Re: [Phys-l] projectile motion lab

[Brian Whatcott <betwys1@sbcglobal.net>]

At 10:49 PM 10/10/2006,  Krishna you wrote:

> [bw]
> p.s. Supposing a Gaussian distribution for a Monte Carlo
> >   seems circular to me: what am I missing?
>
>
> This could be a misunderstanding on my part.  I thought that you did Monte
> Carlo using a distribution that best represented your sample.  Is Monte
> Carlo always done using Gaussian distributions, so that what I said was
> redundant?
>
> --
> regards
> -Krishna
>
> Krishna Chowdary

Realizing that my p.s. appealed more to my prejudices than any rational base
I offer this story of an early Monte Carlo by hand operator - Fermi -
along with anecdotes of Metropolis and Feynman as Marchant repairmen
before Ulam had promoted his gambling uncle to legendary status in naming
 the method "Monte Carlo"; at least in the creation myth that I prefer...

  <http://library.lanl.gov/cgi-bin/getfile?00326886.pdf>

Notice that Fermi makes a random selection from a uniform distribution to
discover the form of a statistical distribution as an output.
 I have to acknowledge the wide menu of distributions that a M-C modeller
 may now choose from in current incarnations, all the same.

And in the toy Monte-Carlo that junior-high students may encounter,
it is the random selection from the uniform distribution in the square domain
that passes the Pythagorean seive for inclusion in the included circle
 leads to an estimate for pi....


Brian Whatcott    Altus OK    Eureka!