Groups | Search | Server Info | Keyboard shortcuts | Login | Register

Groups > sci.stat.math > #10843

Re: statistics in Roberts' paper on Miller

Cross-posted to 2 groups.

Show all headers | View raw

Tom Roberts to Anton Shepelev:

> > [with continuous least squares fitting],
> > you would have been able to avoid combining opposite
> > orientaions and analyse the entire turns, with 15
> > degrees of freedom. With the half-turns combined, the
> > error differences beween opposite orientations are
> > "baked" into the partial curves and uncapable of
> > smoothing out.
>
> Yes, my approach could not handle 15 parameters. I'm not
> so sure that a conventional fitting program would reliably
> converge with that many parameters.

The linear least-squares method has a global optimum, so
that with continuous parameters and partial derivatives
convergence is guarranteed, nor is overfitting an issue. In
your case -- the tighter a fit, the better.

> > > So look at my Fig. 2 and say with a straight face that
> > > you think a signal with amplitude ~ 0.1 fringe can be
> > > extracted from the data.
> >
> > I do not have that Oscilloscopic, Harmonic-analysing,
> > Fourier-transforming vision that you seem to take for
> > granted :-) Yes, it looks awful.
>
> I am skeptical of ANY analysis that claims to pull a
> signal out of noise that is so very much larger.

I persume your experience entitles you to this opinion,
whereas we laymen, lacking statistical intuition, must
content ourselves with formal analysis.

> > The phase would indicate the direction, and the
> > amplitude -- the velocity of the aether wind speed as
> > projected upon the plane of the interferometer.
>
> Sure. That was Miller's intent. But with errorbars so very
> much larger than the variation in the data (Fig. 5), the
> errorbars on velocity will be enormous (and include 0),
> and the errorbars on direction will be much greater than
> 360 degrees. In other words. Miller's approach cannot
> determine the speed or direction of the "aether wind" at
> all.

I for one have not tested Miller's astronomical calculations
yet. If they are as good as he says then it must be a
miracle or one-in-a-million coincidence that the an absent
signal and a strong systematic drift of the device should
have produced a galactially-oriented signal.

> A modern analysis can reduce those errorbars considerably,
> as mine did...

Yet technically, your statistical model and its fitting are
quite simple.

> > The only justification is in the first sentence. Mr.
> > Roberts thinks he should use quantised model parameters
> > because the input data is quantised, whereas I see no
> > logical connection between the premise and conclusion.
>
> My analysis is fitting to find the DRIFT, not the signal.
> The DRIFT is inherently quantized by Miller's data taking.

Quantised are the obsererfations (o) of finge shift (f),
comprising signal s and drift d:

                      o = quantise(f)
                      f = s + d

Physically, none of these values is quantised, but the
method of observation quantises f. What harm can be in
fitting the drift in continuous parameters? If you must,
quantise them after the fitting.

> But before you get all excited about a potential aether
> model, first you must consider how it could agree with all
> the experiments that display quantum effects. No aether
> model to date has done so, and IMHO it seems EXTREMELY
> unlikely that one will ever do so. Electrodynamics is A
> LOT more than light beams and interferometers....

My current purpose being to understand Miller's results and
your analysis of them, I have not considered other
experiments and their reconcillication with an aether
theory.

> > No sceptic has been able to answer how all his
> > measurements made at different times of day and of the
> > year might have conspired to point at a fixed direction
> > in the galaxy.
>
> Experimenter's bias. Miller could not possibly avoid
> imposing his personal opinions, hopes, and dreams into his
> data. Such experiments REQUIRE data taking in a way that
> the experimenter cannot do that, or they are useless.

But experimenter's bias is excluded by design:

   An assistant records these readings in order, on a
   prepared form, starting with the reading corresponding
   to the north or other noted azimuth [...] The observer
   gives no attention to the azimuth. The reading is
   determined by instantaneous visual estimation; it is
   quite impracticable to use any kind of a scale in the
   field of view because the width of the fringes is
   subject to slight variation.
   [...]
   While readings are being taken, neither the observer
   nor the recorder can form the slightest opinion as to
   whether any periodicity is present, much less as to
   the amount or direction of any periodic effect; the
   taking of observations is quite unprejudiced and is
   simply mechanical.

The conjecture that the data from several seasons points to
a single galactical direction due to experimenter's bias
requires extraordinary evidence. The observer did not have a
harmonic analysier and an astronomical computer in his mind
instantenuously to adjust his observations at the time of
making them! I expected you would rather plead an error in
their calculations or, at least, a loose degree of
correspondence between seasons. I have not estimated it.

-- 
()  ascii ribbon campaign -- against html e-mail
/\  www.asciiribbon.org   -- against proprietary attachments

Back to sci.stat.math | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

statistics in Roberts. Was: RAW vs. raw image format Rich Ulrich <rich.ulrich@comcast.net> - 2023-02-27 16:00 -0500
  Re: statistics in Roberts. Was: RAW vs. raw image format davidd02@tpg.com.au (David Duffy) - 2023-02-28 07:47 +0000
    Re: statistics in Roberts. Was: RAW vs. raw image format davidd02@tpg.com.au (David Duffy) - 2023-02-28 09:12 +0000
      Re: statistics in Roberts. Was: RAW vs. raw image format davidd02@tpg.com.au (David Duffy) - 2023-03-13 01:23 +0000
        Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-13 10:32 +0300
        Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-13 10:58 +0300
          Re: statistics in Roberts. Was: RAW vs. raw image format davidd02@tpg.com.au (David Duffy) - 2023-03-14 04:13 +0000
  Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-02-28 12:42 +0000
    Re: statistics in Roberts. Was: RAW vs. raw image format nospam@de-ster.demon.nl (J. J. Lodder) - 2023-02-28 14:22 +0100
      Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-02-28 14:01 +0000
        Re: statistics in Roberts. Was: RAW vs. raw image format nospam@de-ster.demon.nl (J. J. Lodder) - 2023-02-28 15:12 +0100
          Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-02-28 14:57 +0000
            Re: statistics in Roberts. Was: RAW vs. raw image format nospam@de-ster.demon.nl (J. J. Lodder) - 2023-02-28 21:36 +0100
    Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@gmail.moc> - 2023-03-04 00:46 +0300
      Re: statistics in Roberts. Was: RAW vs. raw image format nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-03 23:40 +0100
    Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-03-09 17:07 +0000
    Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@gmail.moc> - 2023-03-19 00:41 +0300
      Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-03-19 00:08 +0000
        Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@gmail.moc> - 2023-03-20 01:08 +0300
          Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-03-19 23:14 +0000
            Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-03-19 23:48 +0000
            Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-20 11:58 +0300
              Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawkxx@nowherel.com> - 2023-03-20 17:23 +0000
                Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@gmail.moc> - 2023-03-24 00:11 +0300
                Re: statistics in Roberts. Was: RAW vs. raw image format "David Jones" <dajhawk18xx@@nowhere.com> - 2023-03-23 23:59 +0000
      Re: statistics in Roberts. Was: RAW vs. raw image format Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-19 14:22 -0500
        Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-20 14:07 +0300
          Re: statistics in Roberts. Was: RAW vs. raw image format nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-20 14:35 +0100
            Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-20 17:01 +0300
          Re: statistics in Roberts. Was: RAW vs. raw image format Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-20 16:20 -0500
        Re: statistics in Roberts. Was: RAW vs. raw image format nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-20 14:35 +0100
          Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-20 17:33 +0300
  Re: statistics in Roberts. Was: RAW vs. raw image format Anton Shepelev <anton.txt@gmail.moc> - 2023-03-03 23:33 +0300
  Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-05 12:48 -0600
    Re: statistics in Roberts' paper on Miller nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-05 21:47 +0100
      Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-08 21:26 -0600
    Re: statistics in Roberts' paper on Miller Rich Ulrich <rich.ulrich@comcast.net> - 2023-03-06 23:08 -0500
      Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-08 21:40 -0600
    Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-08 15:33 +0300
      Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-08 23:11 -0600
        Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-10 01:04 +0300
          Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-11 12:54 -0600
            Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-12 14:25 +0300
              Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-17 13:27 -0500
                Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-19 00:32 +0300
                Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-19 14:45 -0500
            Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-20 16:06 -0500
              Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-21 13:00 +0300
    Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-08 19:11 +0300
      Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-09 11:48 +0300
        Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-09 11:57 +0300
    Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-08 21:09 -0600
    Re: statistics in Roberts' paper on Miller "David Jones" <dajhawkxx@nowherel.com> - 2023-03-09 17:13 +0000
      Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-09 23:26 +0300
        Re: statistics in Roberts' paper on Miller "David Jones" <dajhawkxx@nowherel.com> - 2023-03-10 01:37 +0000
          Re: statistics in Roberts' paper on Miller nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-10 11:25 +0100
          Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-11 00:13 +0300
          Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-10 15:42 -0600
            Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-11 01:32 +0300
              Re: statistics in Roberts' paper on Miller - Data Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-11 12:18 -0600
                Re: statistics in Roberts' paper on Miller - Data Anton Shepelev <anton.txt@gmail.moc> - 2023-03-12 00:13 +0300
                Re: statistics in Roberts' paper on Miller - Data Anton Shepelev <anton.txt@gmail.moc> - 2023-03-20 00:38 +0300
        Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-10 15:25 -0600
          Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-12 00:01 +0300
      Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-10 15:10 -0600
    Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-11 01:14 +0300
      Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-11 13:08 -0600
        Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@gmail.moc> - 2023-03-12 01:51 +0300
    Re: statistics in Roberts' paper on Miller "David Jones" <dajhawkxx@nowherel.com> - 2023-03-15 17:10 +0000
      Re: statistics in Roberts' paper on Miller nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-15 22:36 +0100
        Re: statistics in Roberts' paper on Miller "David Jones" <dajhawk18xx@@nowhere.com> - 2023-03-15 22:18 +0000
          Re: statistics in Roberts' paper on Miller nospam@de-ster.demon.nl (J. J. Lodder) - 2023-03-15 23:28 +0100
      Re: statistics in Roberts' paper on Miller Tom Roberts <tjoberts137@sbcglobal.net> - 2023-03-19 13:46 -0500
        Re: statistics in Roberts' paper on Miller "David Jones" <dajhawkxx@nowherel.com> - 2023-03-19 22:57 +0000
          Re: statistics in Roberts' paper on Miller Anton Shepelev <anton.txt@g{oogle}mail.com> - 2023-03-20 12:05 +0300

csiph-web