On Sunday, November 22, 2015 at 2:38:31 PM UTC-5, Phillip Helbig (undress to reply) wrote:


If I go into a cafe and see three people more than 2 metres tall, I
would be surprised. If I surveyed all the people in a large country, I
wouldn't be. The larger the sample, the more we should expect atypical
members of the sample. Check out papers by Peter Coles and Ian Harrison
using extreme-value statistics. They debunk many claims that extremely
large objects somehow contradict established theory. (Interestingly,
Stuart Coles (no relation, as far as I know) at the University of
Bristol has written a book on statistical modeling of extreme values.)

------------------------------------------------------

But say there is an unexplored island and conventional theory
predicts that because of its likely environmental conditions,
probable food resources and isolation, the inhabitants should be
pygmies.

Then the island is explored for the first time and numerous people
over 2 meters tall are there to greet the explorers.

One could not hide behind crude statistical arguments based on
faulty assumptions, but would have to admit that theory got it
wrong.

RLO
Fractal Cosmology