Re: K random numbers

Ben Bacarisse wrote:
stan <smoore@xxxxxxxx> writes:

<Fisher-Yates stuff snipped>

Kinda makes one wonder what they were doing when this came up.
(algorithm not the time machine)

Fisher was the Fisher who virtually invented modern statistical
methods. In 1938 he had just left the Rothamsted Experimental Station
to take up a position at UCL (somewhat embarrassingly in eugenics)
thereby allowing his colleague Yates to take over as head of
statistics. They published the method in a handbook called
"Statistical tables for biological, agricultural and medical
research". Random permutations are obviously important in
experimental design.

Durstenfeld was the first to publish the algorithm for an electronic
(rather than a human) computer and I think deserves some credit
because the Fisher-Yates method is not what we know today. It was
Durstenfeld who came up with the swap. The fisher-Yates is based on
crossing off items from a list, repeated counting from the start and
skipping items that are already crossed off.

[I know you didn't *really* want to know what they were doing, but I
like this kind of stuff so I chose to take you literally.]

Actually I was curious, but I intended it as rhetorical because I
couldn't imagine anyone had a reasonable answer.

Thanks, you're pretty good. Any word on the time machine?

.

Relevant Pages

  • Re: K random numbers
    ... Are you listening, Mr Fisher-Yates? ... statistics. ... Durstenfeld who came up with the swap. ... skipping items that are already crossed off. ...
    (comp.programming)
  • Re: Unsorting a TStringList
    ... but I ran some empirical tests on the two ... algorithms (Fisher-Yates and your suggestion). ... For a two element set both algorithms perform basically the same. ... for Fisher-Yates, biases in the other algorithm), so Fisher-Yates would ...
    (comp.lang.pascal.delphi.misc)