Breakdown by # of players

[sprmario]

Hi I'm new and have just started playing 3 weeks ago. I am trying to figure out useful PokerTracker stats for SnG tournys and have been looking at how I do w/ the number of players left in the tournament. In the Position stats tab I'll set the min and max # of players to the same and look at the total won or lost w/ those number of players. I have about 150 tournaments in my database now (w/ ok results). I seem to have 3 trouble areas.

10 players. I am totally break even w/ 10 players. I tend to be really tight early as this is when I'm most likely to face unpredicible opponents. I usually don't raise w/ much more than AA, KK, QQ, AKs and AK. I'll play lower PP for sets, and other suited connectors from late position if there are a good number of limpers. I'm break even so this isn't a huge problem.

5 players left. Being tight early this is where I am often finding myself the short stack and closing in on or in level 4. I might be at 500 or so and I have to be more aggressive which is costing me some losses. What concerns me is that I'm significantly profitable when you look at playing at 6 to 10 so I feel as if I shouldn't be here as often as I am. I have a lot of 6th place finishes, which I'm not happy about.

2 players. When breaking it down like this, I'm doing horribly. I am not doing well heads up at all. I am usually the underdog when I make heads up, but even when I do go in as a favorite w/ the chip lead I'm coming out on the short end. I definately need to work on my heads up game because if I could convert a few more 2nd places into 1st places I'd be much more profitable. Just to give you an idea so far this month (i'm going to break my results down monthly) I have 30 tournaments w/ 10 2nd place finishes and 2 1st place finishes. This is pretty consistent using the other 120 tournys as well.

Overall I'm doing well this month 30 tournys, 53.3% ITM, 45.5% ROI (I think this is slightly higher than my true return or ITM).

Basically I'm looking for opinions on whether or not this is a valid way of trying to analyze your game. I know that the blinds might be significantly different in one game vs another w/ a given number of players but over a large sample (250 to 300 tournys?) this should average out. Ideas?