There doesn't seem to be anything against 5...bxc6, doubling towards the centre. This would at least get some mileage out of the early Q move. On general grounds this capture is to be preferred, but in the normal exchange lines this would leave the e-pawn hanging. Here the e-pawn has protection. But it seems clear from Black's play he wanted to avois doubling his pawns!

Inferior? I doubt it. I like this move as although black has to move his queen again back to d8 black gets the two bishops - a long term advantage where as white's lead in development will disipate with correct play.

It was either this move or ...Qd6.

1...d6 is much the better move. Up until here, Black still has a reasonable game, but 7...a6 doesn't even have the prophylactic function of preventing the c6-N being pinned by a B (even then it would be preferable to let the pin happen before playing ...a6, which would throw upon White the need to take a decision).

why withdraw the knight? 11.Be3 If Bxd5, it opens the center for white (with the king still in the center) and makes it a well developed game for white.

And a move that would not be available if Black had played ...dxc6 earlier, let us not forget!

leaving the knight in the center and Be3 would have prevented this

Right now, Black's game looks pretty OK. Rather than looking to a pawn break by ...f6, Black might consider something like ...Ne7, ...c6, ...Rac8 with a view to activating his Q-side majority. White's pawn break at f5 is far in the future, and can be stopped altogether by a timely ...g6 and ...Kg7.

Instead of Re1 I prefer Nd2 f4 Nf3 giving the pawn super protection with f5! to follow but there is nothing wrong with the text move... Just preference.

far1ey's move also sets in train the pawn break at f5, after due preparation, of course. At that, 17.Nh2 might be preferable, so as not to interdict the Queen's coverage of d4, and hence allowing Nc2-e3. The plan would run along the lins of Nh2, f4, Ne3, f5. According to this plan, the R is much better placed at f1.

dxe5 is actually best I think. After Rf5?! f4 the pawn is secure and white can play the excelent Nd4! Bringing the knight to an excellent post.

I agree with far1ey that 21.dxe5 is quite playable, but I'd probably play 22.Nd4, and then follow with 23.f4 (a tempo play). However, Black's Q-side pawns would suddenly have become very mobile and dangerous. On balance, then, I prefer the move actually played, which gives White a lot of pressure down the K-file.

I like 21.Rxe5 because it makes d4 a lot easier to hold. It also gives White the E-file.

There's a lot to be said as well for 22.Qe2 or 22.Re1 or even 22.Nb4. If 22.Nb4 (threatening 23.Rxc7 and 23.Nxd5) 22...c6 23.Ne3 etc.

Possibly too defensively minded. Why not 23.f4, threatening f5?

A good appreciation of this endgame position. If Black captures on e5, Whit's pawn structure is consolidated after dxe5, and he has a passed pawn into the bargain. In Queen and Pawn endings, possession of a passed pawn can often be decisive.

Black's exchange on e5, however understandable, was probably ill-advised. I like 29...Qd6 here, pinning the R against the K (I did think Kg1-h1 was better than Kg1-h2 earlier), and threatening to win a pawn with exchanges. If 30.f4, then ...Rd8 threatens the d-pawn (31.d5 c6; or 31.Re4 Qd5 or ...c5). 30.Qe2 leads to the same result. However, even if White does lose his isolated d-pawn with this line, he would have had by no means a lost game.