Superluminal neutrinos, jokes aside
It has been a lot of fun following the reactions to the OPERA release https://twitter.com/#!/search?q=%23mundaneneutrinoexplanations
Update: People come here searching for “neutrino jokes”. So in addition to the above twitter link, may I recommend some fun anagrams: fast neutrinos – “sensation turf”, “a funniest sort”, and more.
The funniest thing to me is that this is an experimental discovery with an extraordinarily high level of statistical certainty (6 sigma!). But people joke about it. Even the media hype is unusually tactful. I am too very skeptical about the result, but I decided to take the time and do the homework.
I now understand some details of the experiment setup, and realize that there are some erroneous claims in the media. Please keep in mind that I have no formal training in physics whatsoever
I read the official paper by the OPERA collaboration http://static.arxiv.org/pdf/1109.4897.pdf . Parts of it are surprisingly accessible to a layperson like me. I didn’t understand many things. The major things that I didn’t understand are: 1) how exactly GPS was used, and 2) why they averaged the data for all the sessions (“spills”) in their analysis – couldn’t this lead to a loss of potentially significant information? (more on this below).
Btw, I mentioned the OPERA experiment before: https://bbzippo.wordpress.com/2010/06/02/neutrino-oscillation-directly-observed/ and mentioned that Neutrino is an anagram of “inner out”
Now, my “findings”:
Media claims that the measurement was repeated over 9000 16,000 times. No. 16,000 is the total number of neutrinos caught over the course of the experiment (3 years). From what I gathered, all measurements that really matter – the precise measurements of distances and clock synchronization – were conducted maybe just a couple of times.
Media claims that the experiment challenges the theory of relativity. No. The theory does not prohibit superluminal speeds. See http://en.wikipedia.org/wiki/Tachyon .
So why the skepticism? Simple. The experiment does not agree with other experiments. In particular, neutrino detection from supernovas tell us that the difference between the speed of light and the speed of neutrinos is way smaller than that measured by OPERA. And that was measured after they had travelled thousands of light years, versus 500 miles. There are other reasons for doubts of course. E.g. the current established theories don’t favor tachyons, even though they are not prohibited by relativity.
What caught my eye were Fig. 4, Fig. 9 and Fig. 11 from the paper.
Fig. 4 are the protons at the source. The neutrinos at the detector (Fig. 11) have a corresponding “waveform”. And obviously, to find out when the particular neutrino was born we need to match the detected waveform to the source waveform. Obviously, to have an accurate match one needs to rely on steep slopes. The leading and the trailing edge are steep enough, all right. But the “five-peak structure” from Fig.4 gets averaged out after the summation (Fig. 9). I wonder, could they achieve better accuracy had they analyzed individual sessions without averaging? I understand that would be much more work, and I realize that my suggestion might be complete nonsense since I have no clue about this stuff, but still…
Also Fig. 11 is very curious because it shows that before some corrections related to clock synchronization were applied, the neutrinos beat the light by a whole microsecond! And with the correction applied, it became only 60 nanoseconds. It is remarkable that the time corrections and the distributions were analyzed “blindly”: by independent groups of people. That’s cool, but I don’t see how this blind analysis would help eliminate any confirmation bias.