The historical dynamics of 'best fit' EFE

On Mar 5, 3:13 pm, wrote:
In article ,
John (Liberty) Bell wrote:

I am referring here to the Closed Dark Energy model. I first came
across this as a passing reference in one of the accelerating
expansion reference papers we were discussing earlier (perhaps even
the one that you drew attention to. Charles is correct, however, in
the sense that Ned Wright picks up on this possibility and also quotes
cosmological parameters for this optimised solution (with Omega M =
0.55, Omega Lambda = 1.15).

Ned also concludes that this is the best fit, in his arXiv paper of ~
22 January. However, calculation confirms this does predict a
significantly brighter CMB than all other models, including the
classical big bang model (with Omega M = 1 and Omega Lambda = 0)

You're referring to astro-ph/0701584, right? I can't find any such
conclusion in this paper. On the contrary, the paper lists best-fit
models under a variety of different assumptions, and they're all
extremely close to flat. See Table 4, Figure 6, and the discussion
sections in particular, and note the last sentence of the abstract: "A
flat Lambda CDM model is consistent with all the data." In short,
this paper certainly doesn't provide support for your original
statement,

| 5) [recent] evidence of dynamics of accelerating expansion published
| thus requiring (for best fit [closed dark energy model]) that the
| universe is not flat after all.

Quite the contrary. It says that the "concordance" model of a flat
Universe with dark energy is still the best fit.

Yes, you are right. This is even confirmed in the abstract. I think I
may have confused the acronym CDEM where the C stands for
'concordance' with the acronym CDEM where the C stands for 'closed',
given http://www.astro.ucla.edu/~wright/sne_cosmology.html, where Ned
quotes chi^2 values for the original gold + silver Sn1a data set. Here
the best fit IS a closed model, having a chi^2/n improvement
commensurate to what Charles found when testing "Chalky's Law" on that
same data set.

I guess this is why Ned subsequently introduced the GRB data as well,
since this then seems to rule out that closed model over the range
2 z 7

John

On Mar 5, 11:59 am, Oh No wrote:

While I find it relevant to test the standard concordance model with
Omega_k=0, because that is the value (or close to it) indicated by WMAP,
I don't think it is relevant to allow Omega_k to float.

Actually, I would also appreciate a bit more background information on
the pertinence of WMAP, and on how this leads to the conclusion that
Omega_k=0. Somehow I suspect this isn't as simple as saying that we
can accurately predict the CMB energy density at the surface of last
scattering, from theoretical physics, and thus derive the best fit
model of EFE, from the CMB intensity we subsequently measure here.

Or is it that simple?

John


[Mod. note: no, it isn't. -- mjh]

In article , Oh No
writes:

I don't think I would find it interesting. It is of course the case
that, with any given body of data, the more fitting parameters you
introduce, the better the fit you are likely to be able to obtain (you
also make the labour of doing the fit disproportionately greater).

Andrew Little (+co-author(s)?) wrote a paper several years ago which
discussed taking this into account. That is, more parameters will in
general lead to a better fit, as you point out. The question is, when
are such parameters "demanded" by the data. This paper tried to
quantify that. (Ideally, the number of parameters would itself be a
free parameter, and the fit done in a higher-dimensional parameter
space, so to speak.)

In article , "John
(Liberty) Bell" writes:

While I find it relevant to test the standard concordance model with
Omega_k=0, because that is the value (or close to it) indicated by WMAP,
I don't think it is relevant to allow Omega_k to float.

Actually, I would also appreciate a bit more background information on
the pertinence of WMAP, and on how this leads to the conclusion that
Omega_k=0. Somehow I suspect this isn't as simple as saying that we
can accurately predict the CMB energy density at the surface of last
scattering, from theoretical physics, and thus derive the best fit
model of EFE, from the CMB intensity we subsequently measure here.

Or is it that simple?

[Mod. note: no, it isn't. -- mjh]

It is in practice quite complicated. On the other hand, Omega_k (i.e.
the sum of Omega and lambda) is quite strongly measured by the CMB data,
whereas other (combinations of) parameters are not.

On Mar 6, 1:39 pm, (Phillip Helbig---
remove CLOTHES to reply) wrote:
In article , "John

(Liberty) Bell" writes:


Actually, I would also appreciate a bit more background information on
the pertinence of WMAP, and on how this leads to the conclusion that
Omega_k=0. Somehow I suspect this isn't as simple as saying that we
can accurately predict the CMB energy density at the surface of last
scattering, from theoretical physics, and thus derive the best fit
model of EFE, from the CMB intensity we subsequently measure here.

Or is it that simple?
[Mod. note: no, it isn't. -- mjh]

It is in practice quite complicated. On the other hand, Omega_k (i.e.
the sum of Omega and lambda) is quite strongly measured by the CMB data,
whereas other (combinations of) parameters are not.

So..Can anyone elucidate on the how and why, or provide refs? I have
tried (superficially) searching for this severaal times, and got
nowhere.

John

the sum of Omega and lambda) is quite strongly measured by the CMB data,
whereas other (combinations of) parameters are not.

So..Can anyone elucidate on the how and why, or provide refs? I have
tried (superficially) searching for this severaal times, and got
nowhere.

Go to the ADS abstract site:

http://adsabs.harvard.edu/abstract_service.html

Type into the "Abstract words" box

cosmic microwave background geometry flat wmap overview

Click on the "Send Query" button. You'll receive a list of
over 100 papers which have some or all of these keywords
in their abstracts. Scan the list, pick out promising titles.
Click on each one -- for most, you'll see the abstract text.
If _that_ looks promising, read the paper. Either the
official journal text will be available, or, if it isn't, the
preprint text should almost always be available from
the astro-ph site.

There is no royal road to cosmology.

On Mar 5, 11:59 am, Oh No wrote:
Thus spake "John (Liberty) Bell"

I don't actually believe this CDE model myself, but do accept that it
gives a better fit to the Sn1a data, than do the WMAP and best fit
flat models (This model also gives the closest curve to Chalky's Law,
over the Sn1a data range).

It might be interesting if Charles also plugged this model into his
chi^2 calculator, for additional comparison purposes. However, as with
his earlier comparison with Chalky's law, I suspect that all
statistitically significant differences will be erased, as he reduces
the total size of the set.

I don't think I would find it interesting. It is of course the case
that, with any given body of data, the more fitting parameters you
introduce, the better the fit you are likely to be able to obtain (you
also make the labour of doing the fit disproportionately greater). I
strongly suspect that this is what is happening here. It doesn't
necessarily mean that the extra fitting parameter has physical meaning,
just that it gives a better fit for essentially random fluctuations
within the data. In practice, whatever way you cut the data, the sample
size is too small, the error margins too large, and the fits for the
existing laws too good, to expect that one can get a meaningfully better
fit by adding an extra parameter.

In the context of your above comment, it is, perhaps, worth repeating
that Chalky's Law has NO adjustable parameters. Consequently, if what
you say above is true, you SHOULD approve of the results of testing
the available data within the context of this law. After deleting some
of the observational data, you still found that Chalky's law is as
good as/better than the best fit flat EFE (to the limits of
statistical significance). Consequently you should, if what you say
above is true, still give Chalky's Law extra 'brownie points' since
that concordance was achieved with NO tweaking of cosmological
parameters whatsoever.

John

Thus spake "John (Liberty) Bell"
On Mar 5, 11:59 am, Oh No wrote:
Thus spake "John (Liberty) Bell"

I don't actually believe this CDE model myself, but do accept that it
gives a better fit to the Sn1a data, than do the WMAP and best fit
flat models (This model also gives the closest curve to Chalky's Law,
over the Sn1a data range).

It might be interesting if Charles also plugged this model into his
chi^2 calculator, for additional comparison purposes. However, as with
his earlier comparison with Chalky's law, I suspect that all
statistitically significant differences will be erased, as he reduces
the total size of the set.

I don't think I would find it interesting. It is of course the case
that, with any given body of data, the more fitting parameters you
introduce, the better the fit you are likely to be able to obtain (you
also make the labour of doing the fit disproportionately greater). I
strongly suspect that this is what is happening here. It doesn't
necessarily mean that the extra fitting parameter has physical meaning,
just that it gives a better fit for essentially random fluctuations
within the data. In practice, whatever way you cut the data, the sample
size is too small, the error margins too large, and the fits for the
existing laws too good, to expect that one can get a meaningfully better
fit by adding an extra parameter.

In the context of your above comment, it is, perhaps, worth repeating
that Chalky's Law has NO adjustable parameters. Consequently, if what
you say above is true, you SHOULD approve of the results of testing
the available data within the context of this law. After deleting some
of the observational data, you still found that Chalky's law is as
good as/better than the best fit flat EFE (to the limits of
statistical significance). Consequently you should, if what you say
above is true, still give Chalky's Law extra 'brownie points' since
that concordance was achieved with NO tweaking of cosmological
parameters whatsoever.

Oh indeed. It probably got lost in the quantity of reports, but I did
make comment to the effect that Chalky's law deserves "brownie points"
in this regard.


Regards

--
Charles Francis
moderator sci.physics.foundations.
substitute charles for NotI to email