Good News for Big Bang theory

In article ,
Oh No writes:
Most of the low redshift SNe are old discoveries, from a time when they
were mostly found by amateurs with a good memory for galaxies. I don't
immediately have information on where their positions are, but it is
highly unlikely that they are closely correlated.

Make that "highly likely." You are not the first physicist to
overlook practical aspects of astronomy.

Such a phenomenon would so radically contradict fundamental
assumptions

What phenomenon? Seasons? Earth having larger population in
northern hemisphere? Galactic extinction? Non-uniformity of
surveys? Large scale structure in the local Universe? Velocities
from large scale structure?

[Ned Wright] does not see the need for the
removal of SN at less than z0.023, which Riess did for precisely the
sort of reason you have stated. Wood-Vasey removed SN at z0.015.

So all the experts either remove or de-weight the very nearby SNe.
Why do you suppose they do that?

By the way, you aren't using uncorrected heliocentric velocities, are
you? Astier et al. (2006 A&A 447, 31, Table 8) corrects the
velocities to the CMB frame. Have you done that with the other data?
(It won't matter at high redshifts, of course.)

...the inclusion of systematic errors is likely to radically
lower the value of chi^/dof one obtains.

Indeed. As you say, treating systematic errors statistically is
dubious. I appreciate your efforts to combine the data sets and
derive the best conclusions possible, but I don't think a purely
statistical approach has any hope of success. Instead one has to
understand exactly where the systematic differences come from.

Despite that, given your results so far, it seems unlikely that the
existing SN data will distinguish beteen the models.

I believe the indications in the data are that we actually have three
distributions here,

That would be quite an interesting result, if correct.

I'm done with this thread, but if you try to publish your results in
astronomy journals, I expect the referees will be curious about the
same things I am.

--
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA
(Please email your reply if you want to be sure I see it; include a
valid Reply-To address to receive an acknowledgement. Commercial
email may be sent to your ISP.)

Thus spake Steve Willner
In article ,
Oh No writes:
Most of the low redshift SNe are old discoveries, from a time when they
were mostly found by amateurs with a good memory for galaxies. I don't
immediately have information on where their positions are, but it is
highly unlikely that they are closely correlated.

Make that "highly likely." You are not the first physicist to
overlook practical aspects of astronomy.

I don't think I expressed myself properly. It is highly likely that they
are correlated. For example I believe a disproportionately large number
of SN have been found by a single observer in Australia, who is
obviously only observing the night sky in that part of the world. What I
meant was that they are still not likely to be in extremely tight
groups, or even two in the same galaxy if I remember the frequency with
which they occur even vaguely correctly.

[Ned Wright] does not see the need for the
removal of SN at less than z0.023, which Riess did for precisely the
sort of reason you have stated. Wood-Vasey removed SN at z0.015.

So all the experts either remove or de-weight the very nearby SNe.
Why do you suppose they do that?

By the way, you aren't using uncorrected heliocentric velocities, are
you? Astier et al. (2006 A&A 447, 31, Table 8) corrects the
velocities to the CMB frame. Have you done that with the other data?
(It won't matter at high redshifts, of course.)

I am sure Riess does that, but I will double check.

...the inclusion of systematic errors is likely to radically
lower the value of chi^/dof one obtains.

Indeed. As you say, treating systematic errors statistically is
dubious. I appreciate your efforts to combine the data sets and
derive the best conclusions possible, but I don't think a purely
statistical approach has any hope of success. Instead one has to
understand exactly where the systematic differences come from.

I am sure you are right. But of course this means that I would have to
go right back to raw data and not use the prepared tables. It is a much
bigger job. I would have a lot to learn.

Despite that, given your results so far, it seems unlikely that the
existing SN data will distinguish beteen the models.

That is probably true.

I believe the indications in the data are that we actually have three
distributions here,

That would be quite an interesting result, if correct.

I think so. But again attempting to prove it is a much bigger job. If I
could prove it, then the chance of distinguishing between the models
might be quite good - particularly as it looks like the result of doing
this could give a rather unlikely value of Omega and correspondingly low
age for the standard model. Instead of feeling that I can only trust a
dataset of about 125 SN, I would have three datasets with a combined
total of about 300SN, and with much smaller chi^2 values.

The trouble is of course that this kind of analysis with mixed
distributions is generally very difficult and only rarely is it possible
to prove anything. Without clear criteria for which distribution each SN
belongs to, one ends up assigning them by hand. Just possibly, if I went
back to raw data and could separate the statistical and systematic
errors, I may be able to get an adequate criterion, but it is by no
means certain.

I'm done with this thread, but if you try to publish your results in
astronomy journals, I expect the referees will be curious about the
same things I am.

Whether I take this line of research further depends on what time I have
after completing other tasks. For now, all I am intending to attempt to
publish is a statement regarding consistency between theory and data,
together with remarks on what kind of data would be necessary to
distinguish, essentially deferring the prospect of a result until after
SNAP.



Regards

--
Charles Francis
substitute charles for NotI to email