Re: Freight car distribution

Walter M. Clark

--- In STMFC@..., Malcolm Laughlin <mlaughlinnyc@...> wrote:

Posted by: "wmcclark1980" What we have is the problem that always
occurs when using statistics.
Most people won't accept that a relatively small sample size can
forecast much larger populations with extreme accuracy.

That's not what the discussion is about Walter. For that theory
to work you need a population that is homogenous with respect to some
variable. With railrad box cars, we are working with lumpy
distributions with varying degrees of skewness and nothing near a
normal or poisson or other kind of distributionfor which we know how
to calculate statistical significance.

Malcolm Laughlin, Editor 617-489-4383
New England Rail Shipper Directories
19 Holden Road, Belmont, MA 02478

But Malcolm, the population of people and opinions, or even
birthdays, is skewed and lumpy, and statistical methods work for
those. Why not box cars? The result of Tim's and Dave's analysis,
fed into Larry's spreadsheet, matches the available data. That's good
enough for me.

Time stopped in November 1941
Walter M. Clark
Pullman, Washington, USA

Join to automatically receive all group messages.