(50g) Normal Distribution

[Dieter]

(12-04-2018 11:06 PM)Albert Chan Wrote:  I think direct method is good enough (9 digits precision, even with extreme case)

The point is not that you get nine digits, but that you lose three. The program in the HP67/97 thread with the improved was designed to achieve close to machine precision for most cases, i.e. roughly 9 out of 10 digits. If one intermediate result returns only 7 valid digits this is an absolute no-go.

Dieter