Skip to content

nct.cdf becomes 1.0 for nc >= 65536 #3209

Closed
pv opened this Issue Jan 13, 2014 · 1 comment

2 participants

@pv
SciPy member
pv commented Jan 13, 2014

From stackoverflow:

"""
I encounter a problem using the nct.cdf function in scipy.stats!

nct.cdf is the t-Student cumulative distribution function, with shape parameters df (degrees of freedom), nc (non central parameter).

Curiously enough, this function works well for nc<= 65535

YET FOR nc>=65536

You will notice that for any x, and for any nc>=65536:
nct.cdf(x,df,nc)=1.0 which is completely absurd!

For instance:

from scipy.stats import nct
nct.cdf(45,9,65536) # gives 1.0 ????? (It should normally be close to 0)

"""

@richardtsai

I've fixed it when working on a related issue (#2667). A PR will be made soon.

@pv pv closed this in #3222 Feb 24, 2014
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.