4.a4.
Yes, you are right. I spot this plan at once when I replayed the game at home. I don't know why I didn't play that way... It would have been an easy advantage.
Maybe because before my eyes was floating a Schmid Benoni position from a nice win by Dreev from 2006
(
http://www.chessgames.com/perl/chessgame?gid=1410164 )
that I looked through just hours before my game. (1.d4 Nf6 2.Nf3 c5 3.d5 b5 4.Bg5 Qb6 5.Nc3, when after a ..b5b4 the horse goes to a4 attacking the queen and then comes b2b3, Na4b2-c4, according to the included analyzis.)
As you pointed out, in my game it was an even better position for White than after 2..Nf6, so there was an opportunity for an even more straightforward initiative.
6.h3.
I like space advantage and restrain the opponent's pieces. Sometimes it hits back, when after a while, in very tense positions (a lot of pieces still on the board) I would just have to make the decisive move to finish off the cramped opponent but I can't calculate precisely.
With the inclusion of b5 and a6, I didn't see where the c8B could be of much use to Black, who was already a little cramped, in the long run if he didn't try exchanging it for my knight. I felt the bishop is useful attacking on the f1-a6 diagonal, while the f3N could head for c4 later. (or something like that, as in the more normal Schmid Benoni positions, when there is no early b7-b5 as happened in the game)
I don't know if this logic is more appropriate than what you gave. But this is what was going on my mind.
9.Re1 with Bf1.
I have the book Benoni Systems by Raetsky & Chetverik (2005). I had played through a few games in which the rook went to e1 and then the bishop to f1, that's why I played that way. I felt I had every time in the world. Black's formation was comfortable to develop against.
13.Nxe5.
Rybka3 also likes (gives) 13.Nxe5 dxe5 14.a4 for White (as happened in the game) and, after 14..b4, evaluates it as "+= (0.40) Depth:18".
After 13.Nd2 (which it evaluates as being only equal), during the game I had the feeling that it would be not so advantageous to play for f2-f4. I didn't know exactly why. I had the feeling that the plan with my h3-h2B-e1R vs. his piece formation didn't promise much for me. His bishop also could have gone to h6 anytime, blocking f4, for example. And I liked the idea of shutting down the a1-h8 diagonal immediately: another e-pawn, which can be a target, a bishop that is blocked by that pawn, and I can take an f4N anytime. (I had just spent long minutes calculating what could have happened after 12..b4 13.Na4 Qa5 14.c3.) While with a closed a1-h8 diagonal I can play for a2-a4 and no dangerous b5-b4 can attack my knight and b2P simultaneuosly.
I also have the feeling that it would have been hard to win if I had had to win. I was in a favorable situation--I think he would only have accepted a draw if he had foreseen some forced line leading to his defeat... (As happened against another 2300+, months earlier. He was "-+ (5.something)", that is, already being destroyed, when he offered a draw just before a forced tactical shot.)
I wasn't playing for a win, I was just trying to kill off the opponent's initiative before it started and place my pieces on better places with every move, and "let's see what he has". That kind of play is what I love the best and this time the opponent was a good partner in practising it.
Interesingly enough, I have better results against 2200-2399 opponents than against 2000-2199 opponents. The latter often get in a number of suspicious moves, and play even more inferior positions, when I simply can't choose from the promising plans. I lose quite a few "+/-" positions that way.
I am looking forward to your return and comments, which are inspiring.