Question: Some details about ChiSquareSuitableModelTest

Hi,

I seeking for informations on the Statistics:-ChiSquareSuitableModelTest procedure:

1. Once you have choose the number of bins, what strategy does this procedure use to define the bins (equal width, equal probability, other one?).
    
2. It seems the procedure accepts any value for this number of bins and that its correct use then is under the sole responsability of the user. Am I right?

In the book below (but I'm sure this can also be found on the web) there is a detailed discussion concerning "good practices" in using the Chi-Square goodness of fit test: does anyone known is something comparable is used in ChiSquareSuitableModelTest ?

Statistical methods in experimental physics, W.T.Eadie, D. Drijard, F.F.James, M. Roos, B. Sadoulet
North-Holland 1971
Paragraph 11.2.3 "choosing optimal bin size"


Thanks in advance