Compelling evidence for in vacuo light speed anisotrophy
in [Relativity]
|
From: none on 14 Apr 2008 12:01 Surfer wrote: > On Sun, 13 Apr 2008 21:30:12 -0700, none <""doug\"@(none)"> wrote: > >> Surfer wrote: >>> On Sun, 13 Apr 2008 13:12:37 -0700 (PDT), Jerry >>> <Cephalobus_alienus(a)comcast.net> wrote: >>> >>>> The VERY PURPOSE of the experiment was to determine, "What is the >>>> form to be expected?" >>>> >>> That may have been the purpose of the original experimenters. But if >>> Cahill knew the form to be expected, I don't see why that would be >>> relevant to him. >>> >>>> Cahill's highly selective use of one single revolution of data, >>>> throwing out twenty-one revolutions that disagree with his >>>> prejudices, is an atrocity. >>>> >>> I think it depends to what use he put the data. If he just used it to >>> bolster his argument, that would seem harmless. >>> >> If you cherry pick the data that would seem to support you and ignore >> the data that does not, this is called fraud. >> > Cahill did not ignore any data. He presented all the data for readers > to view. He stated though, why he thought one rotation was useful and > the others were not. > The reason he saw it as useful is that it agreed with his bias. 95% of the data did not agree with him. That is bad science. >>> On the other hand, if Cahill selected data and didn't know the >>> difference between good data and bad data, he would not have been able >>> to calculate a useful value for 3-space velocity. >>> >> So "good" data is data that agrees with his prejudice... >> > No. Good data is data that is sooner or later found to be valid. Valid to Cahill means it agrees with his bias. > >> Miller had a signal consistent with 0 km/s >> > If you read Miller's original paper you would realise that you are > misinformed. > You had better look at developments in the last 70 plus years. > From Page 218: > "...A study of numerical results as plotted in Fig. 26 shows that the > probable error of the observed velocity, which has a magnitude of from > ten to eleven kilometers per second, is +/- 0.33 kilometer per second, > while the probable error in the determination of the azimuth is +/- > 2.5%. The probably error in the right ascensions and declinations of > the polar chart, Fig. 28, is +/- 0.5%" > > There is a copy of his paper here. > http://www.scieng.flinders.edu.au/cpes/people/cahill_r/Miller1933.pdf > > The precision of the above values may explain how Cahill's derived > speed of 415km/s, is so close to the values of 420-450km/s required to > correct the flyby anomalies. > > Miller's paper is wrong as Tom Roberts has well illustrated many times. This means all you can say is that the values are consistent with a range of speeds which includes zero. You seem to have no interest in the truth.
From: none on 14 Apr 2008 12:03 Surfer wrote: > On Sun, 13 Apr 2008 21:25:17 -0700, none <""doug\"@(none)"> wrote: > >> Surfer wrote: >>> On Sun, 13 Apr 2008 09:38:23 -0700 (PDT), Jerry >>> <Cephalobus_alienus(a)comcast.net> wrote: >>> >>>> On Apr 13, 9:56 am, Surfer <n...(a)spam.please.net> wrote: >>>>> On Sun, 13 Apr 2008 06:43:03 -0700 (PDT), Jerry >>>>> <Cephalobus_alie...(a)comcast.net> wrote: >>>>> >>>>>> [Cahill] thinks nothing of throwing away multiple experimental >>>>>> runs disagreeing with his prejudices, instead focusing on the single >>>>>> run that seems to show some sort of sinusoidal modulation (you do >>>>>> know the reference, I presume? >>>>> The situation is not so simple. He has been able to derive consistent >>>>> values for 3-space velocity from a number of experiments. >>>> I found the reference that I was referring to. >>>> http://redshift.vif.com/JournalFiles/V11NO1PDF/V11N1CA2.pdf >>>> >>>> In discussing the Joos experiment, Cahill writes: >>>> The data for 22 rotations throughout the day of May 30, 1930 >>>> are shown in Fig.15, and are reproduced from Fig.11 of [15]. >>>> From that data Joos concluded, using an analysis that did not >>>> take account of the special relativistic length contraction >>>> effect, that the fringe shifts corresponded to a speed of only >>>> 1.5 km/s. However as previously noted such an analysis is >>>> completely flawed. As well the data in Fig.15 shows that for >>>> all but one of the rotations the fringe shifts were poorly >>>> recorded. Only in the one rotation, at 11 23 58, does the data >>>> actually look like the form expected. This is probably not >>>> accidental as the maximum fringe shift was expected at that >>>> time, based on the Miller direction of absolute motion, and >>>> the sensitivity of the device was �1 thousandth of a fringe >>>> shift. In Fig.16 that one rotation data are compared with the >>>> form expected for Jena on May 30 using the Miller speed and >>>> direction together with the new refractive index effect,and >>>> using the refractive index of helium. The agreement is quite >>>> remarkable. So again, contrary the Joos paper and to >>>> subsequent commentators, Joos did in fact detect a very large >>>> velocity of absolute motion. >>>> >>>> In other words, out of 22 rotations performed that day, 21 of the >>>> rotations showed nothing remotely resembling a sinusoidal signal. >>>> Cahill throws away data from these 21 rotations and focuses on >>>> the single outlier, even ignoring data from the immediately >>>> preceding and following rotations that were separated from the >>>> 11:23 rotation by only a few minutes. >>>> >>>> This is biased data analysis at its most vicious. >>>> >>> You have a point. However, if police found 22 finger prints at a crime >>> scene of which 21 were smudged and one was clear, would you accuse >>> them of "biased data analysis at the most vicious" if they used the >>> clear one to identify a suspect? >>> >> This is a joke and does not apply. If you take the same measurement >> 22 times, it better gives a consistent result 22 times. >> > It won't if the signal is very weak. >> To pick one out of 22 is called fraud. >> > If he had concealed his rejection of 21 rotations that would be fraud. > But he didn't. He simply gave his reasons why he thought they were > invalid. Readers can make up their own minds whether to agree or not. >>> In some cases I think its valid to distinguish between data of "the >>> form expected" and data not of "the form expected". >>> >> This is true if you are trying to cheat so that your conclustion >> is supported when the data does not support it. >> > It is also true when some data is more reliable than others. > >> The fact that 21 of 22 disagree with the theory gives pretty good >> evidence the theory is wrong. >> > The theory suggests that the phenomenon fluctuates and is very > difficult to detect. 21 of the 22 support that aspect of the theory. > > So the fact that 95% of the data disagree with the model is somehow SUPPORT for that model??? So would it be even better if 100% of the data disagreed?
From: none on 14 Apr 2008 12:06 Surfer wrote: > On Mon, 14 Apr 2008 07:46:06 -0500, Tom Roberts > <tjroberts137(a)sbcglobal.net> wrote: > >> Surfer wrote: >>> [ahill] has been able to derive consistent >>> values for 3-space velocity from a number of experiments. >>> >>> Michelson A.A. and Morley E.W. Am. J. Sc. 34, 333-345, >>> 1887. >> But the variations in their data are not significant, so any "3-space >> velocity" derived from them is likewise not significantly different from >> zero. >> > Nevertheless, MM were able to calculate a value which was later noted > by Miller to be consistent with his own results. So I take this > accusation of "not significantly different from zero" with more than a > grain of salt. If you ignore error sources in an experiment, you are able to get values of whatever you want. Miller is not a reference to base a new theory on. >>> Miller D.C. Rev. Mod. Phys., 5, 203-242, 1933. >> But the variations in his data are not significant, so any "3-space >> velocity" derived from them is likewise not significantly different from >> zero. >> > Miller had complete confidence in his results. And I have complete confidence that I am the smartest person alive. Does that make it true? > > Here is a quote from the above paper. > > From Page 218: > "...A study of numerical results as plotted in Fig. 26 shows that the > probable error of the observed velocity, which has a magnitude of from > ten to eleven kilometers per second, is +/- 0.33 kilometer per second, > while the probable error in the determination of the azimuth is +/- > 2.5%. The probably error in the right ascensions and declinations of > the polar chart, Fig. 28, is +/- 0.5%" > > [The above figures are remarkably precise which may explain how > Cahill's derived speed of 415km/s, is so close to the values of > 420-450km/s required to correct the flyby anomalies.] There is a big difference between precision and accuracy. You can always add digits. You cannot add information. > > Regarding MM, Miller wrote (Page 237): > "Attention is called to the fact that the results obtained here are > not opposed to the results originally announced by Michelson and > Morley in 1887; in reality they are consistent with and confirm the > earlier results" > > >
From: Surfer on 14 Apr 2008 13:09 On Mon, 14 Apr 2008 10:55:06 -0500, Tom Roberts <tjroberts137(a)sbcglobal.net> wrote: >Surfer wrote: >> On Sun, 13 Apr 2008 21:30:12 -0700, none <""doug\"@(none)"> wrote: >>> Miller had a signal consistent with 0 km/s >>> >> If you read Miller's original paper you would realise that you are >> misinformed. > >Miller himself reported a non-zero result. A modern analysis of his RAW >DATA shows he was mistaken. His original data are available, and a basic >computation of errorbars from them shows that his result is not >statistically significant (i.e. errorbars using his raw data and flawed >analysis). > You are entitled to that point of view but I disagree. Your errorbars were derived by assuming that a perfect signal would have a certain form so that the magnitude of deviations from that form could be taken as indicating the magnitude of measurement error. However, in the dynamical 3-space view, fluctuations in dynamical 3-space velocity can be expected to cause the signal to fluctuate even before it is measured. Therefore although it is possible to estimate an average form for the signal, deviations from that form are not necessarly indicative of measurement error. Hence your error analysis has resulted in error bars that are far too large for this type of signal. Miller who could directly investigate sources of measurement error in his equipment, estimated probable errors as follows. From Page 218 of Miller's paper: http://www.scieng.flinders.edu.au/cpes/people/cahill_r/Miller1933.pdf "...A study of numerical results as plotted in Fig. 26 shows that the probable error of the observed velocity, which has a magnitude of from ten to eleven kilometers per second, is +/- 0.33 kilometer per second, while the probable error in the determination of the azimuth is +/- 2.5%. The probably error in the right ascensions and declinations of the polar chart, Fig. 28, is +/- 0.5%" > >Moreover, there is a clear explanation for why he was fooled: >his analysis technique FORCED his systematic error (primarily >interferometer drift) to have a sinusoidal form with a period 1/2 turn >(just as any signal would have). He took enough data so that this >systematic drift can be modeled quantitatively, and after subtracting it >there is no "signal" left. The result of a MODERN analysis of Miller's >data is 0 with an errorbar of 6 km/s: > http://www.arxiv.org/abs/physics/0608238 > >Regardless of Miller's errors, there is no reason TODAY to believe his >non-null result. This remains true in spite of Cahill's claims. > > >Tom Roberts
From: Surfer on 14 Apr 2008 13:33
On Mon, 14 Apr 2008 09:01:36 -0700, none <""doug\"@(none)"> wrote: > >You had better look at developments in the last 70 plus years. > >[Surfer wrote:] > >> From Page 218: >> "...A study of numerical results as plotted in Fig. 26 shows that the >> probable error of the observed velocity, which has a magnitude of from >> ten to eleven kilometers per second, is +/- 0.33 kilometer per second, >> while the probable error in the determination of the azimuth is +/- >> 2.5%. The probably error in the right ascensions and declinations of >> the polar chart, Fig. 28, is +/- 0.5%" >> >> There is a copy of his paper here. >> http://www.scieng.flinders.edu.au/cpes/people/cahill_r/Miller1933.pdf >> >> The precision of the above values may explain how Cahill's derived >> speed of 415km/s, is so close to the values of 420-450km/s required to >> correct the flyby anomalies. >> >> >Miller's paper is wrong as Tom Roberts has well illustrated many times. > Tom Roberts' paper http://www.arxiv.org/abs/physics/0608238 is interesting, but his errorbars were derived by assuming that a perfect signal would have a certain form so that the magnitude of deviations from that form could be taken as indicating the magnitude of measurement error. However, in the dynamical 3-space view, fluctuations in dynamical 3-space velocity can be expected to cause the signal to fluctuate even before it is measured. Therefore although it is possible to estimate an average form for the signal, deviations from that form are not necessarly indicative of measurement error. Hence his error analysis has resulted in error bars that are far too large for this type of signal. It is better to trust Miller, who could directly estimate sources of error in his equipment. > >This means all you can say is that the values are consistent with a >range of speeds which includes zero. > Only if you assume that the only possible 3-space is one that resembles a calm and stable ether. If you admit the possibility of a dynamical 3-space, then Tom Roberts' error analysis becomes inappropriate and Miller becomes credible. |