Arp objects, QSOs, Statistics

[Wangler]

Would you mind showing your working for the Group a and Group b calculations please?

Which of BAC's formulae and inputs did you use?

As presented by BAC:

P <= ((2*nk)/3)^r * (deltaz1 * deltaz2 * ... * deltazi=r)

where

nk is the number of Karlsson values in the range 0-3,

r is the number of quasars,

and deltazi is the distance to the NEAREST Karlsson value of the ith quasar,

assuming the distribution of quasar z is uniform in the range 0-3.

I have a spreadsheet I could post, I suppose.

First, I took all QSO redshifts, and subtracted from them the galaxy redshift.

Then, I obtained the absolute value of the difference of each QZO from the nearest Karlsson value. This is what BAC calls deltaZr.

Then, I used the formula presented above.

I am not certain I did it right, but what's what I did.

 

[DeiRenDopa]

Would you mind showing your working for the Group a and Group b calculations please?

Which of BAC's formulae and inputs did you use?

As presented by BAC:

P <= ((2*nk)/3)^r * (deltaz1 * deltaz2 * ... * deltazi=r)

where

nk is the number of Karlsson values in the range 0-3,

r is the number of quasars,

and deltazi is the distance to the NEAREST Karlsson value of the ith quasar,

assuming the distribution of quasar z is uniform in the range 0-3.

I have a spreadsheet I could post, I suppose.

First, I took all QSO redshifts, and subtracted from them the galaxy redshift.

Then, I obtained the absolute value of the difference of each QZO from the nearest Karlsson value. This is what BAC calls deltaZr.

Then, I used the formula presented above.

I am not certain I did it right, but what's what I did.

Thanks.

What values did you use for the Karlsson peaks?

What value(s) did you use for n_k?

I can reproduce your Group a result (more or less), but not your Group b one ...

 

[DeiRenDopa]

I am shocked to find half of our students are below average!

Clearly they are not from Lake Wobegon.

 

[Wangler]

Thanks.

What values did you use for the Karlsson peaks?

What value(s) did you use for n_k?

I can reproduce your Group a result (more or less), but not your Group b one ...

DRD, here is my Excel file. Calculations for this case are on Sheet 2. It is very sloppy, as I just slapped things in there without rhyme or reason. You will be able to see what constants I used for the Karlsson values, if nothing else.

 

[DeiRenDopa]

Thanks.

What values did you use for the Karlsson peaks?

What value(s) did you use for nk?

I can reproduce your Group a result (more or less), but not your Group b one ...

DRD, here is my Excel file. Calculations for this case are on Sheet 2. It is very sloppy, as I just slapped things in there without rhyme or reason. You will be able to see what constants I used for the Karlsson values, if nothing else.

Thanks.

In Group b, the first object's redshift (z = 0.607041) has been transcribed incorrectly (z = 0.608041); in Group a, the (small) difference between our results arises from your lowest Karlsson peak being 0.3, and mine being 0.06 (there is also, thus, a knock-on difference due to the number Karlsson peaks).

This is gratifying; it seems that you and I have both understood how to do a calculation according to 'the BAC approach' the same way (except for whether there is a Karlsson peak at 0.06 or not).

Oh, and also whether you need to subtract the z of the 'parent' galaxy from the z's of each of the 'quasars predominantly along the minor axis' first, or just use the observed quasar redshifts.

Perhaps BeAChooser could weigh in on this?

BAC, are Wrangler and I doing the calculations, according to 'the BAC approach', correctly?

Is there a Karlsson peak at 0.06, or not?

Do you subtract the z of the 'parent' galaxy from the QSOs' z's first, or not?

 

[Dancing David]

Have you found a good citation for the Karlsson peaks, or are they more acurately Burbidge peaks?

The ones I posted earlier seemed, um, rather broad to be peaks, more like large plateus.

 

[DeiRenDopa]

Have you found a good citation for the Karlsson peaks, or are they more acurately Burbidge peaks?

The ones I posted earlier seemed, um, rather broad to be peaks, more like large plateus.

.

That's a question I already asked BAC (in a somewhat different form), but have yet to get an answer (not that I'm holding my breath, there are rather a lot of 'BAC unanswered questions').

BAC earlier gave Karlsson, K. G., "Possible Discretization of Quasar Redshifts", (1971) Astronomy and Astrophysics, Vol. 13, p. 333 as the original source for these. In just one quick search I turned up a similar, though different, set of numbers (from http://www.haltonarp.com/articles/origins_of_quasars_and_galaxy_clusters ); no doubt it you went looking you could find several other versions ...

In any case, it may not matter much; if BAC chooses to confirm that Wrangler (and I) have understood 'the BAC approach' sufficiently well to be able to correctly calculate what he calls probabilities, then it may be easy to show there are some odd things indeed about this approach (or, more likely, that I have, once again, failed to understand it).

 

[Wangler]

Any word from BAC yet?

Has he confirmed we are doing the calculations right?

Have we proved, without a doubt, that quasars are ejected along minor axes of galaxies, with quantized redshifts?

I can see the abstract reference now:

"Statistical Confirmation of QSO/Galaxy Associations: Let The New Age Begin", BeAChooser, DeiDenRopa, Wangler"

 

[Dancing David]

Well, if this is like last year, he is very busy right now. (That is my assumption, he is out on the road, in the bush or something?) On the another problem with the bang thread he disappeared for quite a while. He then returned two to three months later in another thread.

 

[DeiRenDopa]

Any word from BAC yet?

Has he confirmed we are doing the calculations right?

Have we proved, without a doubt, that quasars are ejected along minor axes of galaxies, with quantized redshifts?

I can see the abstract reference now:

"Statistical Confirmation of QSO/Galaxy Associations: Let The New Age Begin", BeAChooser, DeiDenRopa, Wangler"

Hmm ...

Actually, I think we are well on the way to showing, in concrete terms, using explicit examples from both BAC's chosen dataset and others, that his approach does not do what he claims it does.

I'll be posting more material over the next few days, the results of my investigations.

However, it could be that no one, other than BAC, has actually understood his approach, despite the dozens of hours he obviously spent on developing it and writing it up for 'publication' in the JREF forum ...

 

[Wangler]

Hmm ...

Actually, I think we are well on the way to showing, in concrete terms, using explicit examples from both BAC's chosen dataset and others, that his approach does not do what he claims it does.

I'll be posting more material over the next few days, the results of my investigations.

However, it could be that no one, other than BAC, has actually understood his approach, despite the dozens of hours he obviously spent on developing it and writing it up for 'publication' in the JREF forum ...

Does this approach hold any water at all?

I will try to complete my analysis of your other data sets, and correct my calculations of the first sets including the 0.06 Karlsson peak.

Then we can compare the final answers.

 

[Dancing David]

After looking at NGC 3256 of APOD, I thought, hey here is one that ought to be spitting out QSOs and there fore in the Arp catalogue. Here is what I did find
http://www.google.com/search?hl=en&q=NGC+3256+QSO


Gosh, more converging evidence that QSOs are galaxies.

Hmm, what was that definition of a QSSO, oh, "object conveniently located close to an Arp galaxy!"

 

[DeiRenDopa]

Taking Group a) and b) from post #439, the 7 'Karlsson peaks', and the two distributions in post #440, and using the BAC approach without correcting for the 'ejecting galaxy' redshift, I get the following 'BAC probabilities' (two significant figures in all cases):

a) Amaik: 0.011; Karlsson: 0.0041; 'regular': 0.00082

b) Amaik: 0.30; Karlsson: 0.00018; 'regular': 0.0011

'Correcting' for the putative ejector, I get the following 'BAC probabilities':

a) Amaik: 0.011; Karlsson: 0.0034; 'regular': 0.0094

b) Amaik: 0.16; Karlsson: 9.2x10^-5; 'regular': 0.0017

I have also found a set of seven 'peaks' (let's call then 'DRDS peaks') that are pretty cool.

Here are the four 'BAC probabilities' for these 'DRDS peaks':
without 'correction': a) 0.00053, b) 1.9x10^-6with 'correction': a) 0.00034, b) 2.6x10^-6
Next, some calculations of 'BAC probabilities' using the seven datasets in post #444 (they will all be 'without correction' calculations).

 

[Dancing David]

Wow, that is like watching CNN parse the democratic nomination.

I am fuddled. What does it mean again?

 

[Wangler]

Wow, that is like watching CNN parse the democratic nomination.

I am fuddled. What does it mean again?

I think it means Obama won................

Actually, those are the probabilities associated with groups a and b.

Basically, what are the probablilities that they are a certain distance from certain "quantized" redshift values?

I need to get going and finish my analyses for comparison.

 

[Dancing David]

I just find the peak thing strange (like when they start to sling the stats on CNN), you have a distribution for the red shift of QSOs at a small range of magnitude and then you look at the distance of the QSO from the 'peak' in that magnitude range. Except that it isn't a peak, it is a platue and nobody is using a distributive standard deviation. So BAC's, Burbide, Arp and Karlson are playing games.

What is the SD for the range of a 'peak' and how many of these QSO redshifts fall outside that SD.

And I bet that it will approach a bell curve as the number of QSOs in the magnitude slice increase in the base of data.

Funny stuff.

 

[DeiRenDopa]

I just find the peak thing strange (like when they start to sling the stats on CNN), you have a distribution for the red shift of QSOs at a small range of magnitude and then you look at the distance of the QSO from the 'peak' in that magnitude range. Except that it isn't a peak, it is a platue and nobody is using a distributive standard deviation. So BAC's, Burbide, Arp and Karlson are playing games.

What is the SD for the range of a 'peak' and how many of these QSO redshifts fall outside that SD.

And I bet that it will approach a bell curve as the number of QSOs in the magnitude slice increase in the base of data.

Funny stuff.

Look at it like this ...

I said, repeatedly, that I did not understand how BAC did his calculations, nor what they were supposed to mean (as in, what legitimate conclusions can one draw from the calculated numbers?).

BAC said, repeatedly, that I was being deliberately evasive, and other things.

Wrangler posted the results of his calculations done using the BAC approach. I was able to reproduce them (other than for some easily-fixed errors).

That gave me confidence that anyone - well, me at least - can take the BAC approach and get objectively reproducible answers.

That also means anyone can apply the approach to samples that are 'quasars' near bright, low redshift spiral galaxies (whether active galaxies or not), along any axis; to samples that are 'quasars' in randomly chosen fields, along any axis; to samples of made up 'quasars' (generated by random numbers, using an algorithm that has nothing to do with real quasars); to all the above, with 'peaks' other than 'Karlsson peaks'; and so on.

If it turns out that Wrangler and I continue to get the same answers ('BAC probabilities') in all cases, then we can use the many different kinds of samples to ask questions about what the 'BAC probabilities' actually mean (well, we've always been able to ask questions; now we will have some objective, reproducible data to use in our questions).

As I understand it - and I freely confess that I may not, at all - a 'BAC probability' is the probability of finding quasars with that particular set of redshifts, near a position on the sky, along a particular axis. Further, BAC has insisted (I think ... I could be wrong) that:

* BAC probabilities will be lower, much lower, for quasars predominantly along the minor axes of bright, low redshift spiral galaxies

* but only for 'Karlsson peaks'.

Of course, we have a sample that is too small for any decent testing, but maybe it will be indicative.

I just noticed one thing I haven't done - see if I can reproduce BAC's numbers on NGC 3516 and NGC 5985 'quasars' (he posted the results of his calculations for only these two, right?); I'll have a go at that later today ...

 

[DeiRenDopa]

From post #444:

Set ONE: 0.0547, 0.58, 0.720, 1.084, 1.12463, 1.41, 1.903, 2.05812, 2.10, 2.244, 2.6068, 2.98, 2.99, 3.0347, 3.14.

Here are the 'BAC probabilities' I calculate:

"Amaik peaks": 0.0020

"Karlsson peaks": 1.7x10^-6
"regular peaks": 8.4x10^-7
"DRDS peaks": 0.00090

How are you coming along with your calculations, Wrangler?

 

[Wangler]

From post #444:Here are the 'BAC probabilities' I calculate:

"Amaik peaks": 0.0020

"Karlsson peaks": 1.7x10^-6
"regular peaks": 8.4x10^-7
"DRDS peaks": 0.00090

How are you coming along with your calculations, Wrangler?

I am sorry to have fallen behind. I am buried by data at work that I have to get back to our QA Dept., as well as working on a report or two.

My evenings have been consumed with hockey playoffs.

I will try to get my stuff done no later than Friday evening.

Hopefully BAC can look at this stuff this weekend?

 

[DeiRenDopa]

From post #444:

Set TWO: 0.467142, 0.520394, 1.08367, 1.21187, 2.58, 2.90609.

Here are the 'BAC probabilities' I calculate:

"Amaik peaks": 0.0094

"Karlsson peaks": 0.043

"regular peaks": 0.0066

"DRDS peaks": 0.0012

And

Set THREE: 0.150221, 0.352114, 0.369104, 0.64384, 0.84003, 0.86495, 0.947653, 1.09562, 1.20564, 1.56234, 1.67964, 1.97046, 2.04535, 2.05185, 2.24764, 2.99602.

Here are the 'BAC probabilities' I calculate:

"Amaik peaks": 2.9x10^-6
"Karlsson peaks": 9.7x10^-7
"regular peaks": 4.6x10^-8
"DRDS peaks": 4.4x10^-5