Peter:The original formula was not completely correct in that it involved some approximations which were not completely valid. The adjusted formula should be more correct, but is too complex to calculate precisely. My estimate of the result given by the new formula is in the same region as the observed figures for resh. But in fact I would be rather surprised to see a precise fit, for the basic assumptions I made in the hypothesis, that verb roots are randomly distributed with no tendency to avoid homonyms, seem so unlikely.
I'm having an intuitive reaction that something doesn't seem right
here. I don't know enough to be able to tell what is wrong, but if the
formula is correct, it should not be giving the obviously incorrect
response for the resh.
I sent a copy of the spreadsheet with some explanatory info. to aI would be interested in any response.
friend who has a PhD in statistics, but he hasn't responded yet.
Archive powered by MHonArc 2.6.24.