Sample size and reliability of flops seen percentage?

[Adde]

How big a sample is needed to draw any conclusions about [flops seen %]?

If I have PokerTracker data that says a player is seeing 30% of the flops, how reliable is this number given a sample size of 10/20/30/40/50/100/etc hands? Anyone know of a formula to calculate this?


Adde

[bigpooch]

Don't put too much reliability on that statistic! It is
much more important to make what you believe (and better
if it is in reality!) are decisions with the largest +EV.
Flop% doesn't take into account that you are in the blinds,
whether the game is now 9-handed or 8-handed or game
conditions have changed so you can see more flops (maybe the
game has become very loose passive) assuming you are playing
at a ring game. There are so many factors that are much
more important than this number.

[Adde]

I see your point.

Still, I think there has to be some validity in knowing the flops seen % contra not knowing it. (And as for my case, the blinds are counted for, using [voluntary pay money to see the flop]), and I only use 8-10 player tables in the stats).

If you have a sample of 20 hands and a player is seeing 50% of the flops, isn't it more probable that he is loose than not?