Dear all, This is probably off topic, but I did not know where to turn and since most of you have an excellent knowledge of statistics, I thought of sending my question to this list. I was wondering whether there is some literature on (or a way to calculate) how many data points one would need to constrain the shape of a PDF to such an extent that this can be statistically differentiated from another given PDF. I am aware that this depends on many things and the answer is not straightforward, however my problem is very simple and would need a very simple answer. I would really appreciate your help Thanks Tibi