🚀 go-pugleaf

1) Most statistical methods attempt to generalise from a set of data
   by making assumptions about the underlying population (e.g. `Normal
   with unknown mean & variance'), estimating the unknown parameters
   (here mean & variance) and using the properties of the fitted
   distribution (here Normal) to, for example, predict future data.

   `Bootstrapping' goes straight for the jugular (bypassing the brain
   almost entirely :-) and uses the observed data distribution as an
   estimate of the population distribution (i.e. probability mass 1/N
   at each of the N data points).  Properties of the assumed underlying
   population are inferred from the corresponding data properties.
   (And if it's too hard to do analytically, simulate by generating
   further `bootstrap samples' from the data distribution).

   Bootstrapping is the ultimate in maximum likelihood estimation
   (attempts to produce a Bayesian analogue are, IMHO, weird).

2) Biasedness of the SD is hardly mentioned because it's too hard.
   (And in almost any practical analysis, biasedness is the least
   of my worries anyway).

                -- Ewart Shaw
-- 
J.E.H.Shaw,  Department of Statistics,  |  JANET:  strgh@uk.ac.warwick.cu
             University of Warwick,     |  BITNET: strgh%uk.ac.warwick.cu@UKACRL
             Coventry  CV4 7AL,  U.K.   |  PHONE:  +44 203 523069
yacc - the piece of code that understandeth all parsing