Re: A challenge to non-SRians

From: Paul Miller (aksuited_at_yahoo.com)
Date: 08/01/04 

Date: 1 Aug 2004 05:06:33 -0700

> One does not risk money until the conditions of the bet have been agreed
> upon. So, we need to agree on the outlines of the condition of the bet.

Agreed.

> Actually, it is vice versa. One must first understand what one considers
> acceptable 'confirmation' statistics, before one designs the details of the
> experiment.

Ok, well presumably it's in both our interests to make the experiment
as accurate as possible, as long as the experiment is still feasible
to carry out.
 
> > You've just said, 'it doesn't take high
> > precision to demonstrate SR is wrong'. Ok, let's do that then. Let's
> > design an experiment which does just that, and bet on the outcome. I'm
> > willing to bet $100,000 to your $1000, on the outcome of that
> > experiment. Are you?
>
> Yep. Now, I need to know what statistical confidence level you will accept
> as a 'win'/'lose' point.

> There are no general requirements for 'rigour' for a journal article.

Well, I realise that; I'm just flagging up the obvious point that we
need a well-controlled experiment, that's all. There's nothing
underhand intended here.

..snip..

> Perhaps you really are a neophyte to the SR arena. So, I'll give you a
> general description of how the experiment will work in the case of SR, and
> in the case of non-SR. However, I'll still need from you the confidence
> level that you require for the win/lose critierion -- before the details can
> be identified.
>
> Here's the simplest that I can think of (it may need to be more complex,
> depending upon your win/lose criterion).
>
> A source of light pulses will be located at a position (S). There will be a
> target/detector/reflector (T) placed a long distance away (some miles).
> Each location will have a recording clock, synchronized to the GPS time (no
> e-synching allowed). A light pulse will be sent from the source to the
> target and back (at least two 'bounces'*). The arrival times will be
> recorded. The orientation of the source-target will be changed, or
> duplicated with other source-receiver combinations.
>
> According to SR, the time-of-propagation will always be exactly the same on
> both trips (S-T and T-S). According to aether theory (LET, for example),
> the time-of-flight can vary.
>
> Let us say, for example, that the anisotropy in the CMBR background
> indicates a local aether 'rest' coordinate system (which is commonly viewed
> as reasonable -- at least among aetherists). Then there will be a
> difference of about 350 kps in the light travel speed. Or, about 1 part in
> 1,000 of the expected light speed. Hence, our system must be able to time
> differences of less than 10 ppm of lightspeed. Since light moves about a
> foot (0.3 meters) per nanosecond, the accuracy of our available clocks --
> combined with your win/lose criterion -- will determine how long the
> baseline must be.
>
>
> * This will eliminate any systematic errors due to timing measurement
> variation for emission/reception.
>
>
> What precision will you accept? I'd wish to see something on the order of 1
> part in 10,000 of light speed (30 kps).

Ok, I'll have a think about your proposal, and get back to you about
the desired precision. All sounds reasonable to me so far.

Paul