I would go for 12...d4, locking the central pawns and blocking White's fianchettoed bishop.

This is fine for Black. Nor did White have any choice but to exchange, at that.

Good move. (I think you mean, a pawn chain from g7 to d4.

Bear in mind, though, the the chain is a means to an end, and in any case is merely incidental here. For one thing, opening up the board by 14...dxe4 was not that bad an option, whilst maintaining the tension by leaving the pawns where they are was also not adverse to Black's interests. I like 14...Qd7, here, with the threat 15...Bg4, trapping the WQ! Now, if White were slipshod in averting the threat, this could happen: 15.Qf3? Bg4 16.Qe3?? d4! After 14...Qd2 15.Qd1 (say) 15...f5 Black would still have a tremendous attack. White piece minus would be too crippling.

The closed pawn centre has no effect on an exactly equal exchange: however, you are correct, exchanging equal pieces when you are a piece down is generally a bad idea.

The closed pawn centre does have an effect: White is giving up a good bishop for a bad one. Clearly this move is motivated by the fact that the bishop had no mobility on b2, but perhaps it could have been redeployed to d2 via c1.

I like the pawn capture. That added pawn presence does Black's centre no harm whatever!

Well...there is the problem if the d5-pawn being very backward and weak, although if Black maintains its defence it is more an advantage than a disadvantage.

The d5 pawn supports a pawn push to e4. Had Black taken with the bishop, the pawn remaining on c6 would be rather more backward and weak. (Not that such considerations really matter when you're a piece up.)

This time I would have preferred the bishop recapture. Why? Take a squizz at those lovely, pristine, light squares about White's King! The bishop would be very hard to shift from that central posting, and then Black could set about attacking that weakened K-side.

19.Rab1 is clearly much better: the text leaves the hapless a1-rook completely trapped and useless.

If White is about to push the b-pawn, the a1 rook will not be useless for long! Had this rook remained on e1, it would have found itself biting on the granite of Black's pawn chain. Fatcat, I suggest you read up on mysterious rook moves.

I think this one is a bit too mysterious... With 19...f5, Black could begin a decisive attack. His next move in my view wrongly allows White's knight to enter the game at b5.

I don't think such was White's intention. Anyway, 19... a5 made White's plan of activating the rooks a lot harder.

White's knight manouevre hasn't achieved much. Better, after 20...Qf7 (cunning lure, that) would have been 21.b4 axb4 Rxb4, and White might really be able to begin active operations on this flank.

! smashes White's pawn-structure apart, and takes advantage of the immobile rook on a1.

Now was the time for some "Hope Chess". There's no way in the long run of saving those rearward pawns, so why not try a trick or two? Here goes: 23.Re1!? Now, there is an immediate threat, here: 24.Nxd4, forking Q and B. If Black plays 23...Qxb3 then 24.Nxd4 Qb6 25.Nxe6 Qxe6 26.f4 - and Black is faced with another fork threat! Or 23...Qxd3 22.Rad1 Qxb3 23.Nxd4 exd4!? 24.Rxe6 and Black can't take on d1 owing to 24...Qxd1?? 25.Qxd1. I agree: this doesn't actually change anything very much, but it does put the frighteners on, and offers Black a few fun ways of going wrong.

Argues a commendable flexibility of mind. Just here, though, the simple 24.Qxf5 seems much preferable. White would even be suggesting he might go after the back d-pawn by 25.Qe6+; 26.Qxd5.

And trapping the a-rook again.

The a-rook is now needed to protect a4, though admittedly, even after 25. Rab1 Black should prefer to take the d-pawn.

The title of the annotation did warn us...

No... I didn't know that. I just have to see how that feat was accomplished!   If Black were to blunder a whole Queen for nothing hereabouts, he ought still to have good chances...

After 27.Rb7? Be4 ought to have ended White's bid for counterplay (28.Qd7 Bxb7).

Or deflection/distraction. But same principle involved.

The side ahead on material should welcome exchanges. 30... Rb2 is strong.