I have recently begun the habit of annotating my own games, as opposed to pressing the 'Computer Analysis' button. To practice, I have tried to deeply annotate a game below. The variations have not been checked with an engine, so I'm sure there are plenty of mistakes. If you find one, please let me know.

A couple of questions.

Why is 9. f5 given an exclam if it loses a pawn?

Why are you trading pieces when you are down material?

Okay, so I'm probably more prone to mistake than you are, but I have a couple of questions/comments:

I agree with your view of Bc4, although it is a matter of taste.

In your analysis of 10.Ng5, you give exf5 Bxf5 Qe1+ Kd7 as a possible alternative. Why does Black have to play Kd7 and not something like Ne7? Is this because of possible overloading with Bg5 and Nd5?

What was the point of 15.Bd5? There probably weren't too many good moves, but you could have tried something else...I think =/

Anyway, I gave it my best as far as analysis goes. I'm bad at this kind of stuff, so you may actually end up correcting my comments...

@mnag: f5 is a common pawn-push in the Grand Prix Attack, with the point of trying to break open Black's Kingside defense. In some cases, it resembles a gambit, but usually White gets reasonable compensation for the pawn.

This basically transposed to lines common to the Grand Prix attack in the Sicilian, except that white lost a tempo and black played bd7.  F5 is the common way to pursue the attack.  I doubt that the lost tempo in exchange for bd7 helps white, but anyway, I bet if you look at those games you can get a sense of how the white attack is typically conducted.  I used to have a short book by Hogdson with a good amount of material on this system. 

Yes, Nd5 and Bg5 would win the knight unless Black resorted to Bf8.

15.Bd5 was to try to cover g2 at least and I couldn't find a good move anyways.

Good point. I don't know what I was thinking.

First of all, I agree with your observation after 8.Bc4!? that Black's Bd7 is a mixed blessing for him. It's not real development if your developed piece is in the way.

I like your fighting spirit with 10.Ng5?! but I'm afraid you're right: 10.exf5 is just a better move (as is 10.Bf4, I think). In fact, I think you're doing fine after 10.exf5. After 10.exf5, Bxf5 (10...Nxf5 11.g4! doesn't seem better) 11.Ng5, Nh6 you have 12.Re1+. I think White has a small advantage here, and it looks like the kind of game you wanted to play.

Your suggestion of 13.Bf4! seems right on to me. GM Jacob Aagaard, in his book "The Attacking Manual," has a chapter entitled "Bring all your toys to the nursery party." Without even reading the book you can see that he means for you to get all your pieces into the attack.

Unfortunately for you, your opponent played well (I hate it when they do that!) and his Be5! was a nice move, after which you were a bit worse.

I play the Grand Prix variation myself from time to time, and it's a great way to refine your attacking technique. Unfortunately, I've learned the most from the good moves I missed, and the painful defeats that followed. So this should be a valuable game for you!

As I think about it, I think the reason for your faulty 13th move is psychological. If you had believed in your position, you would have been thinking: What can I do here to strengthen my attack? Instead, I think you were thinking: I'm a pawn down! I'd better get it back!

Very true. I was afraid that my attack was petering out, so I decided to win my pawn back. I think part of the reason was also because I was playing a NM.

On a side note: several people seemed to approve of 8.Bc4, so now it is the bobobbob variation! It's definitely worth a second try though.

To me you played too directly and wildly (I'm not a big fan of Bc4 there, and if you wanted to play with the bishop on c4, why not put it there right away instead of giving time for black to play ...Nd4?). You were pinning all hopes on an attack but you let black get control of the center so even with his king on d7 it wasn't a big deal because his pieces were actually the more active. I don't know if black was better after ...kd7 (well after nxf7? I think he was, don't know about bf4), but his position was certainly comfortable, probably at least equal. After you took the pawn you allowed black's pieces to completely activate, so bf4 probably was better, and you have reasonable compensation but nothing too special.

From what I've observed, Bb5 seems to be played more often than Bc4 these days. And it's not clear to me that playing Bb5 and then Bc4 represents a pure loss of time. Black's B on d7 is not very well placed. I do think 8.Bxd7+ is probably the best move, but Bc4 is an interesting try.

No, the main problem I have with it is that black wins time with ...Nd4 (counter a wing attack with one in the center, and better yet, with tempo!), not so much with ...Bd7.

In the book "The Chess Openings for White, Explained" by Alburt and Dzindzichashvili,  one of their mainline variations goes 1.e4 c5 2.Nc3 Nc6 3.f4 g6 4.Nf3 Bg7 5.Bb5 Nd4 6.O-O! With the comment "White freely trades off his white squared Bishop, which only gets in the way of his attack." So theory apparently doesn't consider Black's Nd4 to be a serious problem

That's 6 0-0 though, not 6 Bc4. I would agree the bishop, when moves like d3 are played, it tends to get in the way of things, since on c4 it can be kicked easily, making white try to be very aggressive, while on g2 like in a closed sicilian it's somewhat inactive, though it does have the good point of eyeing the center at all times and protecting the kingside.