HP Prime Calculator - NORMALD_CDF and NORMALD_ICDF Problem

[TexasEngineer]

The normal distribution functions do not work as documented.  The inputs are mean, standard deviation, and either X or probability depending on the function.  In all cases, the functions ignore the input values of mean and standard deviation and assume 0 and 1 for these values.  Is there a planned fix for this problem?

[uklo]

Hello, try updating the software version to 5106. Mine works fine, in both cases the result is correct. For example NORMALD_CDF(0,1,2.32) = .989829561331 and NORMALD_CDF(1.5,.8,2.32) = .847318406167

[TexasEngineer]

---

@uklo wrote:
Hello, try updating the software version to 5106. Mine works fine, in both cases the result is correct. For example NORMALD_CDF(0,1,2.32) = .989829561331 and NORMALD_CDF(1.5,.8,2.32) = .847318406167

---

I have Version 5106.  The function works fine apparently if you are in Textbook mode.  I use RPN mode and the functions do not work as indicated in RPN mode.

[BartdB · Accepted solution]

Hi,

You may wish to post your problem here:

https://groups.google.com/forum/?fromgroups#!forum/comp.sys.hp48

There are members of the HP calculator team that browse that newsgroup and the Prime is discussd there.

[TexasEngineer]

Thanks for the suggestion.  I received the following response from Joe Horn:

Not a bug.  Both of those functions allow a different number of arguments (either 1 argument or 3 arguments; see their help screens). Therefore you have to specify how many arguments you want them to take from the stack if you want them to take 3 arguments instead of the default 1 argument.  The syntax for this is normal_cdf(3) and normal_icdf(3).  Do the same for all functions which are able to take different numbers of arguments.  E.g. MAX(7) consumes 7 stack levels and returns the max value from them.  Ans then returns all 7, which not even the HP 50g can do (its LASTARG returns a maximum of 5 arguments).