[MNpoker]

Here's a way to calculate EV and Variance using your Session notes.
Whether you want to believe the curve is normal or not is up to you, but according to statistical theory it really should be pretty close.

A) Column A be the number of hands played. (Each session will go on one row)
B) Column B be The number of Big Bets won / lost in the session.
C) Column C = Column B ^ 2 / Column A
D) Get the Sum of Column A (The total number of hands played)
E) The sum of Column C
F) The number of sessions (This will equal the number of rows)
G) Find the value of The Sum of Column C * 1 / N <-- the number in (F)
H) Find the sum of columnB / sum of column A. THIS IS YOUR AVERAGE WON PER HAND
I) Take the number found in (E) * (H) ^2 / (F)
J) Take the number in (G) - (J) <-- This is your VARIANCE PER HAND

Now to find your Confidence interval.
95% (Probability all at the top) = 1.645 * (Number of Hands * Variance Per Hand) ^ .5
For 90% use 1.282, 99% use 2.326, 99.9% use 3.090

So For example if your E(BB) per hand is .02
And you Variance Per hand is 3.5
And you have 20,000 hands.

You have a 80%* chance that your Expected win rate for 20K hands is between 20,000 * .02 +/- 1.282 (3.5 * 20000)^.5 =
400 +/- 339.18 OR between .3 and 3.7 BB per 100 hands.

* Note this is 2 sided so we goto 80%; 10% chance over the top and 10% under the bottom.

Or in other words A player with these figures has a 90%+ chance of actually being a positive player overall (and not just lucky). It's actually 93.47%.

Disclaimer: As people mentioned before these numbers assume you are playing the same game v. the same players and you are playing at the same skill level throughout.
Though if you are improving and the games on average are the same the number you come up with should actually be low.