Higgsy wrote:Michael Mozina wrote:http://www.nbi.ku.dk/gravitational-waves/
FYI, It turns out that Ian Harry had a bug in his code which is why he wasn't originally able to produce the results of the paper that was critical toward LIGO. The authors of the paper that was critical to LIGO pointed out Ian's code problem.
That's true.
So in addition to the five LIGO methodology problems which I already pointed out, they have a sixth problem in their methodology as well. First they have to cherry pick only the "juiciest" data that gets them to the almighty five sigma figure, then they have to overlook the correlation of noise between the two detectors and *pretend* that only gravitational wave signals can be observed by both detectors within a 10ms window. The whole LIGO claim is one giant circular feedback loop that is based upon a 1/2 dozen *questionable* assumptions, all of which skew the results in their favor.
Why bother? The Danish group already looked at Harry's Python code and they already had to explain to him exactly which errors he made. I therefore know that Harry's original assessment was flat out wrong, and he's from LIGO. I already know that LIGO is as a prone to error as anyone else, and Harry's other commentary is based on false assertions that are based on buggy software routines.
That only makes me wonder all that much more about that "missing veto" which LIGO won't talk about, especially it's original *****intent******, it's "high confidence" rejection of the exact signal in question, and LIGO's complete lack of type of quantification with respect to any "unsafe veto".
Let's look first at 10: where he shows that correlations exist if the data is not properly whitened.
In other words, we already know for a fact that signals *can be* correlated between the detectors, so LIGO's oversimplified assertion that the correlation of any signal *requires* gravitational waves is falsified. LIGO has to go to *great* lengths to try to remove correlation from the detectors *before* we can even decide if one specific "noise" is correlated or not correlated between the detectors. The correlation claim can therefore *not* be used to exclude "blip transients" from further consideration because we can never be certain that we have eliminated 100 percent of the correlation of ordinary background noise between the two detectors, and the background noise is often *far greater* than the candidate signal.
The whole LIGO claim of "discovery" is based upon the *flimsiest* of possible arguments rather than being based upon any type of *visual* or neutrino confirmation of their claims!
This sure looks like Joseph Weber claims all over again. He claimed to have 311 "confirmations" before his results were overturned because they couldn't be replicated. It's deja vu all over again. This time around, the new 'Weber bars' cost hundreds of millions of dollars to build to test LIGO's claims, and for all I know they will *destroy* LIGO's claims rather than confirm them.
Then 11: shows that whitening the data removes most of the correlations although there is still some non-random phase in Livingstone as the whitening did not perfectly remove all tones.
In 13: he reproduces this behaviour with coloured Gaussian noise
In 15: he shows the actual data of the event with the signal and the whitened noise with the signal removed. With the signal the cross correlation is -200 and with the signal removed the residual maximum is -0.4 at 12.5ms with no min/max at 7ms.
In 16: he uses Creswell's signal conditioning with different bandpassing filters. Now there is a minimum in the residual data at 7ms of -12 (compared to the signal cross-correlation of -200), but expanding the time slide shows that there are other minima and maxima at other delays including a +13 at 130ms which is way outside the maximum propagation window. So the minimum at 7ms is unexceptional with this bandpass filter - it looks like a stochastic effect. The signal cross correlation is way higher than any cross correlation in time sliding the residual noise so even after making these corrections, it seems that there is no real problem with residual noise correlation at the event delay.
The Danish team went to great lengths in their video to point out the problems with "templates' being applied to different data and to explain why they went to great lengths to avoid them where possible. LIGO has to go to *ridiculous* lengths in order to "cherry pick" out only the very "best" data which they can find to support their claim, and they have to reject more than 2/3rds of the original data to get themselves to the all important 5 sigma figure. They then have to massage the hell out of the remaining data set using filters galore, and even then we have correlated noise problems which remain in the data.
LIGO then claims that three events which they can find *must* be caused by gravitational waves only because the noise happens to show correlation, when we know for a fact that noise can be correlated between the detectors, and *is* correlated between the detectors! This whole LIGO claim reeks of "bad science", "wishful thinking", and blatant confirmation bias, at every level of the process, including the way that LIGO *processed the data*!
Give me a break!
Without a visual or neutrino confirmation of their claim, this whole ''discovery" claim is just a textbook example of confirmation bias on a stick and it's a mirror image of the Joseph Weber fiasco.
Why did LIGO even start with BH-BH mergers if they were so difficult to visually confirm? Why not start with one of the other types of events which LIGO claimed they would be able to observe as well, which would be more straight forward to visually verify? Why? Because LIGO can't find any celestial event that actually ties back to any type of LIGO noise pattern, that's why. This "discovery" claim is like a "hail Mary" pass, with the express intent of circumventing any type of need for visual confirmation which is exactly why it's almost identical to Weber's bogus claims.
After reading the Danish team's assessment of the correlation problems, I have no confidence whatsoever that any one of these three specific signals in any way requires the introduction of gravitational waves to explain. Period.
They nailed LIGO to the wall (again).