George Gamow talked about the incredibly small probability of a car
tunneling into a garage. I believe the probability is incredibly
small, meaning it *is* nonzero *and* I won't believe anyone's number
for it as being believable within an order of magnitude - of its
order of magnitude! We aren't yet wrong, Margaret.
Paul's icepick problem is easy. My roommate at Cal came home with a
much harder variant. His prof asked him what the expectation value
was for the number of times an ideal ping pong ball could be bounced
from a similar fixed ball when dropped optimally from an initial
vertical separation of ten radii. I didn't solve the problem for him
(I hadn't had QM at that time) so we did the experiment and
concluded the answer was less than two. I did the (messy) calculation
a year later, but I forget what I got.