Little explorations with HP calculators (no Prime)

[pier4r]

(04-03-2017 06:50 PM)DavidM Wrote:  Despite the closeness in fitting (\(R^2\) is 0.9999 or better in both cases), the question still remains as to whether rounding and/or ceiling the results will provide a useful replacement for the table lookups. Hopefully these (or something like them) will give you what you need!

Thanks once again. Already your first function of the two posted was great, I'll check those of yours.

My need is not so strict, what I need is: I would like that a certain variance modifier appears more than other modifiers, so I put probability tokens in the middle (little modifiers) rather than on the tails, like a triangle. The ideal would be a gaussian bell but that's too refined.

So, since rounding makes things ok for the interval steps of 25 units between two modifiers, the problem can happen only if an approximate formula gives too much chances to a modifiers that should not have that much. But if the approximate formulas are pretty tight, since we are talking about thousands of units (I choose large amounts on purpose), then it should be fine. Anyway I'll test once I have the calculations with the modification finished.

Could be a nice side exploration.