sci.astro
[Top] [All Lists]

Re: swift grb data rules out beamed theory

Subject: Re: swift grb data rules out beamed theory
From: "sean"
Date: 26 Mar 2006 10:32:37 -0800
Newsgroups: sci.astro
Craig Markwardt wrote:
> "sean" <jaymoseley@xxxxxxxxxxx> writes:
> ...
> > Note  from my quotes that I have stated that rapid
> > variations in the optical lightcurve should be
> > seen if the time scale of the sampling were short
> > enough. One of my theoretical criticisms of current
> > procedures and beamed theory was that the power law
> > fitting of optical data employed by the astrophysics
> > community was in fact a misguided procedure
> > that disguises the true variability of magnitude within
> > the optical lightcurves by incorrectly smoothing out
> > the observed variability in the the optical lightcurve,
> > particularly in early time lightcurves where
> > flucuationsare more easily seen due to the relative
> > brightness of grb afterglow.
Note that Craig seems to be in denial that power
law fitting takes variable brightness observations
and fits them to a straight line power law.
Especially considering that Craig himself has insisted
many times that the decay rate is best described by
a smooth power law decay with at most 1 break.
Here is Craigs quote from years ago trying desperately
 to show how my non `power law` approach is just not
acceptable
"..Also, as has been pointed out, trying to infer that
something is a "peak" when the data are as noisy
and as sparse as they are, is in my view a dubious
practice.  Simply "connecting the dots" will lead
to *a possible* solution, but ultimately a very
*low probability* one a priori.
A more appropriate approach would be to start with
a featureless model of the decline (say, a power law),
then add a gaussian or some other simple
parameterization of the putative peak.  By computing
the F-statistic, one can then find out how significant
the additional peak is, statistically speaking,.."

And here is Staneks reply that this method as endorsed
by Craig is basically, well unscientific as regards grb`s..

"....but with clear short-timescale variations, as
reported before by Morris et al. (2006b). Trying to
describe these erratic events with smoothpower-law
fits is often a dubious statistical proposition..."

Note how Stanek clearly criticises Craigs approach
and endorses my argument that the small time scale
variations in optical data has to be accepted
without ANY power law smoothing.>
> Note that fitting of a power law model in no way smooths out the data.
> The data is the data, and the model is the model; they are two
> separate things.  Thus, your "smoothing" criticism is irrelevant.
>
> It's also rather ironic that you are claiming that Stanek's work is a
> vindication of your "theory." Your "theory" involves the X-rays and
> optical light curves being time-delayed copies of the gamma-ray light
> curve.
Ironic? He confirmed my prediction regarding
power law. So yes its Ironic that a supporter
of beamed theory publishes findings supporting
my theory.
Note also that 060218 has exactly the time delayed
copies I predicted taht you have so often claimed as
being unsubstantiated and impossible theoretically.
Here in fact is a quote from one of your own colleagues
on Swift team
"...  Althoughthere are unusual aspects to this event
(cf. Gehrels GCN 4787)....We also draw attention to
the chromatic nature of the Swift light curves.The BAT
emission peaked substantially earlier than the XRT
emission,which preceded the UVOT emission peak... "

Now I know you arent that good with concepts Craig but
what he means above is that gamma peaked first , then X
then optical. Just as I always predicted (theres many
quotes if you need them) and contrary
to what you calimed was possible. In fact contary
to what your collegue admits to what is predicted
or explained by beamed theory.>
> On the other hand, while the optical afterglow of GRB 060206
> has multiple peaks, the gamma-ray light curve has a single smooth
> peak.  Not a "copy."
To start with you are ignoring your own gcn posts
from Swift gcn4697 ...".  The highest energy bin shows
a possible double peak structure, with the first peak
at T0-1 second..."  Didnt he say  `two peaks` there?

But the sampling in gamma is limited as its only a
10 sec or so duration burst in BAT. So my guess ,
is that if you were to compare similar shaped gamma
profiles from bursts where BAT is 50 seconds only
an ostriche would insist that there isnt similar
finer multipeaked structures in shorter BAT times.
Go look at some old BATSE curves where the graphs
have better sampling. There is always finer structure
detail in better sampled BAT lightcurves.
And anyways even looking at 060206 bins one can see
hints of further multiple fluctuations.>
> Or, while the gamma-ray light curve for GRB 060210 has multiple peaks
> spanning hundreds of seconds, the available optical light curve has
> only a single peak.  Again, not a copy.  Your theory is not really
> vindicated at all.
The `flat part` you refer to is  only about 1/15
the time of the available optical data.And only
the first 7000 seconds is available in Staneks
paper at least so its undersampled. And you
ignore the fact that elsewhere in the optical curve
there *are* other rebrightenings. So the overall
structure in optical is spikey not flat as you say.
In fact even though its still not well sampled Id
say over all even if one excludes the initial rise
in OT e there are at least 4 spikes recorded in
optical.
That is if you dont try to smooth them away with
power laws.>
> ...
> > I have supplied illustrated graph samples at my website
> > www.gammarayburst.com since 2000 of how if one takes into
> > account the observed variability in the lighthcurve data one
> > can extrapolate numerous, rapid and pronounced variabilty
> > in brightness of the grb afterglow as it decays.
> > As you can see, my critics not only said that my theory was
> > wrong and proof of this was that their evidence (power law
> > chi squared smoothed decays ) showed that there were no
> > fluctuations as predicted by their beamed theory.
> > They alsotried to suggest that my attempts to argue
> > that one shouldnt apply power law smoothing was unscientific.
> > As you can see if you look at Staneks quotes, in fact now
> > the only acceptable way to analyse grb optical decay is by
> > using the very methods I proposed back in 2001 and not
> > the misguided method of power laws as championed by my narrow
> > minded sceptics since 2001.
>
> Sorry, but your claims are not really convincing.  You would have to
> show that somehow the data sampling has suddenly gotten more rapid now
> than before, in order to resolve the fluctuations.  I don't think it's
> the case.
>
> In my opinion, what *is* different in the Swift era, is that observers
> are performing much more rapid follow-up observations, so they are
> seeing more early-time fluctuations that nobody was even trying to
> observe before.  Also, you are focussing on very high redshift bursts
> which have been intrinsically time-stretched at all wavelengths, by a
> factor of (1+z) ~ 5 or so.
Why do I have to show data sampling is or isnt
more rapid? Thats an irrelevent and confused point.

And secondly I dont just use Swift era data. How could I?
I made the prediction that it was already there before
Swift. Its only you guys who needed it spelt out in
big letters before you were willing to accept
they were the norm. And even now you seem to
be in denial that power law fitting is unsuitable.
Despite Staneks paper.
>
>
> > Here are samples of my quotes
> > below. In addition my quotes below implicitly
> > predict that X ray lightcurves will also exhibit dramatic
> > rebrightenings and at later times than expected under
> > beamed theory allows. This prediction is also
> > rubbished as impossible and unverifiable by one critic
> > below. Yet as we know now from several recent Swift
> > bursts (outlined in Staneks paper) that this is not
> > only possible but most likely the norm
> > My first  quotes below are from my 1st post from the
> > thread `Swift grb satelitte` started on nov16 2004
> > the url is below, unwrapped ...
> > http://groups-beta.google.com/groups?as_umsgid=40f80aa1.0411160844.8b2fe5@xxxxxxxxxxxxxxxxxx
> >
> > Sean Nov 16 2004...
> > " the shorter the time frame of the exposure of the CCD
> > the more detail will emerge..... more `peaks` will emerge
> > in shorter ccd exposure times for SWIFT. This will give
> > the appearance of more numerous rapid rebrightenings
> > than current *theory allows*.."
>
> See comment above about sampling now vs. then.
>
> > "...SWIFT will also see these rebrightenings always
> > occuring at later times in longer wavelengths. Ie/ a
> > rebrightening observed in UV will appear to peak
> > slightly later in optical.
> > If SWIFT observes a burst with enough detail in its
> > Gamma X UV OT filter bands it should be possible to
> > chart features that first occur in gamma then appearing
> > seconds later in X..."
>
> Note: no evidence presented in this post of any type of delays.  I am
> also aware of quite a number of GRBs that have been observed
> simultaneously by Swift in gamma-rays and X-rays, and there are no
> delays.  Also, there are obvious case like GRB 041219a which have no
> apparent delays between the gamma-rays and optical, as I've mentioned
> before.  [ I've seen a talk by Tom Vestrand which suggests the optical
> emission is a combination of prompt + delayed emission, and he
> presented the "Green's" function kernel which described that
> response. ]
Your first comment shows how as usual you like
to fabricate false information, despite knowing that
what you say is incorrect. Case in point is my
repeated reference to you in particular and your
acknowledgement that youve referred to the graphs that
I have at my website. For 6 years Ive posted
a graph using NASA data that shows the delays in
time between the different bands in 970508.

Which  grbs do you refer to vis a vis no delays
between x and bat? Wheres your eveidence that there
werent delays on the order of 1/1000th sec?
Ill like look at the data first. But I imagine they
decayed fast in all bands and the delay
is too small to measure in microseconds accurately.
As I have clearly pointed out before.
Remember I say that grbs can be very short with
very little delays and very long with long delays
like 060218.
Regarding 041219 . Prompt emmision isnt neccesarily
a given. Take gcn 2876 and 2870 . Isnt that k=15.5
(2 mins post burst) k=14.9 (0.8 hrs post burst)
and k=15.5 (1.5 hours post burst)? Unless I missed
something that looks like it increased in K between
2 mins and 0.8 hours post burst?
Also all the J band observations ( 2872 and 2865 )
seem to still be at earliest 2.8 minutes post burst
Only the raptor R  observations which start
during the burst show a faint OT at R=19 and
nothing brighter than that at later observations
But R as my model predicts would have peaked
before K and after gamma which puts that peak
in R somewhere between end of burst and 0.8 hours.
And there are no R band observations from end of
burst to 0.8 hours.
Anyways as I say K shows a peak after the gamma
afterglow faded so thats not prompt emmision.
You are wrong again.
>
>
> > ".. This will be a
> > progression directly proportional to wavelength so that
> > if it takes 10 seconds for the `spike` to move from
> > 1nm to 10nm then it will take 100 seconds to move
> > from 10nm to 100nm... "
>
> But I'm also aware that you made two different claims, delay
> "proportional to wavelength", and delay "proportional to the
> difference in wavelength," but you could never understand that these
> are quite incompatible predictions.
>Oh your not on about this again? You dont
even know the difference between..."delay
proportional to wavelength", and delay
"proportional to the difference in wavelength,"
yourself .
Its a non argument and an irrelevent point.
For instance tell me. If you think that the
two statements above are different then heres a
test of your intelligence. Which of your two
above different statements  explain this example..

If X is at 1nm Y is at 10 nm and z is at 100nm
X peaks at 1 sec Y peaks at 10 sec and Z peaks
at 100 sec.
Which of your above 2 terms  describes this scenario?
Bet you dont know.
>
> > Craig Markwardts post of mar 28 2005 of the same
> > thread as my quotes...
> > "...Scientifically speaking, if you were claiming a
> > "good fit," then you would be giving proportionality
> > parameters, confidence limits on those parameters,
> > and goodness-of-fit measures (i.e. chi-square statistics).
> > I note that you continue to ignore such formal measures
> > and instead speculate wildly...."
>
> My point still stands.  If you are claiming a "good" or "better" fit,
> then you should be able to back that claim up with formal and
> quantitative statistical measures.  Otherwise you are just propounding
> subjective mush.
(Ooh nasty boy Craig missed his potty again.)
You are confused here. The fact is Im
claiming one *shouldnt* be using fitting
at all. And Stanek agrees with me.
>
>
> > .. and here Craig tries to push the theoretical approach
> > of power law smoothing as being correct over my prediction
> > that in fact there is significant small scale variabilty in
> > optical lightcurves that has been erased by `smoothing`...
> >
> > (Craig Markwardt april 11 2005)"..
> > And the point is that
> > right now you are using a trial-and-error approach which:
> > (a) ignores many possible combinations of valid parameters
> > which may not be so flattering to you; (b) ignores
> > uncertainties on measurements; and (c) ignores possible
> > systematic biases. "
>
> Note: no mention of "power law smoothing," so your claim is
> irrelevant.  But my points still stand: you can't ignore systematic
> biases in your analysis approach, which you had (and have) been doing.
I sort of thought you`d try to pretend that
all your waffling on about chi squared and
parameters wasnt anything to do with fitting
or power laws once you found out Id proved
you wrong.

So heres a quote of yours...
"...Simply "connecting the dots" will lead
to *a possible* solution, but ultimately a very
*low probability* one a priori.
A more appropriate approach would be to start
with a featureless model of the decline
(say, a power law),.."

Its odd that . You know , Im sure you just
used the words power law there.>
>
> > Craig Markwardt feb 22 2005 sci.astro..
> > " In fact, if you
> > had read some of the literature, you would have found
> > measured hard-to-soft gamma-ray lags to be significantly
> > less than 0.1 seconds.  (Norris 2000).  But it appears that
> > you have not even tried to consult the literature.  When you
> > get your wavelength values right, you will see that you
> > would predict significant lags down to 0.5 keV or to
> > optical... "
> >
> > As you can see above, Craig stands by his claim that
> > my predictions of time lags between hard and soft high
> > energy observations of  than 0.1 seconds are incorrect
> > and not possible. As seperate 060218 data shows he is
> > completely wrong and the data verifies my predictions
> > that some grbs will have significantly longer delays
> > between all wavelengths (including in X)than believed
> > possible by current beamed theory.
>
> I note that you don't have a response to the Norris (2000) paper,
> which is a systematic study of many GRBs.  Meanwhile you chose to
> cherry-pick one GRB to showcase your "theory."
Ill read his paper and get back to you on that.
But regarding cherry picking Its better for me to
have predicted and explained it then that of
your theory which not only doesnt explain it but
it doesnt  predict it either.
>
> Let's look at that one, GRB 060218, a little more closely.  It's
> certainly true that the burst had delayed X-ray and optical/IR
> emission.  However, one case does not *prove* a theory.  In fact, that
> event appears to have all the signatures of a supernova explosion,
> including an optical supernova spectrum.  Thus, the most likely
> interpretation is that GRB 060218 was in fact a supernova.  Does your
> "theory" predict the spectrum of a supernova?  I think not.
>You think its an SN and try to get data to show
that, but as the gcn for thatburst show its a
questionable SN-grb as Ive pointed out to George
due to dispute over it being an off axis burst.
And where are these spectra? How many were made
and how many fit exactly which other SN sample?
You`ll have to show some spectral evidence to back
up your spurious claim because that burst has
elements which cannot be explained by grb or SN
Anyways you pretend that the only verified proof
I have is 060218.
What a joke. You havent read a word Ive said ever.
Every single GRB on record fits my theory.
There isnt one you can name that shows clear
contrary evidence to any of my predictions.
If you think they do, cite that evidence here.
Dont make things up.
> > Below are a  few more quotes from M Hardcastle, J Lazio
> > and C Markwardt from an even earlier thread ...
> > ...`beamed gamma ray bursts`Sat, Oct 27 2001 9:09 am...
> > They try here to argue that my criticisms of using power
> > law smoothing is unscientific and unverifiable
>
> ...
> > (And finally Craig Markwardt replies with an
> > argument that we now now to be misguided at best
> > below about how the only way to analyse grb
> > optical lightcurve decays is to dogmatically
> > adhere to chi squaring and power law smoothing
> > of the data to make sure it fits a decidedly
> > useless beamed theory.
> > Below Markwardt insinuates my approach of including
> > the fluctuations in the observed lightcurve is dubious!
> > Do you still think so now Craig? )
>
> Let's see what I said:
>
> > C Markwardt...
> > Also, as has been pointed out, trying to infer that
> > something is a "peak" when the data are as noisy
> > and as sparse as they are, is in my view a dubious
> > practice.  Simply "connecting the dots" will lead
> > to *a possible* solution, but ultimately a very
> > *low probability* one a priori.
> > A more appropriate approach would be to start with
> > a featureless model of the decline (say, a power law),
> > then add a gaussian or some other simple
> > parameterization of the putative peak.  By computing
> > the F-statistic, one can then find out how significant
> > the additional peak is, statistically speaking,
> > compared to the overall decline.  If it's significant
> > at the >95% level, then it may be worth considering
> > further..." !
> > (yeah right Craig..maybe you should take a look at
> > Staneks paper.)`
>
> I did, and while it's a great paper, it's not really relevant to the
> discussion.  The light curve we were discussing at the time was noisy
> and sparsely sampled.  Mild fluctuations have long been known in
> afterglow light curves, and the Swift era we are certainly sampling
> some larger amplitude fluctuations (see above for some reasons).
Yes. Thats great science there Craig. Really
laughable this quote... "not really relevant to the
discussion"
I *only* refer to its findings in my post.
I posted only to cite Staneks findings.
This thread here was started only to
highlite how his paper confirms my prediction
on real fluctuations in the lightcurves.

>
> Coming back to your claims from above:
>
> > ... about how the only way to analyse grb
> > optical lightcurve decays is to dogmatically
> > adhere to chi squaring and power law smoothing
> > of the data to make sure it fits a decidedly
> > useless beamed theory.
>
> Did I say the "only way" to do the analysis was to fit a power law?
> No.  I suggested an appropriate approach, given the presence of a
> noisy light curve.  Did I say anything about "smoothing"? No.  Did I
> say to use a "beamed theory?"  No.  So far you have completely
> misinterpretted what I said, and thus your criticisms are irrelevant.
>
So I take it now you do agree with me and Stanek
that power law fitting using chi squared
parameters etc is incorrect in grb lightcurves?

> > ... Markwardt insinuates my approach of including
> > the fluctuations in the observed lightcurve is dubious!
> > Do you still think so now Craig? )
>
> Did I say that including fluctuations was dubious?  No of course not.
> What I *did* say was that trying to infer peaks from a noisy and
> sparse light curve by connecting the dots was a dubious practice.  In
> *fact*, I offered a possible approach to admitting fluctuations (by
> adding them to the model and seeing how well they improve the model in
> a quantitative way).  Not the only way, but "a" way, in the face of
> compliant data.

Rubbish again. Heres your quote...

C Markwardt...
"..Also, as has been pointed out, trying to infer that
something is a "peak" when the data are as noisy
and as sparse as they are, is in my view a dubious
practice.  Simply "connecting the dots" will lead
to *a possible* solution, but ultimately a very
*low probability* one a priori. ""

Didnt you just use the word dubious there?
Didnt you Craig??
Anyways your wrong . And Stanek has now confirmed
that my methods are the best scientific methods for
grb analyses.
And, he has supplied a good example of how GR can be
violated. Although he doesnt realize yet.

Sean


<Prev in Thread] Current Thread [Next in Thread>