Re: Inferences from a normal distribution
From: Richard Ulrich (Rich.Ulrich_at_comcast.net)
Date: 01/01/05
- Previous message: Josh: "Inferences from a normal distribution"
- In reply to: Josh: "Inferences from a normal distribution"
- Next in thread: Josh: "Re: Inferences from a normal distribution"
- Reply: Josh: "Re: Inferences from a normal distribution"
- Messages sorted by: [ date ] [ thread ]
Date: Fri, 31 Dec 2004 20:36:24 -0500
On 31 Dec 2004 16:30:28 -0800, "Josh" <josh.newspersona@gmail.com>
wrote:
>
> A friend of mine tried to convince me that because returns on stocks
> are normal, there is no way to predict price movements.
Of course "returns on stocks" predict price movements --
That is called Insider trading, and you can go to jail for it.
So, I'd say that this premise is simply wrong....
> I am inclined
> to believe that you can't predict price movements, but I'm not sure
> that this is why.
I think the notion is -- if the outcome was determined
by a whole lot of tiny influences (which maybe were not
measurable), then it should be distributed 'normally.'
Therefore, if an outcome is distributed as normal, maybe it was
unpredictable. I think this is the same style of logic by which
Von Daniken proves that space aliens build the pyramids.
- If a simple predictor is itself distributed normally, then its
linear outcome also would be 'normal.'
[snip, other stuff]
-- Rich Ulrich, wpilib@pitt.edu http://www.pitt.edu/~wpilib/index.html
- Previous message: Josh: "Inferences from a normal distribution"
- In reply to: Josh: "Inferences from a normal distribution"
- Next in thread: Josh: "Re: Inferences from a normal distribution"
- Reply: Josh: "Re: Inferences from a normal distribution"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
|
