Statistics

[Don Smith <dasmith at rotse2_physics.lsa.umich.edu>]

Hi,

Look, I don't really want to get into a debate about relative risks or trying
to really quantify how risky it is to drive to the store, but I do feel
constrained to point out that you guys are not doing the statistics properly.
You can't point to the fact that you (or anyone you know) hasn't died in a car
accident in umpteen trips, while the two shuttles have exploded.  That's a
posteriori statistics, and you need to use Bayes's theorem to do it right.  The
chance of *someone* winning the lottery is 100%, the odds of *you* winning the
lottery are vanishingly small.  So it's possible I may be wrong in saying
driving is more risky; I made that as an off-the-cuff comment.  I confess I
have not tried to calculate the odds.  So I retract the comment, since I won't
back it up.  However, that doesn't change the fact that you're calculating the
statistics incorrectly.  :-)

On to other topics,
-- 
Don Smith                           Robotic Optical Transient Search Experiment
donaldas at umich_edu                                 http://xte.mit.edu/~dasmith/

---------------
Unsubscribe by going to http://www.actwin.com/OtR/