Plasma Cosmology - Woo or not

ben m said:
If anything, then, this helps cosmology, by bringing the imprecise modern X closer into agreement with the precise CMB X.
Click to expand...

I see. You were just joking when you claimed that "all of those predictions are confirmed".

Because you weren't sticking the actual measured values of X and Y into your equations but something else entirely ... numbers that would give you the answer you want. ;)
 
BeAChooser said:
I see. You were just joking when you claimed that "all of those predictions are confirmed".

Because you weren't sticking the actual measured values of X and Y into your equations but something else entirely ... numbers that would give you the answer you want. ;)
Click to expand...

Considering you never presented any numbers, theory or evidence to support how the PC actually works in the hundreds of posts you made, this is the best response.

21 meter salute:
:id::id::id::id::id::id::id:
:id::id::id::id::id::id::id:
:id::id::id::id::id::id::id:
 
BeAChooser said:
I see. You were just joking when you claimed that "all of those predictions are confirmed".

Because you weren't sticking the actual measured values of X and Y into your equations but something else entirely ... numbers that would give you the answer you want. ;)
Click to expand...

Good heavens. Read my post again, the post you're quoting. Read it carefully. All of those words are important and they're there in a certain order.

We are talking about hypothesis testing. Do you know each of those words? Say them out loud. Here is how hypothesis testing works.

Pick any value of X.
Pick any value of Y.

(Are they the "real" values? Not yet, this is still hypothesis testing. Does it matter which values you pick? Not yet, we'll eventually scan through all values of X and Y and test them all.)

Have you picked a value of X and of Y? Assuming these values were true, plug them into your physics model. Use the laws of physics in your model to predict what the data would look like. (Read this carefully. Do you understand? For example, "a hypothetical universe with X = 10 and Y=0.00001, undergoing its own GR-obeying Big Bang, would have (insert physics calculations here) a 200K CMB with 10^-2 fluctuations at L=40 and L=80"

Now look at the real world. Do the data and the hypothesis-prediction look the same? (Read this carefully. Do you understand? For the numbers above, the real world has a 2.7K CMB with 10^-5 fluctuations at L=120.)

Calculate how similar they look and write down that number. Smaller numbers indicate good agreement. Bad agreement, like you see above, gives you a large number. This number is called the "test statistic". For the prediction above, maybe the test statistic we calculate is 10,000,000,000. So we write down:

X=10, Y=0.00001, test statistic = 10,000,000,000

So, ooh, that's not a very good model, is it? But never mind for now. Go back to "pick a random value" and keep scanning until you've tested all possible X and Y.

Do you understand? I really don't want you to misread this. Please read it two or three times, very carefully.

The LCDM model has the following property, exactly like I said. You make the above table, trying out every possible XYZT and and plugging them into the Big Bang model's GR initial conditions. Every time you generate a new XYZT set, you look at the CMB and see if it matches. Somewhere in that table is the following:

X=0.07 Y=0.23 Z=0.75 H=70, test statistic = 10 (poor match)
X=0.07 Y=0.25 Z=0.73 H=70, test statistic = 4
X=0.06 Y=0.24 Z=0.75 H=70, test statistic = 3
X=0.06 Y=0.25 Z=0.74 H=70, test statistic = 3
X=0.05 Y=0.24 Z=0.76 H=70, test statistic = 2 (mediocre match)
X=0.05 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)
X=0.04 Y=0.25 Z=0.76 H=70, test statistic = 2 (mediocre match)

There's one and only one value of XYZH for which the Big Bang predicts the properties of the CMB.

Let's try again on a totally different data set. Let's run Big Bang theories, try all possible XYZH, and make predictions for Large Scale Structure. Look at that table. Somewhere in there is this:

X=0.05 Y=0.25 Z=0.78 H=70, test statistic = 3 (mediocre match)
X=0.05 Y=0.24 Z=0.76 H=70, test statistic = 1.5 (fair match)
X=0.05 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)
X=0.04 Y=0.25 Z=0.76 H=70, test statistic = 3 (mediocre match)

There's ALSO a value of XYZH for which the Big Bang theory correctly predicts the properties of the LSS. Look---it's the same values!

Repeat that exercise for the Lyman-Alpha Forest, and in that table is:

X=0.05 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)

Repeat that exercise for the CMB polarization map, and in that table you find:

X=0.05 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)

And so on. Read my post again:

Listen carefully: if I tell you that the Universe today has X g/cc baryons, Y g/cc dark matter, and Z g/cc dark energy, and a Hubble constant H (all measurable), then BBT tells you that the CMB should be a blackbody at 2.73K, nearly-but-not-exactly isotropic, with perfectly Gaussian fluctuations with a peak-y angular power spectrum, and from the four numbers X,Y,Z,H you can predict exactly where all of the peaks are. Then it tells you that the CMB should be polarized, with the polarization all in the E-mode, and it tells you how to predict the TE and EE cross-power angular spectra.

All of those predictions are confirmed.
Click to expand...


Did you skip the "listen carefully" part?

Did I accidentally say "BBT tells you the ratio of red stars to cluster gas"? Did I say "BBT tells you what fraction of baryons are easily visible?" No I didn't, because it's not true. I said what I meant to say.
 
ben m said:
We are talking about hypothesis testing.
Click to expand...

Oh, I see. You were merely talking about "hypothesis testing", not the "certainty" you first suggested with your original statement. Because I sure got the impression that your *hypothesis* had been "confirmed" by the X and Y you had before this new information came to light. :p
 
BeAChooser said:
Oh, I see. You were merely talking about "hypothesis testing", not the "certainty" you first suggested with your original statement. Because I sure got the impression that your *hypothesis* had been "confirmed" by the X and Y you had before this new information came to light. :p
Click to expand...

You read that post awfully quickly. I wanted you to read it slowly, and carefully, and sleep on it and read it again.

Yes, I was talking about hypothesis testing. When you find that one hypothesis matches a large swath of data, this is called "confirming the hypothesis". The LCDM hypothesis is considered "confirmed" because it matches so many data points using these unique values X=0.05, Y=0.25, Z=0.75, H=0.7. That's the same thing as saying that we have confirmed that those are the real values. (With the usual caveats of all science.)

Let's make point #2 clear. The fact that the Universe is 5% baryons means (as far as we can tell) that it's 4% cluster gas and 1% stars. The values 0.4 and 0.1 are not part of the LCDM model. LCDM would be equally consistent with 5% stars, or 4% stars, or 3% stars, or 2% stars, or 1% stars. Your new piece of data seems to be telling us that the 5% baryon-fraction is slightly richer on the stars-side---so it's 1.2% stars and 4.8% cluster gas? Fine with me, and fine with LCDM.

You want to add star-fractions to the LCDM test? Fine, go ahead. Hypothesize that the stars are S and the cluster gas is C, so S+C=X. You've got a five-parameter model. Compare all possible lists of S,C,Y,Z,H to (say) the CMB data. You'll get a table that looks like this:

S=0.01 C=0.04 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)
S=0.02 C=0.03 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)
S=0.03 S=0.02 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)
S=0.04 S=0.01 Y=0.25 Z=0.75 H=70, test statistic = 1 (perfect match)
S=0.01 C=0.04 Y=0.26 Z=0.74 H=70, test statistic = 2
S=0.02 C=0.03 Y=0.26 Z=0.74 H=70, test statistic = 2
S=0.03 S=0.02 Y=0.26 Z=0.74 H=70, test statistic = 2
S=0.04 S=0.01 Y=0.26 Z=0.74 H=70, test statistic = 2
etc.

... and which tells you that (a) cosmology data doesn't care about star counts, and (b) star counts are not a good way of testing cosmology.
 
(Continuing the above pedagogical-model, let me comment on falsifiability. What would happen if the very best entry in our whole table was, for the CMB data,

X=0.07 Y=0.26 Z=0.72 H=70, test statistic = 50
X=0.05 Y=0.26 Z=0.74 H=70, test statistic = 20
X=0.03 Y=0.26 Z=0.76 H=70, test statistic = 45
(and all other entries were much worse)

and the very best entry in the whole table for LSS data was

X=0.50 Y=0.50 Z=0.00 H=70, test statistic = 100
(all other entries much worse)

and the very best entry in the whole table for CMB-pol data was

X=0.10 Y=0.1 Z=0.0 H=40, test statistic = 1000

That's what falsification looks like. The hypothesis-testing process could have falsified the LCDM hypothesis entirely. It could have been falsified if those test-statistics had been crappy. It didn't. It could have been falsified if different datasets disagreed on the best input parameters. It didn't. It kept spitting out these gorgeous agreements (statistic = 1) and it always put them at the same place in the XYZH table. How did that happy? Bad luck? Not fine-tuning---remember, we tried all possible XYZH combinations, we didn't cherry-pick the best line. The test statistic is a standard mathematical calculation, there's nothing to tune. Not merely by adding parameters---the CMB data, all by itself, tells us that "no dark energy" is a poor model. If you think that "lambda=0.75" is some nonsense that we made up to match CMB data, how did the LSS data happen fit perfectly to lambda=0.75? How did the SZ data also do it? How did the supernovae also do it? Remember, we tried comparing lambda=0 to the SNe data, and the SNe data told us "no, that doesn't work, but lambda=0.75 does." )
 
dafydd said:
Resorting to abuse is a sure sign that you have lost the argument.
Click to expand...


Assigning two practically equivalent words to someones forum nickname is not abuse :rolleyes:

Where has Mr Invictus gone with his reply to his little "challenge"? I'm not going to respond at a tangent to everyone's replies else I end up not being able to follow through on points raised, and this seems to greatly annoy DRD especially, which I would not dream of doing any further as we get along so well. Much rather stick to one argument at a time.
 
Zeuzzz said:
Assigning two practically equivalent words to someones forum nickname is not abuse :rolleyes:

Where has Mr Invictus gone with his reply to his little "challenge"? I'm not going to respond at a tangent to everyone's replies else I end up not being able to follow through on points raised, and this seems to greatly annoy DRD especially, which I would not dream of doing any further as we get along so well. Much rather stick to one argument at a time.
Click to expand...

So here you go Zeuzzz,

1. Provide a model.
2. Show how the data fit the model as well or better than the standard model.

This is where the PC/EU fails everytime, not talking about 'wouldn't it be great' but 'what can you actually show'.

Seriously.

The issue is that you have been promoting these ideas and never are able to provide 1 and 2 together.
 
Zeuzzz said:
Assigning two practically equivalent words to someones forum nickname is not abuse :rolleyes:

Where has Mr Invictus gone with his reply to his little "challenge"? I'm not going to respond at a tangent to everyone's replies else I end up not being able to follow through on points raised, and this seems to greatly annoy DRD especially, which I would not dream of doing any further as we get along so well. Much rather stick to one argument at a time.
Click to expand...

The ball's in your court, Z. You never replied to any of the posts below. Go back and read them (not just the quotes below, they're incomplete), reply, and we'll see where we are.

sol invictus said:
Very well, lithium abundance it is. There is indeed a tension in BBC with Li abundances, although to find out how serious it is I will need to do some reading. Probably some other forum members can help with that.

For the PC side, you've provided two papers by Eric Lerner. Can I assume that - in your view - the theory presented in those papers is (part of) PC, and that the results in those papers were derived correctly from the theory?
Click to expand...

sol invictus said:
You see Zeuzzz, that's why I'm waiting for your response to this:



But as usual ben m is more eloquent than me, so:
Click to expand...

sol invictus said:
No Zeuzzz, it doesn't work that way. You asserted that PC is more predictive than the BB. Now it's time to back that up. So you pick one prediction of PC, we'll take a look.

It has to be specific and quantitative and cosmological. It has to differ from the BB either in the prediction or in the mechanism, or both. And most importantly, you have to say - in advance, before we analyze it - that it is a core part of PC, that the success of PC depends significantly on it, and that if it were falsified, it would falsify or significantly weaken the case for PC as a theory of cosmology.



Make up your mind - pick something and take a stand. That's where you failed last time, and that's probably where you'll fail this time too. Science is about making falsifiable predictions, that's what distinguishes it from religion.
Click to expand...
 
Zeuzzz said:
Lithium abundance.
Click to expand...

a) What laws of physics does Lerner use to predict the lithium abundance?

b) How many free parameters are there in the initial-state he applies these laws to? ("A priori we didn't know the primordial densities of C,N, or O, nor the cosmic ray proton flux, nor cosmic ray proton spectral index" would be approximately five parameters.)

c) What data points (with error bars) are used to determine the values of the unknowns in (b)? What is the chi^2 (or other standard test statistic) showing the strength of the agreement between these data points and the model?

d) What other data are predicted by the same parameter set you isolated in (c)?

The statement---"PC predicts the lithium abundance"---is equivalent to the statement that the (c) data points outnumber the (b) unknowns, that the (c) chi^2 is acceptable, and the desired model is not otherwise ruled out by some data in (d). Are you going to convince us of that?
 
Its like an excited terrier with a bone.

If the theory is gonna get pwnd I'd rather it was done at the hands of an actual scientist, or maybe even the guy I actually asked here, thanks anyway, ben.
 
Zeuzzz said:
Its like an excited terrier with a bone.

If the theory is gonna get pwnd I'd rather it was done at the hands of an actual scientist, or maybe even the guy I actually asked here, thanks anyway, ben.
Click to expand...

I am a nuclear/particle astrophysicist.

Sol is a theorist.

This is one of the rare times I know the topic better than he does; pardon me for taking an interest.
 
ben m said:
I am a nuclear/particle astrophysicist.

Sol is a theorist.

This is one of the rare times I know the topic better than he does; pardon me for taking an interest.
Click to expand...

HA!!! Call to authority, if I ever saw one!
 
You must log in or register to reply here.

Back
Top Bottom