the formula appears to list the records sequentially as they appear in the data set.

So if I want 25 results for letter "T" it list the first 25 records it finds, anything after this is hidden.

Therefore if I have a total of say 150 records that are coded "T" number 1 - 150 the formula seems to list values 1 - 25 while hiding 26 - 150 thus not random.

Hope that makes more sense.