The results of Prism's binomdist function does not match Excel's.

The Prism Statistics Guide (that comes with Prism 4) on page 139, and the Fitting Models  book on page 326, mistakenly describe the equivalence between Prism's and Excel's binomial function.

The description of binomial function is correct (it is the cumulative function of k or more "success" in N trials). But the Excel's function BINOMDIST(k, n, p, true) calculates the "cumulative distribution function, which is the probability that there are at most k successes". 
 
So the equivalence between Prism's function (on the left) and Excel's (on the right) is: 
 
     BINOMIAL(k, n, p) = 1 - BINOMDIST(k, n, p, true) + BINOMDIST(k, n, p, false)
 
 
 

(This was fixed in the third printing of the manuals.)