Minimal observed hands to identify a loose player

[Python]

I wondered how many hands pokeroffice/tracker has to log of a specific player to tell for sure, if he is loose or not.

I've calculated a little chart for that. On the x-axis, there are the observed hands and on the y-axis, there is the observed percentage of seeing a player on the flop.
A loose player is defined as being more than 30% on the flop.
If the observed percentage is above the red or green line, the player is a loose player with a probability of 95% resp. 99%.
(The line wobbling is caused by low statistics using binomial distributions).


Derived rules of thumb:
after 20 hands => min. 0.6
after 40 hands => min. 0.5
after 100 hands => min. 0.4