The Standard formula for pIntMax is correct

[John Gustafson]

The Posit Standard (2022) says that pIntMax, the largest consecutive integer-valued posit value, is

ceiling(2^floor(4(n–3)/5))

and someone pointed out that the "ceiling" function is unnecessary because the quantity being rounded up to the nearest integer is already an integer. And I agreed, and thought we had made a mistake.

But you need that ceiling function, if n is allowed to be as small as 2. You wind up with 2^(–1) = 1/2. But the value of pIntMax  for n = 2 is 1. The ceiling function fixes that one case. I'll attach a graph of what floor(4(n–3)/5) looks like as a function of n, for clarity.

John