Standard Deviation for VPIP?

[Spicymoose]

I am trying to figure out when I can have good approximations of my VPIP stat on opponents. For example, if I see someone has a VPIP of 90 after only 20 hands, although my sample is small, I usually think that their true VPIP has to be at least over 40 50% of the time. If their VPIP is 60 after 50 hands, I also think they are pretty loose.

Is there anyway to get an approximation for a confidence interval of VPIP after X hands? I would like to know how confident we can be after 10, 20, 50, 100 hands.

Thanks,
Spicymoose

P.S. I play 6-handed incase that would make a big difference.

[Girchuck]

Take VPIP from all of your sessions, and calculate the SD for them.
That should give you a very good estimate of true SD of your VPIP, which you can use as a starting point.
Then do the same for your opponents with most hands and sessions in your database. See how the numbers go.
Find the average.