced element" that makes an attack successful. Ok, I am writing as if I were an expert, and I am not, I'm aware  ))

Well considered remarks for all that, and if it turns out you are wrong, we all learn something. It looks as though Rh3 and Qh5 might yet be feasible, but Black can contain that attack easily. White might have to try to explode the Joblock by h2-h4, but it leaves whatever attack might come of it very localised. Can White switch pressure from one wing to another (that c7-square is still under pressure)? Even with less space than White, Black seems to have the better lateral communications.

Joanne is using a defensive version of the famous Blockade strategy popularized by Aron Niemzowitsch. With White's doubled pawns, it's particularly effective.

As White here, I'd think about Qe2-h5... I'm not sure if there is an attack there because of Black's rook and queen put the h-file pressure has to account for something! [hmm, reading on I realise that Qe8 was a possibility for Black...]

But why "head" for e8 when you can play there in one move?

Overfinessing, perhaps? Possibly Black wanted to retain touch with the c7-square until the WW of the W (WQ) began its journey to the K-side. No: I think Joanne was being far sighted here. The h-file is too narrow a front upon which to advance: White really needs the g-file as well. The Queen on c8 goes to entrenching the pawn at g5 very strongly against a possible attack via h4. In fact that g5 bastion has become immensely strong. I believe it was the First Lord of the Admiralty who, when asked about the possibility of an invasion of England by Napoleon Bonaparte's vast armies: 'I do not say they cannot come. I say only that they cannot come by sea.'

Contrast those bishops! The Black bishop is the baddest of bad bishops: all but one pawn on its dark squares; practically no mobility worth a hoot; it is little more than an overlarge pawn. The model of virtue on the White side has plenty of space to move in, but dare not go near the spearpoints projecting from Black's compact position. 'Bad' bishops can often make very good defensive pieces.

Wthout the help of the h6 rook, white's attack doesn't reach its full power... with a white rook on c1, it would have been another story. In my opinion, Ba5 should have been prepared with - say - Qc2 or something.

Maybe Freddy was already muttering: 'will no one rid me of that turbulent pest on e3?!'

This game already looks like a draw at this point - how is either side expected to break through here?

Black is not expecting to break through. For her a draw is as good as a win in a game like this. This is one advantage a lower rated player has, by the way: to use the possibility of a draw in order to induce the stronger player to over-reach and compromise his game.

Agree with Ion. Against higher-ranked players I'm always content with a draw and try to blast the material off the board and force the position to either simplify to nonexistence or lock up so there are no attacking lines. Then if the higher ranked player gets frustrated and sacrifices to open the board I have a material advantage and swapping down accentuates that. Joanne in this game hasn't swapped down but she has effectively locked up the board.