Evidence of Gravitational Waves, or Confirmation Bias?

[Michael Mozina]

I have no idea what you're talking about here. I am not party to whatever discussion you are referring to and in any case I see no relevance to our discussion here.

It's relevant in the sense that my analogy clearly demonstrates that cherry picking data sets allows one to make about any claim under the sun with respect to a "candidate signal" and the concept of "statistical noise" can be misleading. In LIGO's case their statistical noise is devoid of most of the more likely environmental influences that could cause that signal, so their sigma calculation cannot be used to eliminate environmental influences.

If the sigma isn't related to the cause in any physical way, it's just a useless math exercise in the final analysis. LIGO started by *stripping out* the most likely environmental causes of the signal and cherry picked only the data they wished to "compare" their candidate signal to. They "eliminated" statistical (cherry picked) background influences, but they couldn't possibly eliminate *all* environmental influences that way.

You see the thing that really puzzles me Michael is all this fuss you are making here (and on other forums apparently) about how the sigma of the detection doesn't relate to cause, and when I point out to you that it doesn't relate to cause, it can't relate to cause and no-one expects it to relate to cause, you start saying it's useless.

In this case then, where's the "discovery" if it's not related to cause?

The reason I'm puzzled is that determining the probability of whether an apparent signal is signal or random noise is completely routine, not just here but in every branch of physics. I'm amazed that someone like you who takes such an interest in physics doesn't know that. I don't understand why this is at all controversial.

It depends on the specific situation. If we're talking about particle physics "discoveries", the mere presence of something that cannot be generated by (non cherry-picked) statistical noise might actually constitute a "discovery". In this case there was a cherry-picking process up front, and the "statistical noise" therefore means almost nothing with respect to eliminating other potential "causes" of this "chirp" signal.

You measure the statistics of your static random noise (arising from shot noise in the electronics, photon pressure noise and so on) and you determine the probability that that static random background could give rise to the thing you think is a signal, and you express that probability as a sigma.

In terms of removing "photon pressure noise" from consideration, that's fine. In terms of removing *all environmental noise*, it's a useless figure.

(Oh and by the way, in determining the statistics of the random static noise, you have to do so at a time when the system is quiet and not subject to transients, whether true signal or transient noise. See below)

You really should look my "aliens did it" analogy. It clearly demonstrates the danger of cherry picking the data set.

(There is also continuous noise associated with specific sources such as mains and its harmonics, resonances in the mirror suspension systems and so on which you can notch filter out of the data since these are tones).

Well, I already capitulated to Selfsim with respect to using that sigma to eliminate various *internal* (to the equipment itself) potential influences, but it no way can it be used to eliminate any *external* causes from consideration. It's actually "questionable" even in terms of internal noise because of the blatant cherry picking problem. LIGO only selected their "best case" data set, not the full data set. It's still a cherry picked data set with respect to various internal influences too.

Now let's say that, as in this case, the probability that the putative signal could arise from the static random background noise is very low - the statistics of the backround noise compared with the putative signal indicates that the signal would only arise from the random noise once in 100,000 odd years. OK, so you have a signal, but you're not out of the woods yet. You still have to eliminate the possibility that your signal is caused by either intrumental or environmental transients - earthquakes, lightning strikes, heavy trucks moving about near the experiment, North Korean nuclear tests, doors slamming in the office, blip transients etc etc. In this case LIGO has a large number of channels with accelerometers and magnetometers and seismometers and microphones and so on and so forth to eliminate envirionmental transients. Once you have done all that, and the environment is clean and the kit is running clean, you can at last say that you have a proper signal.

Why are they even comparing this signal to a "clean" data set in the first place if you're trying to eliminate environmental contamination from being the cause of the signal? Keep in mind that this exact signal was first *vetoed* by the software which was in place at the time of the signal. This also brings up the blatant confirmation bias problem inherent in LIGO's methodology. All other potential influences are said to be 'eliminated' from further consideration based upon a lack of an observed signal in some type of *external* equipment or measurement, whereas all celestial origin claims are simply given a free pass without respect to any elimination process at all. There's not even a potential category for "unknown cause". Why? How can that even be considered a "scientific" methodology if the method favors all celestial origin claims and disfavors all other potential causes of the same signal?

But you're still not out of the woods, because the next question is did the other site see the same signal, and what was the coherence between them? If you do time slides of the data from the two sites does a slide within a few milliseconds give a big response? If only one site sees it, throw it away. If both sites see something but the time difference is too great, throw them away. In this case you have a very strong signal which has a high degree of coherence with a physical time difference between the two sites. You're still not out of the woods, because you have to match the signal you now have with merger templates, and if, and only if, your coherent signal appearing at both sites matches a template for inspiral, merger and ring-down, then you think you might have a candidate.

Since they had *just* completed a major upgrade of this equipment, and they were only in the engineering run of this new configuration, how could they even know for sure that ordinary "blip transients" are *never* detected by both detectors, or that that they *never* "chirp" (change frequency)? These two "human decisions" look to be highly questionable choices considering that this equipment was much more sensitive than previous versions, and they had not even completed the engineering run. They could have just as easily have "assumed" that blip transients can sometimes be observed in their new equipment, and they can sometimes "chirp" (change frequency). That seems like a purely *arbitrary choice* that was specifically intended to skew the results, just like the cherry picking, and just like the confirmation bias of having no exclusion method for celestial claims.

But the main point is that LIGO is using sigma estimates in a completely routine, normal and boring way, and I am surprised that you are exercised by it.

I'm exercised by it for three main reasons. They blatantly and intentionally *cherry picked* the data set, they didn't use the whole set of data. They also have no way to exclude environmental data this way. They have no way to exclude blip transients this way either without making two *questionable assumptions* about them which could not possibly be known at the time of the signal.

In addition to those three problems, they have a *huge* problem with confirmation bias because their methodology favors all celestial origin claims, and disfavors all other potential causes of candidate signals. Without a visual/neutrino confirmation requirement, there is simply no way to be sure where this signal came from.

As I understand it the veto has been explained - it came from an uncalibrated strain channel that should have been labelled unsafe, but wasn't, and was manually labelled unsafe afterwards.

Nobody outside of LIGO even knows why that veto was added or what specific type of "environmental' influence is was trying to detect. We don't know why it rejected this exact signal with a 'high confidence' figure. We don't know what the term 'safe" even means in any quantified way. Maybe it did 'safely' remove the exact environmental factor that LIGO was originally trying to filtered out with this veto. Why specifically was that veto even added in
the first place? What was it designed to do?

I'm not sure you understand what vetoes are and what they are for. Vetoes are not intended to dictate conclusions.

Except that's exactly what they did with respect to claiming to have eliminated environmental noise from consideration. Frankly I'm also *appalled* that they gave a whitewashed account of veto events. I've never seen such a thing before.

They are an aid, not a supervisor. The data is noisy, with high levels of static and transient noise. Multiple channels look at the strain data in multiple ways and trigger potential candidates - dozens per day. Everyone would go mad if they had to analyse all those candidates, so automatic vetoes are used to say "ignore this, it's probably noise; ignore this, it's not a big enough signal; ignore this, it's the wrong shape". Nothing says that people can't over-ride a veto, if on analysis they disagree with it. But in this case there wasn't even a valid veto in the first place as it came from an uncalibrated strain channel, that should not have been allowed a veto in the first place. There was no valid veto (by the way, it wasn't measuring the environment - it was a strain channel).

Why *specifically* was that veto written in the first place, and what was it trying to remove so it didn't drive folks mad? Why did it achieve a "high confidence" figure in the first place, and how "safe" was it to simply override that veto? How 'safe' is safe? 80 percent safe? 90 percent safe? 5.1 sigma safe? How does anyone know that it didn't correctly veto out exactly what it was designed to filter out in the first place?

That only brings us right back to their blatant confirmation bias problem. They eliminated *non-celestial* claims based uponi a *lack* of external corroboration in external hardware. They then blatantly *failed* to put their own celestial claims through that very same process of elimination. They simply rigged the system in their favor.

And it brings me back to the point that you know what you're looking for and when you see it, it's not unreasonable to say so. I fail to see where the confirmation bias is.

The confirmation bias comes in when there's no elimination method applied to celestial origin claims as there is for any and all other claims of origin. They eliminated ground movement and EM field influences from consideration based upon a *lack* of confirmation in some other external equipment. They looked and looked and looked for evidence to support a celestial event and they didn't find one. In fact, they haven't found one in any of the three events that they claim are related to celestial events! How likely is it that they'd really be 0 for 3 if these really were celestial in origin?

The only way you can even try to build such a case is to begin with a special pleading fallacy. We couldn't see them because......

If you're talking about counterparts, it's always been considered that seeing a counterpart of a neutron star merger would be great physics, but no-one ever said as part of the LIGO experimental design that seeing a counterpart is needed to claim a detection - and in fact no counterpart or a very difficult to detect counterpart is expected from a BH merger.

They're 0 for 3, and according to the Nature article, they have 6 more "candidate signals" and I've not heard a single peep about *any* of them enjoying a single shred of visual support. What confidence can I possibly have in such a claim? You've claiming that two entirely invisible entities emitted multiple solar masses of energy in a quarter of second, but they magically did so without emitting enough EM emissions to be seen from Earth. Doesn't that sound even the least bit suspicious to you? Even if black holes themselves to not emit light, the plasma around the black hole tends to be brighter than any other plasma in space. That last claim even suggested something about "spin" angles not matching which would likely result in *massive* interactions of the accretion disks. If they happen to be charged, you could end up with discharges taking place inside the accretion disks too.

This isn't like particle physics discovery where *any* type of deviation from the standard model is a "discovery" of something new. Particle physicists also don't "cherry pick" their data.

I really don't know what you mean by cherry picking. What data was cherry picked?

They had over 40 days of data to work with but they only selected 16 of the "cleanest" days they could get their hands on, presumably to get the highest sigma possible.

Then where's the "discovery" in this paper?

The detection of gravitational waves caused by a black hole merger as inferred from the fit between the data and the theory and the elimination of all other likely causes,

They didn't eliminate environmental causes with that sigma!

using an instrument designed to detect gravitational waves.

It also "detects" all sorts of stuff that has nothing to do with gravitational waves.

See figure 12:
http://iopscience.iop.org/article/10.10 ... 34001/meta
https://arxiv.org/abs/1602.03844

They even mention that this blip transient event fits with a neutron star/black hole merger pattern.

Well, it's a strong signal, and it's compared with the best fit neutron star/BH merger template, which it doesn't fit very well (too short on the inspiral side, no ring down),

Very well? This is where the quantified aspects look to be far less "quantified" than one might hope, particularly if we're claiming "discovery".

and of course it only appears at one site.

But how can we know that that something like a long distance electrical discharge *between* the two detectors, or rare power grid problems, or whistler wave from high in the atmosphere might not influence both detectors at once? Both detectors just got a *major* upgrade that made them 10 times more sensitive than ever before in terms of distance, and a 1000 times more sensitive in terms of volume space.

There is no frequency evolution and it doesn't match a BH merger. So, no, blip transients don't match the candidates at all.

Except that exactly not what they claimed.

Since they're in the engineering run to test the capability of their heavily upgraded equipment, how could they possibly *know* yet that blip transients cannot *ever* be detected by both detectors at the same time? The same question applies to their bald-faced *assumption* about blip transients *never* "chirping" occasionally. If because of various environmental factors, it's just a "rare" blip transient that chirps and can be observed by both detectors, it would still end up being classified as a gravitational wave "discovery".

And if I had bread I could make a sandwich if I had some cheese. No blip transients have ever been detected that match an inspiral, merger and ring down signal in both detectors at a physical delay, and especially not a BH merger signal.

That depends *entirely* as to whether you include the 3 in question (print) and the 6 other ones they claim to be reviewing into the category of "blip transient", or you claim they are "gravitational waves". It's a purely human decision to make that distinction, and ultimately it's still an "act of faith" in the absence of any corroboration from any satellite in space or on the ground.

The bottom line here is that there is no way to be sure that any particular candidate signal is celestial in origin without some type of visual or neutrino confirmation of that claim. Without it, the candidate signal should simply end up in an "unknown origin" category at best case, certainly not a 'discovery of aliens' category as in my simplified example.

But this is your standard, not the standard set by the experiment, and I have to say that I think yours is unreasonable. [/quote]

It's the *scientific* standard to not use a methodology which favors one explanation over another. In this case their "experiment" didn't favor anything, but their "methodology" certainly favors all celestial origin claims, and certainly disfavors everything else. A level playing field is a reasonable standard. A non level playing field is not a reasonable standard, particularly when you're trying to claim that you've "discovered" something new.

First of all BH mergers are not expected to have a counterpart, or if they have it would be weak.

Says who? The one published paper that I cited in my PDF predicts that the merger will emit more energy in the EM spectrum than in gravitational waves.

Second, these events are very distant.

Yet you claim to "see" them in "gravitational waves", but quite miraculously not a single one of the three published events manages to emit enough EM radiation to be observed in *any* satellite or neutrino detector on Earth or in space.

We worked out together a difference in apparent and absolute magnitude of 40 for z=0.2. So they will be faint (at least BH mergers will be faint - NSBH and NS mergers are expected to be brighter).

Were did you get your original brightness figure from?

Third they are very short lived, particularly BH electromagnetic counterparts if they are detectable at all.

How do you "know" that the EM emissions would be "short lived"? Supernova events last a long time due to supposed "time dilation", and yet not one of these events was even seen in gamma ray burst detectors or neutrino detectors either, neither of which requires a particularly long signal.

And fourth, they require the good luck to have a powerful telescope looking in exactly the right direction at the right time.

Gamma ray burst detectors can pinpoint a powerful event in very short order and they are often used to point other instruments. The fact that LIGO's job remains "difficult" doesn't give them the right to cut corners.

So counterparts are not regarded as necessary for a detection.

That specific choice is pure confirmation bias, and instantly favors celestial claims and disfavors all other possible causes of the signal which *require* a counterpart or they are rejected.

You disagree and I don't see how that disagreement can be resolved by discussion. But it is you against the entire community.

Yet that entire community has been 0 for 3 in terms of providing any visual support of their claims, while I remain 3 for 3 in my prediction that since the signals are not likely to be celestial in origin, there is therefore no EM 'counterpart' to find. So far I'm batting 100 percent, while the entire community is batting 0. All it would take is for even one of the three events to have be identified and it would be obvious that LIGO was right, and I was wrong, but alas the whole of astronomy has been unable to demonstrate that these signals are celestial in origin.

I must say that the third "non detection" of an EM counterpart only made me more confident in my position, as well any additional non detection of EM counterparts in future papers on this topic. How is you claim even "falsifiable" or "testable" in comparison to mine *other than* your complete inability to produce one with an EM counterpart? How long do I have to wait to know if I'm right, and why shouldn't they have to wait before they claim that it's a "discovery" in the first place?

However, LIGO has engaged all along with the light astronomy folk, and sooner or later I expect a counterpart will be detected. I don't see anything wrong with you expressing thoughtful caution and saying that you don't accept the discovery until then (although the community disagrees), but I do think some of the things you're banging on about, sigma and the veto in particular, are ill-conceived.

The sigma problem isn't really even the worst part of their methodology. The worst part of their methodology is that it is unable to differentiate itself from my interpretation of the "cause" of these signals being unlikely to be celestial in nature. The fact they can't even falsify their own claim says it all. It's a *terrible* methodology that they are using which is why they're batting 0 for 3 and I'm already 3 for 3.

They also claim to have 6 more possible candidate signals, and they would definitely publish one that had a visual confirmation if they had one, so we know that if the next paper doesn't have an EM counterpart, I'm probably already 9 for 9, and they're probably already 0 for 9. We'll have to wait and see what they publish next and see if it includes a EM counterpart, but I must say that I'm becoming more confident in my position every single time they fail to produce any visual support of their claim, or any neutrino support of their claim.

If I'm right, this is only going to become more obvious once three detectors are online, and nobody can find an EM counterpart when the region of the sky is well constrained. That's *assuming* that even all three detectors see the same signal at the same time, which might not even happen if the real cause of the signal is environmental in origin.

I think LIGO jumped the gun and they painted themselves into a difficult corner. Who's going to now admit a mistake if they indeed made one?

[Zyxzevn]

Gravitational wave 2017 RAW data


I put it next to the 2015 one.

Again we see the Amplitude Modulation. But somehow the H-detector (red) has a
more constant resonance. Maybe there is an error in the downloaded data?
But for now I assume this data is correct.

This fairly constant amplitude is actually a positive point, because now we can
use the modulation frequency to test for gravitational waves too.
Gravitational waves are supposed to change the timing of these pulses.
So if any time-shift occurred, we would see this in the modulation signal.
And, for now, I do not see any changes in the timing of these signals.
Attention: No time-shift means a NULL-result for the gravitational wave.

The modulation signal is caused by LIGO's laser recycling, which works like an echo.
Its base frequency is 268 Hz, due to the path the laser takes.
The interferometer can also recycle the laser signal, which can cause echoes on
in time-multiples of the base frequency.
So the interferometer can produce echo signals of 134 Hz, 88 Hz, 67 Hz, 44 Hz. etc.
Assuming the laser itself only causes positive feedback of the echo.

Additional sources for signals are the mirrors themselves, which are hanging in the air
like gongs, while being bombarded with photons.

The non-linear noise caused by the echoes,is in the same frequency range as
the theorized Gravitational Wave signals.

This means that the many echoes in the LIGO system can indeed cause false positives.

[Zyxzevn]

There is now an AMA on reddit by the LIGO scientific collaborators.
See link

I added a question about how they dealt with the echoes in their detector.

[Cargo]

I think it's over already and the amount of HandWavium used by the LIGO team is probably on the order 5E22 Solar Masses.

The Folly of Man.

[kell1990]

Here you go:

http://www.scirp.org/journal/PaperInfor ... erID=71246

Face it. *Without* any visual or neutrino confirmation of your claim about these signals being celestial in origin, you're *guessing* about the real cause of the signal. You haven't got any real evidence that the signals in question are related to celestial events. They look to be *power grid related* events.

I originally figured that the signal was a whistler wave in the magnetosphere, but after reading their paper, I'd be inclined to agree with them that it's probably a power grid related phenomenon. That's even closer to home than my original idea, and the filters show there is a correlation between them.

Of course that paper attacks the whole LIGO technology and the whole concept of gravitational waves, so it's even more "skeptical" of your claims than I am.

>Me Well, I am even more skeptical that Mozina about the existence of black holes. I can almost guarantee you that they do not exist; rather, what you see in the telescopes is actually a Bennett pinch. Wait until the Webb telescope goes up and see. It is how the universe regenerates itself. One day "scientists" will discover this.<me

Your derogatory comments have been dismissed.

The fact of the matter is that in spite of a huge amount of press coverage the LIGO team doesn't really know what it has "discovered". To be sure they think they have found evidence of a pair of twirling twin black holes, but there is a huge problem with that: An earlier press release stated that a massive black hole (even if it existed) could have hurled our a planet the size of Neptune.

What's the matter with this guess? It is that black holes with their supposed mass (and gravitational attraction) could capture anything within its realm, including light itself. So how could such a body such as Neptune be "hurled out"? Not possible, according to the current definition of black holes.

Maybe some of you guys need to go back to the drawing board and try to come up with a better guess.

[Zyxzevn]

Maybe some of you guys need to go back to the drawing board and try to come up with a better guess.

I think that this signal is just an echo of some noise.

[kell1990]

There is now an AMA on reddit by the LIGO scientific collaborators.
See link

I added a question about how they dealt with the echoes in their detector.

Here are Reddit's requirements for responses:


Submission Requirements

Directly link to published peer-reviewed research articles or a brief media summary
No summaries of summaries, re-hosted press releases, reviews or popular reposts (over 100 upvotes)
Research must be less than 6 months old
No sensationalized titles, all titles must include the model where applicable
No blogspam, images, videos, infographics
All submissions must be flaired and contain a link to the published article, either in the submission link or as a standalone comment.

Comment Rules

On-topic. No memes/jokes/etc.
No abusive/offensive/spam comments.
Non-professional personal anecdotes may be removed
Arguments dismissing established scientific theories must contain substantial, peer-reviewed evidence
No medical advice!
Repeat or flagrant offenders may be banned.

Got that? You need a peer-reviewed paper to even be considered, and various other humdrum.

This is how they shut out competing ideas. See how easy this is? Goebbels would be proud.

[Higgsy]

Gravitational wave 2017 RAW data


I put it next to the 2015 one.

These traces are completely uninformative unless you label the axes. I assume they represent strain against time. So if the y-axis is strain is it of the order of 10-5, or 10-15, or 10-25? If the x-axis is time are we looking a 10msec, or 1 second or 100 seconds? Unless you label the axes, these traces are meaningless.

Again we see the Amplitude Modulation. But somehow the H-detector (red) has a
more constant resonance. Maybe there is an error in the downloaded data?
But for now I assume this data is correct.

What is the frequency of the your so-called amplitude modulation? I don't mean what do you guess it is - I mean in the actual data what is the frequency you see? You do realise that LIGO has sensitivity plots in equivalent strain, which are, in effect, a Fourier transform of the background noise which contain strong tones (not 268Hz) which are known and accounted for?

This fairly constant amplitude is actually a positive point, because now we can
use the modulation frequency to test for gravitational waves too.
Gravitational waves are supposed to change the timing of these pulses.
So if any time-shift occurred, we would see this in the modulation signal.
And, for now, I do not see any changes in the timing of these signals.
Attention: No time-shift means a NULL-result for the gravitational wave.

No. First of all there are no "pulses" in the strain signal (there are a few known tones), and secondly even if there were pulses, the time shift from a gravitational wave is so tiny that it is unmeasurable. For a signal with peak amplitude of strain 0.5x10-21, typical of signals detected so far, the time shift would be unimaginabbly small at 6.7x10-24 seconds. That is why LIGO uses interferometry to measure the strain.

The modulation signal is caused by LIGO's laser recycling, which works like an echo.
Its base frequency is 268 Hz, due to the path the laser takes.
The interferometer can also recycle the laser signal, which can cause echoes on
in time-multiples of the base frequency.

No. I have explained before that that is not possible. The LIGO interferometer is based on a continuous working, not a pulsed source, and the Fabry-Perot cavities for signal and power recycling do not and cannot introduce any modulation at 268Hz or any other frequency. You are completely wrong about this. How do you know there is a tone at 268Hz other than wishing it so? You haven't measured 268Hz in the traces above, have you? Whereas LIGO's own noise sensitivity shows no tones at 268Hz. No noise there at all.

So the interferometer can produce echo signals of 134 Hz, 88 Hz, 67 Hz, 44 Hz. etc.
Assuming the laser itself only causes positive feedback of the echo.

No, it can't. You are completely wrong.

Additional sources for signals are the mirrors themselves, which are hanging in the air
like gongs, while being bombarded with photons.

You are right here, there are resonances from the mirror suspension which is at a distinct tone and is notch filtered out.

The non-linear noise caused by the echoes,is in the same frequency range as
the theorized Gravitational Wave signals.

You clearly have no idea what the term non-linear means, but there is NO noise created by the power and signal recycling. None.

This means that the many echoes in the LIGO system can indeed cause false positives.

No, they can't.

[Higgsy]

It's relevant in the sense that my analogy clearly demonstrates that cherry picking data sets ...

I think LIGO jumped the gun and they painted themselves into a difficult corner. Who's going to now admit a mistake if they indeed made one?

I don't think there is any point in continuing this discussion. I have already said that if you personally want to withhold acceptance of the discovery until a counterpart is detected, because a signal that looks exactly like a BH merger might appear in both sites at the same time and be caused by an unknown, never before seen glitch, then that is fine. But with regard to the other subjects, signal versus noise sigma, "cherry-picking" and the veto, you are simply not listening. Every time I have explained something to you, for example the fact that sigma does not relate cause, and no-one (except maybe you) expects it to, you just come back with the same unreasonable beef. It has nothing to do with eliminating transients, it never did and it never will, neither in LIGO or in any other phyics experiment in the universe, and those transients are eliminated by other processes. I took you through a simplified version of the whole process which you simply ignored. So for whatever reason, you seem to have an irrational commitment to your misconceptions about sigma, measuring steady-state noise and the veto, which it is clear that no amount of discussion can shift.

But I am intrigued. I believe you are disussing this on other forums - aren't other people telling you the same things as I am? If multiple people are telling you the same thing, don't you think there might be something in it?

[Higgsy]

The fact of the matter is that in spite of a huge amount of press coverage the LIGO team doesn't really know what it has "discovered". To be sure they think they have found evidence of a pair of twirling twin black holes, but there is a huge problem with that: An earlier press release stated that a massive black hole (even if it existed) could have hurled our a planet the size of Neptune.

What's the matter with this guess? It is that black holes with their supposed mass (and gravitational attraction) could capture anything within its realm, including light itself. So how could such a body such as Neptune be "hurled out"? Not possible, according to the current definition of black holes.

I don't know what specific observation you are referring to here (perhaps the hypothesis that the supermassive BH at the galactic centre rips and ejects stars which the recoalesce into planet sized lumps?), but black holes are known theoretically and experimentally to accelerate matter in their vicinity that does not fall into the event horizon. Obviously, if it does fall in then it's lost forever, but not all matter accelerated by a BH falls in. Google black hole jets for more info.

[Michael Mozina]

It's relevant in the sense that my analogy clearly demonstrates that cherry picking data sets ...

I think LIGO jumped the gun and they painted themselves into a difficult corner. Who's going to now admit a mistake if they indeed made one?

I don't think there is any point in continuing this discussion. I have already said that if you personally want to withhold acceptance of the discovery until a counterpart is detected, because a signal that looks exactly like a BH merger might appear in both sites at the same time and be caused by an unknown, never before seen glitch, then that is fine.

What do you mean by "never before seen"? LIGO has routinely observed blip transients in both detectors, and they seem to fit the same basic mathematical mass merger models they're comparing various signals to. All that would be required is for the increased range of the new detectors to allow for them to see blip transients in both detectors at once, and ordinary "blip transients" end up being classified incorrectly as "gravitational wave signals".

Since they were still in ER8 mode, and presumably *testing* their upgraded equipment, I don't even understand how LIGO had enough time to collect enough raw data in the new detectors to make any determination about what blip transients might look like in the upgraded equipment. It seems like they just "assumed" two things about blip transients that they couldn't actually "know" (with confidence) about them.

Worse however is that they *cherry picked* the statistical noise that was already filtered and stripped of environmental noise, and therefore their sigma means absolutely nothing with respect to eliminating normal environmental noise as the real cause of these signals.

But with regard to the other subjects, signal versus noise sigma, "cherry-picking" and the veto, you are simply not listening.

It seems like it's you that aren't listening from my perspective. If they cherry picked only the data, then the sigma figure is useless as it relates to eliminating environmental processes from consideration. It's just that simple.

Every time I have explained something to you, for example the fact that sigma does not relate cause, and no-one (except maybe you) expects it to, you just come back with the same unreasonable beef.

Since they're using that sigma figure to claim that this is a "discovery", it's not really an unreasonable beef at all. If they left in the ordinary environmental noise and still had a five sigma confidence figure, I might be impressed. As it stands that sigma figure means absolutely nothing in terms of supporting their claim as to cause.

If they *did* have a visual confirmation of one of these signals, it might actually be possible to calculate the odds that these two events (signal and visual verification) are related based on their triangulation methods and expected distance measurements. Without any visual however, they will *never* be able to tie the sigma back to a specific cause.

It has nothing to do with eliminating transients, it never did and it never will, neither in LIGO or in any other phyics experiment in the universe, and those transients are eliminated by other processes.

Sure, but that sigma figure does *NOT* eliminate any of the other environmental factors *besides* blip transients. If you read my paper, you'll find that the sigma issue is more directly related to LIGO's lack of any ability to eliminate environmental factors as being responsible for the signal.

The blip transient problem is a separate issue (and section) entirely because presumably LIGO cannot filter out blip transients in the first place. That problem goes back to their *two primary assumptions* about blip transients, while still in an *engineering* run phase of highly upgraded equipment. I have no confidence in their two main assumptions about blip transients, A) that they cannot *ever* be observed by both detectors at the same time in the upgraded equipment, or B) they don't ever produce a "chip" type signal. Without making those two assumptions, there's simply no way to logically eliminate blip transients as being the actual cause of the signal. Since they had no real history yet using these upgraded detectors, I see no logical basis for their "assumptions" in the first place. That's my complaint about blip transients.

I took you through a simplified version of the whole process which you simply ignored. So for whatever reason, you seem to have an irrational commitment to your misconceptions about sigma, measuring steady-state noise and the veto, which it is clear that no amount of discussion can shift.

I didn't ignore it, but you did not address my main points. LIGO didn't ever actually eliminate environmental influences as being the actual cause of this signal. LIGO even got a "veto" of this specific signal within 18 seconds of uploading it for processing in fact, although the peer reviewers were never given that information to start with. The peer reviewers were erroneously told that there were no vetoes present within an hour of the event.

LIGO also refuses to even explain why that veto was originally added, and what it was expected to "detect" in the first place. They didn't explain how that specific veto or why this specific veto achieved a 'high confidence' rejection of this specific signal either, nor did they offer any mathematically quantified definition of the the term 'safety' with respect that veto removing what it was designed to remove in the first place.

But I am intrigued. I believe you are disussing this on other forums - aren't other people telling you the same things as I am? If multiple people are telling you the same thing, don't you think there might be something in it?

Well, you all seem to admit that the sigma figure is irrelevant and unrelated to cause. That doesn't help your case IMO, it destroys any claims of 'discovery'. Nobody has really acknowledged or admitted to the blatant *cherry picking* of data that went on. Nobody has even touched the blatant confirmation bias problem with a ten foot pole. LIGO claimed to have *eliminated* all other environmental factors based on a *lack* of confirmation in external equipment, but they gave their own claims a completely free pass! That's just a biased methodology from start to finish, which is most likely the main reason that LIGO has been completely unable to visually validate a single one of their claims. It's only going to get worse with three detectors online because they won't really have anymore logical excuses for *not* seeing anything, and not detecting any excess neutrino emissions either.

You don't find it the *least* bit suspicious that I'm already 3 for 3 in terms of predicting that they couldn't and wouldn't be able to visually verify their claims of a celestial origin of this signal, and LIGO is already 0 for 3 in terms of offering any external corroboration for even a single claim? If I go 10 for 10, and they go 0 for 10 would that change your mind?

You act as though I'm personally obligated to agree with you simply because you offer me a couple of paragraphs that don't even address my primary concerns, but can you even tell me what would cause you personally to "doubt" LIGO's claims that these signals are celestial in origin. How does that work?

What happens later this summer if only two of the eventual three detectors sees a similar "chirp" signal, but it's not seen by the third detector? What if they can't find a visual even with a well triangulated and reasonably small search area? Would that cause you to "doubt" their claim? What would it take to sway you since you seem completely unconcerned about their blatant confirmation bias problem, and their lack of any external corroboration for any of their several claims?

[Higgsy]

If they left in the ordinary environmental noise and still had a five sigma confidence figure, I might be impressed. As it stands that sigma figure means absolutely nothing in terms of supporting their claim as to cause.

I am not going to get drawn into discussing this with you for the reasons that I have explained in my previous post. I'll just point out that your statements above illustrate the problem that you have with listening, thinking about and understanding what people are explaining to you, and the reason why discussing it with you is pointless.

The idea that any competent physicist would include transient signals in the characterisation of steady state random noise is beyond ridiculous. Why in the name of Beelzebub would anyone do anything so fabulously stupid? I have told you times beyond count that the sigma figure relates to the confidence that the signal under investigation is a signal and does not arise from random steady state noise. Everybody in the world except you understands that. And there you go again criticising the sigma estimate becuase it doesn't support cause. But I have explained over and over again that the sigma does not relate to cause in the past, now and in the future in any experiment that any physicist has conducted, is conducting and will conduct; and yet you keep repeating the same utterly absurd complaint over and over again as if no-one had ever patiently explained to you what the sigma estimate means. Incorrigible.

[Michael Mozina]

If they left in the ordinary environmental noise and still had a five sigma confidence figure, I might be impressed. As it stands that sigma figure means absolutely nothing in terms of supporting their claim as to cause.

I am not going to get drawn into discussing this with you for the reasons that I have explained in my previous post. I'll just point out that your statements above illustrate the problem that you have with listening, thinking about and understanding what people are explaining to you, and the reason why discussing it with you is pointless.

You are evidently confusing the concept of "listening" to you with "agreeing" with you. They aren't the same concept. I hear what you're saying, but that only demonstrates that there is no evidence of a "discovery" of gravitational waves, just a "discovery' of a "signal" of unknown origin. I see *zero* evidence to support LIGO's claim as to the origin of that signal and I see no valid way for them to have eliminated ordinary environmental factors as the cause of that signal in that sigma calculation or any other way for that matter.

The idea that any competent physicist would include transient signals in the characterisation of steady state random noise is beyond ridiculous. Why in the name of Beelzebub would anyone do anything so fabulously stupid?

Since they're claiming to have ruled out all environmental noise, they should have included that environmental noise in their tests data. They didn't do that. They stripped out all environmental factors that hey could, and *pretended* to have eliminated environmental noise from further consideration.

I have told you times beyond count that the sigma figure relates to the confidence that the signal under investigation is a signal and does not arise from random steady state noise.

So what? It could easily arise from any number of *known environmental factors*.

Everybody in the world except you understands that.

Oh, I understand it too, and I also accept that what you're saying is true, which is exactly why I reject their claim of having "discovered' gravitational waves. They only 'discovered" a signal who's source they didn't identify with any confidence figure of any sort.

And there you go again criticising the sigma estimate becuase it doesn't support cause.

Nothing else in their paper does either. The only hypothetical evidence you seem to have offered is nothing more than a curve fitting exercise that also happens to match various blip transient events, and which could also be produced by a half dozen other physical processes for all I know.

But I have explained over and over again that the sigma does not relate to cause in the past, now and in the future in any experiment that any physicist has conducted, is conducting and will conduct; and yet you keep repeating the same utterly absurd complaint over and over again as if no-one had ever patiently explained to you what the sigma estimate means. Incorrigible.

If it can't be used to demonstrate any confidence in their claim as to *cause*, then what good is it? This isn't like particle physics where anything which deviates from the standard model automatically counts as a 'discovery". There's that small matter of ruling out everything else as a possible cause of the signal, and ruling in their own claim which LIGO simply did not do.

The five sigma number was fudged from the start. If they weren't gaming the system, they would have at least used all 48 days of "non-vetoed" time. They didn't even do that. They cherry picked only the "best" data in terms of getting that sigma figure over the 5 mark so they could call this a "discovery'. If it was only a 2 or 3 sigma figure, it would be able to be classified as a "discovery", so they gamed the system until get trumped up a 5+ sigma figure to pretend that it has some bearing on this being a 'discovery' of gravitational waves. It's an irrelevant number and an entirely irrelevant figure with respect to their own claim as to cause.

About the only thing we seem to agree on to this point in time is that the sigma figure doesn't demonstrate a damn thing to support their claim as to the cause of this signal. It doesn't even eliminate ordinary background noise as the cause of this signal.

[Cargo]

black holes are known ... experimentally to accelerate matter in their vicinity that does not fall into the event horizon.

Wait. They have experiments with Black Holes now?! Where is this Black Hole they experiment with and how did they create it?

[kell1990]

The fact of the matter is that in spite of a huge amount of press coverage the LIGO team doesn't really know what it has "discovered". To be sure they think they have found evidence of a pair of twirling twin black holes, but there is a huge problem with that: An earlier press release stated that a massive black hole (even if it existed) could have hurled our a planet the size of Neptune.

What's the matter with this guess? It is that black holes with their supposed mass (and gravitational attraction) could capture anything within its realm, including light itself. So how could such a body such as Neptune be "hurled out"? Not possible, according to the current definition of black holes.

I don't know what specific observation you are referring to here (perhaps the hypothesis that the supermassive BH at the galactic centre rips and ejects stars which the recoalesce into planet sized lumps?), but black holes are known theoretically and experimentally to accelerate matter in their vicinity that does not fall into the event horizon. Obviously, if it does fall in then it's lost forever, but not all matter accelerated by a BH falls in. Google black hole jets for more info.

Yes, well I understand all that. But the definition of a black hole seems to vary according to need here. Either it is an all-attracting object, or it is (most likely) a figment of someone's imagination.

The Law of Conservation of Mass and Energy is in play here. Nothing is "lost forever"; neither mass nor energy.

Whenever an object (a baryonic mass) approaches and is caught up in a Bennett pinch (or a Z pinch) it begins to be drawn into the pinch via a pair of twisted electromagnetic arms. As is plunges into the pinch it begins to be dissected, or more aptly, disassembled. The binding energies are released (which accounts for the radiation from the area) and as it descends into the pinch it is accelerated up to relativistic speed. Each "segment" (or layer) of the pinch contains the same amount of energy, and when it gets down to the tiniest level an enormous amount of energy is applied to the "baryons". They are stripped naked. Nothing but mass. This is what you think is a "black hole".

Then they pass through the pinch and begin to reassemble themselves via a charge component that they gather from having passed through the pinch.

The "quarks" (if that's the right word) become atoms; the atoms become molecules, and so on up the chain. Ad infinitum.