Header 1

Our future, our universe, and other weighty topics

Friday, February 19, 2016

LIGO Doubts Will Persist Unless Replication Occurs

Last week the billion-dollar LIGO project announced the discovery of gravitational waves. It is interesting to examine the very absurd double standard followed by mainstream science when it comes to instantly accepting announcements such as this (based on a single rather questionable observational occurrence) while rejecting evidence for phenomena that is based on decades of experiments.

First, let's look at the LIGO observational event reported last week, and some reasons why it may not be so bulletproof. The scientific paper is here. The observational event occurred on September 14, 2015 at 9:50:45. It lasted only a tenth of a second. Two LIGO observatories recorded an ultra-faint spike at this instant, one in Hanford, Washington, and another in Livingston, Lousiana. The signal was so short-lived and faint that it is best described as a momentary micro-blip.

But there are several reasons for being skeptical about the claim that both observatories observed a pair of distant black holes merging. For one thing, the signal in Louisiana apparently occurred while there was no one in the control room of the observatory. Two scientists had left an hour ago, as reported in this press account. This raises security questions about whether someone could have been messing around in the control room, unobserved.

We actually know that sometimes LIGO would have false signals “injected” into it, by something called the “blind injection group,” but we are assured that this was not such a false signal. But are we sure? The New Yorker reports that the scientists investigated whether one of their colleagues could have faked the signal:

Reitze, Weiss, Gonz├ílez, and a handful of others considered who, if anyone, was familiar enough with both the apparatus and the algorithms to have spoofed the system and covered his or her tracks. There were only four candidates, and none of them had a plausible motive. “We grilled those guys,” Weiss said. “And no, they didn’t do it.” Ultimately, he said, “We accepted that the most economical explanation was that it really is a black-hole pair.”

That is a little less than airtight. Apparently at least four people could have faked the signal, but we are asked to dismiss such a possibility solely on the rather questionable assertion that “none of them had a plausible motive” and the fact that they were questioned.

We should also ask: could other people have faked the signal, other than these four? We would expect such a possibility if the data was stored online, in database servers that could be hacked. Since the reported observation event covered less than a second, it would seem to be an easy job to fake such a thing. If a standard SQL database was used to store the data, someone would merely need to create a short SQL script with a few lines of Update statements. Once you login to a remote server using Telnet or some other utility, you would merely need to login to the database and run the script. Repeat the process on a second server, and you've done everything you need to fake things. Some foreign hacker might have done that from his living room. How can we be confident that such a thing didn't occur? American corporations suffer from all kinds of weird hacking incidents from other countries.

According to this account, an expensive gravitational wave observatory has just been approved in India, because of the observations reported by LIGO. So conceivably foreigners may have had a motive for hacking.

There are other possibilities that don't involve fraud. One possibility is that the incredibly sensitive equipment simply didn't work quite right. That seems all too possible. At this link the equipment is described: “The interferometers consist of suspended mirrors, which reflect the laser beams which are used to sense tiny mirror motions — 1/10,000th the diameter of a proton — caused by the passage of a gravitational wave.” Why should we have much confidence in something so sensitive, measuring something so microscopic that we should doubt the ability of anyone to accurately measure it? These are perhaps the most sensitive measurements that have ever been attempted.

I may also note that the event occurred at the very beginning of a major upgrade. Every software manager knows that new releases or major upgrades are often the source of various types of bugs.

Another possibility is that an earthly signal was detected. The scientific paper claims that such a possibility was considered: “Exhaustive investigations of instrumental and environmental disturbances were performed, giving no evidence to suggest that GW150914 could be an instrumental artifact.” That doesn't sound airtight at all, and the scientific paper gives very few details of how such an investigation occurred. The paper says that they looked for an earthly source that could have produced the signal, and didn't find one. How does that rule out such a possibility? They could have overlooked some earthly source that produced the signal. Since the LIGO observatories are kilometers long, we should expect that they should be sensitive to seismographic events originating from our planet itself.

There are other explanations that could explain the data without gravitational waves being involved. A software error or hardware error could explain the data. There could be a bug somewhere in the very complicated LIGO software that produced a false alarm, due to some programming error.

As for the very brief match of the signals between the two observatories, they are not an exact match, although they look about 95% similar.  Below is the graph from the scientific paper, showing observations from the different observatories during a time of less than a second.  Even though the data from the Hanford observatory has been "shifted" and "inverted" (possibly to make it look like a close match to the data from the Livingston observatory), there is not an exact match between the lines.

Such a match could have been found from a “data mining” database query looking for a match. If you give me two random sources of data (such as stock prices in 2006 and bond prices in 2015), collected continuously over months, then with the right SQL query I will probably be able to find some tiny time slice where the data matches up, purely by coincidence, giving a false impression that the same thing was being observed. 

There is another reason for doubt. A scientific paper co-authored by a huge team of scientists estimated that a project such as LIGO should produce 40 detections of gravitational waves each year, from “compact binary coalescence sources.” But so far the LIGO project has announced only one detection, the event of September 14, 2015. Where are all the other such events, that are supposed to be happening almost once a week? This may cast doubt on the LIGO announcement, until such time as other observations are made.

There are two parts of the LIGO announcement. The first is the claim that gravitational waves were detected (which as we have seen is on rather shaky ground). The second is the claim that these waves were caused by a merger of distant black holes. The second claim is speculative, and is not well supported by the evidence.

The scientists had no direct evidence that the claimed signal came from a distant black hole. What they basically did is to do some calculations showing a hypothetical scenario by which a black hole merger might have produced the described signal. But that is not at all the same as showing that such a hypothetical scenario was the actual cause. Given a gravitational wave observation, there are always many possible ways of explaining it astronomically. We are reminded here of the BICEP2 affair, when scientists triumphantly claimed that the signals they detected came from the dawn of time. It was later shown that just such a signal could have been produced by dust. There was no way of even telling from which direction the LIGO signal was coming, so the scientists did nothing to show that the signal came from an exact spot in the sky where black holes are known to exist.

You would think that facts such as these would cause our scientists to be cautious. Science is supposed to require repeated observations, not one-shot wonders. But the scientific community has thrown caution to the wind in this matter. Based on a single rather questionable observational event, the scientific community has acted as if LIGO is proof of gravitational waves and proof of a black hole merger. The first claim is shaky, and the second claim is extremely shaky.

We can contrast this with the situation in regard to evidence for ESP (extrasensory perception). Gathered for over a century, the observational evidence for ESP is currently vastly greater than the one-second LIGO evidence. The Society for Psychical Research and other organizations started publishing experimental data as early as the 19th century. At universities such as Duke University, researchers such as Joseph Rhine spent many years doing experimental research that repeatedly produced spectacular successes, such as the extremely convincing Rhine-Pearce experiments and the equally convincing Pearce-Pratt experiments discussed here, getting results with a chance likelihood of about 1 in 10 trillion. Even more compelling (as described here) was the 73% accuracy rate recorded by a professor at Hunter College, with a person located at another location, making 1850 guesses that should have produced by chance a success rate of only 20%. Also very compelling are ESP tests in recent decades using sensory deprivation studies involving a technique called the ganzfeld technique, and some recent tests with ESP and autistic children, as reported here.

But have our mainstream scientists accepted these results, decades of convincing evidence? No. They keep demanding that more airtight tests be done, no matter how airtight is the evidence. But the same mainstream scientists will instantly accept some “gravitational wave” finding based on a single questionable observational event lasting a tenth of a second. It's a ridiculous case of a double standard. When our mainstream scientists have something that they want to believe in, they seem to have an extremely lenient standard. But when they have something they don't want to believe in, they adopt an exclusionary standard a thousand times more stringent.

It's hard to imagine a more outrageous double standard: if it's something our scientists don't want to believe in, then a hundred years of compelling observations are dismissed as “no evidence”; but if it's something they do want to believe in, then a single questionable observation event lasting a tenth of a second is counted as conclusive proof.

We can imagine if a country club operated in a similar way. It might work like this. If a white person tried to enter the front door of the club, the person guarding the door would simply ask “Are you a member?” and would let in the white person if he answered “Yes.” But if a person of color tried to enter the door, the guard would demand to see a birth certificate, a driver's license, a Social Security card, a work photo ID, and a college diploma. And when all those were produced, the person of color would still not be admitted, on the grounds that there was still doubt about his identity.