"clipka" <nomail@nomail> wrote:
> Still there must be some mathematical solution to this problem...?!

Looks like this is (or at least comes very close to) what I need (particularly
the "calculate sample size" function):

http://www.quality-one.com/products/inttools.php

If I understand it correctly, you feed it the number of failures observed during
testing, the "confidence level" (i.e. how sure you want to be about the result)
and the desired reliability (i.e. the percentage of failures to be expected "in
the field"), and it spits out the "sample size", i.e. how often you must have
made the test in order for the number of observed failures to be considered
acceptable.

Can anyone come up with a forumla for this?

Post a reply to this message