Note: is derived as theoretical value of error!
(replace in denominator by to get experimental
value).
So we only get something different by using the full maximum
likelihood method, if some are small and Poisson distribution is
not in regime where the distribution is essentially Gaussian.
is large
method.
is derived as theoretical value of error!
(replace 
in denominator by 
to get experimental
value).
are small and Poisson distribution is
not in regime where the distribution is essentially Gaussian.