At 4:00 PM -0800 3/2/00, [log in to unmask] wrote:
>We'll perform a simulation to determine which strategy works best. The exact
>mechanics, rules, and verification methods for the simulation will be worked
>out to our mutual satisfaction before the test, but it will essentially be
>done by generating a list of random evenly-distributed "winning" values from
>[1,2,3] of appropriate length (say 10,000 or so), picking a random
>equally-distributed "selection" value from [1,2,3] for each item in the
>list, discarding one non-winning value, then switching the selection to the
>remaining value and comparing it to the winning value.


Gavin has already posted a BASIC program that does exactly what you
describe. His simulation consisted of 1000 iterations, but it might as well
have been 1000000. The results were something like 634-366 in favor of
switching doors. Looks to me like the evidence is in. You have a 2/3 chance
of winning if you switch doors, and only a 1/3 chance if you stay put.

How about we let the goat behind door #2 get some sleep, shall we?


Barry Lake                                 [log in to unmask]
Allegro Consultants, Inc.                  www.allegro.com
(408)252-2330