Snidely submitted this gripping article, maybe on Wednesday:

The statistical comparison that counts is not at all meaningless: can the
crew expect to get to spoace, perform their mission, and return?

Statistical comparisons delaing with individual components may be
problematic, but I don't see a difference in counting mission failures.

Crap, I need to fire my proof reader ... the statistical properties of
that component are not staying within guidelines.

/dps

--
"I am not given to exaggeration, and when I say a thing I mean it"
_Roughing It_, Mark Twain