>Does a random sequence contain information? I certainly hope so:
>
>Regularly, in most of my lectures, I refer to a "random sequence" of
>random variables. If they do not contain information, then I, and my
>fellow statisticians, have done a fantastic job of hoodwinking the world.
>(I'll keep my paycheck, thank you. :-)) The entire subject of statistical
>inferences is based on the premise that random sequences CAN be gleaned
>for useful information ("information" used in a nontechnical sense). Apart
>from this reality, most of modern science could shut down. (And Glenn
>could stop telling his company where to dig its next oil well -- allowing
>him even more time for the reflector. :-))
>
>It is true that most statistical models assert that a random sequence
>contains a COMPONENT of "noise" -- devoid of information. Perhaps this is
>what some discussants are thinking about. Beyond this, any useful model
>contains a component for "signal" or "the true measurement" - apart from
>noise, or whatever, depending on the context. Sometimes the "signal,"
>itself, is permitted some randomness as well.
Thanks. I thought I was saying that, albeit in other words, but I guess I
should have checked with you first!
Art
http://chadwicka.swau.edu