Re: [Phys-L] [SPAM] Re: animal magnetism?

[brian whatcott <betwys1@sbcglobal.net>]

Think of it this way: you might call it the Financial Market Modeller's 
Peril.

You start with an experimental data set, it might be a time series of 
stock prices.
You then start to examine factors which may influence the observed 
changes. It is almost certain that you will find some time series which 
will some time delay will provide an excellent statistical fit.    But 
because you have been bitten many times before (and likely as not you 
have an engineering or science background)   you will run this predictor 
with a sequence  of the same data set that you put aside for this 
purpose, and the statistical fit vanishes.    To put it another cynical 
way: if you try hard enough you can justify linking any data set to 
another data set.

Brian Whatcott   Altus OK



On 1/16/2014 7:14 AM, Anthony Lapinski wrote:
> Of course I did!
>
> I'm just trying to understand this dog/magnetism thing. Makes little sense
> to me. I don't really believe it, despite the "scientific evidence" I've
> read...
>
>
>
> Phys-L@Phys-L.org writes:
> > On 2014, Jan 15, , at 13:21, Anthony Lapinski <Anthony_Lapinski@pds.org>
> > wrote:
> >
> > > I believe moss is about sunlight and shadows.
> > I suspect Anthony did see the :-)   .
> >
> > bc did! and is still smiling.