A recent game I played:
White played the grand prix attack without Nc3 inserted which allowed me to play a pawn sacrifice with d5 for lead in development and active piece play. Later my adv. in development turned into a little attack against his king, however in the critical moment I should have just got the pawn back with a microscopically better position. I missed some tactics as well and got under huge time pressure which eventually costed me the game. This was a 15min game by the way
e4 c5 2.f4 d5 
This is supposed to be the refutation of white's line. Thus if white wants to enter the grand prix attack it's better with Nc3 first which stops the d5 advance.
3.exd5 Nf6This is the whole point! If white defends the pawn with c4 Black gets quick development and pressure on the d file and the d4 square for his pieces after an e6 pawn push. White can't really allow black to take on d5 unhindered as black would be simply better in that case because of white's weak d pawn and squares on the d file
4.Bb5+Logical, white is willing to trade his lightsqured bishop before playing c4. Not only this gets the bishop outside the pawn chain it also helps white to catch up in development.
Bd7 5.Bxd7+ Qxd7 6.c4 e6 7.Qe2This seemed like a contraversial idea for me, although if I am not mistaken it's the main line. The problem with this move is that what will eventually happen is I'll get a rook on the e file way faster than white can develop and play e5 with his queen so exposed to potential tactics.
On the other hand an endgame if white doesn't play Qe2 but goes 7.dxe Qxe+ Qe2 Nc6 is reasonable for black as white's d pawn is awfully weak.
I would really like to hear your thoughts on this oneBd6 8.dxe6 fxe6 
The e pawn isn't really much of a problem as I'll get rid of it in time by opening up files in the center. White can hardly stop that as I am ahead in development
9.d3 O-O 10.Nf3 Nc6 11.O-O Rae8 12.Nc3Finally the time has come for e5. Black's pieces just spring to life and white is likely to loose his extra pawn just because of the pressure
e5 13.fxe5?I don't like this at all. Just seems conceptually wrong. Black is ahead in development yet white is willing to open up the game for black!

Besides white trades his only developed pieces

Better was 13.f5!
analysis diagramalthough it gives the pawn back it blocks black's pieces and allows white to catch up in development. Black retains a slight edge I think however gets some squares to work with too: d5 and e4 look good home for the knights
Nxe5 14.Nxe5 Bxe5 15.Ne4 Ng4?!A miscalculation

Why on earth didn't I trade another of his developed pieces?%^?$!@%!*?

A clean cut way of transforming my lead in development to a clear positional advantage was 15. ...Nxe4! and black will get his pawn back with excelently placed pieces. For example 16.Rxf8+ Rxf8 dxe Qd4+ Be3 Qxe4 and black is clearly better. There are some minor tactics which I leave for the reader

If you are not sure, ask! I'll gladly answer
16.Rxf8+ Rxf8 17.g3 Bd4+ 18.Kg2 Nf2 Computer analysis
Hidden Text Below - [Show it] - [Hide it Again]
According to Fritz10 18. ...Qe6! wins for black!

With a very simple threat of Rf2+ as the knight is pinned. Pinning that knight is the key it seems. I completely missed that both during the game and during analysis. So it turns out that Ng4 wasn't so bad! On the other hand I just didn't see this follow up

Thus more important is the fact that my thought process that lead to the move was wrong.
19.Be3 Qh3+?!
It does look scary for white! But this was overly optimistic on my part as I didn't really see a clear way my attack to succeed. I should have just regained the pawn with 19. ...Bxe3 Qxe Qxd3 which would probably lead to a draw. The rest of the game is just me trying to get an imaginary attack which just doesn't work and misplaces my pieces on the kingside. White can always give up a pawn on the kingside just to get the queens traded and leave my pieces misplace there.
20.Kg1 Bxe3 21.Qxe3 Ng4? 22.Qe2 b6 23.Re1 h6 24.b4!? Rf5 25.bxc5 bxc5 26.Nxc5?This only lead to a perepetual check which I certainly should have accepted! Somehow I thought I was getting mated

This explains the caveman
Kh7?
27.Ne4 Rh5 28.Rf1 Nxh2and here I lost on time
What I've learned:
1. Stop playing "hope chess"! It's surprising how many times I repeat that to others and still do the same mistake myself. This is a very bad habit I have, lazyness...

it can be noted in my analysis as well, although I am slowly trying to correct this
2. 15. ...Nxe4 wasn't that hard to find, was it? This is pretty much the same as point number 1. Try to calculate the resulting complications better. As I pretty much gave up on this move because I looked at it like 5 seconds and saw that the e pawn is defended. Checks and captures, checks and captures, checks and captures....
3. Time managment! I'll go read an article by Dan Heisman or something
