[Python-ideas] random.choice on non-sequence

Stephen J. Turnbull stephen at xemacs.org
Thu Apr 14 04:16:06 EDT 2016

Greg Ewing writes:

 > The limit of p = 1/n as n goes to infinity is zero.
 > Events with zero probability can't happen. I don't
 > know how it can be made more rigorous than that.

Sorry, as Steven pointed out, events *that are modeled as* having zero
probability happen all the time.  What you should not do with such
events is give them positive (ie, atomic) weight in statistical
calculations (= theory), and therefore you should not *predict* their
occurance as individuals (unless you have no fear of becoming a
laughingstock = practice).

 > Infinity: Just say no.

I thought your name was Ewing.  Is it really Brouwer?<wink/>

(Don't bother, I know.  The reference to Brouwer is close enough for

[1]  https://en.wikipedia.org/wiki/Glossary_of_owarai_terms

N.B. "Straight man" is not an accurate translation.

More information about the Python-ideas mailing list