MiniBooNE Results

I won’t bother to write up something about the background of today’s MiniBooNE results, since Tommaso Dorigo has already done a better job than I ever could. He also provides a link to where the live feed of the seminar will be, starting at 11am CDT. I’ll be in class at that time, but an hour later will try and attend the local seminar here at Columbia featuring Mike Shaevitz discussing the results.

And the result is….

No mu-neutrino to electron-neutrino oscillations of the sort that would explain the LSND result and require an extension of the Standard Model (beyond giving masses to the 3 known neutrinos). MiniBooNE was designed specifically to look for this, and has successfully ruled it out at 98% confidence level. They do see something anomalous in their data at low energy, but it is not compatible with being due to the kind of neutrino oscillations they were looking for. It’s also true that they just first got a look at this data two weeks ago, still have a lot of work to do to see if there is some sort of background contamination they hadn’t expected at these energies, or something they didn’t know about low energy cross-sections. Maybe it will take them a while to sort this out, but the bottom line is that what they were looking for is definitely not there.

Press release here, paper to come soon.

Update: For an excellent description of the result from Heather Ray, one of the MiniBooNE experimenters, see this guest posting at Cosmic Variance.


Warning: serious people should stop reading now, the rest is a low form of entertainment.

For something truly hilarious, you really should be following Lubos’s continually evolving misunderstandings of this experimental result, which he has taken as a reason for launching into another bizarre rant about me and Lee Smolin. As near as I can figure out, Lee and I are responsible for the misguided idea of designing an experiment like MiniBooNE to check into the possibility that LSND was seeing evidence of a sterile neutrino. His posting keeps changing (its URL is miniboone-confirms-lsnd, title now “Miniboone Refutes LSND”), with the early versions saying:

Evidence for several types of neutrino oscillations have been known for a decade or more. That includes atmospheric neutrino oscillations, solar neutrino oscillations, and a lab experiment called LSND in Los Alamos.

A simple oscillation in between two neutrino flavors – electron neutrino and muon neutrino – was a natural candidate to explain the observations but it couldn’t explain details of the LSND data which is why the LSND results were questioned. Another natural candidate was a two-flavor oscillation that includes a sterile neutrino, a new kind of neutrino without a charged partner.

Today, Fermilab’s MiniBooNE experiment has confirmed that the LSND results were correct and a more subtle explanation than the simple two-flavor oscillation is necessary. The result rules out the possibility that the observed oscillation is a two-flavor oscillation involving a new sterile neutrino. Their results indicate that there is something surprising that doesn’t fit the most obvious model.

He does seem to have more recently gotten a clue about this and noticed that it doesn’t confirm the LSND results, editing his posting and adding the standard obsessive rant. I see that in his previous posting (about a Harvard faculty meeting), according to him the proposed new Harvard core curriculum states that “All of science education must lead to increasing food production for the working class in the next 5 years [added to please Peter Woit]”.

It seems that I am not only determining the course of neutrino experiments, but also setting the Harvard core curriculum. My powers are truly immense…

Update: Lubos has now changed the file-name from “miniboone-confirms-lsnd.html” to miniboone-refutes-lsnd.html, and deleted the comments from people explaining to him that he was confused. The new version starts off with:

I have erased several comments that only increased the amount of confusion, changed the filename to break links from crackpots’ blogs, and hope that the text below is now more or less OK.

Last Updated on

This entry was posted in Uncategorized. Bookmark the permalink.

62 Responses to MiniBooNE Results

  1. Well, Peter, thank you: as they say around here “drink what you like, it’s on me”.

    I think it will be quite interesting to see if they kill their own upgrade or what…


  2. Shantanu says:

    Unfortunately I am getting a “User Limit Exceeded” message. looks like
    lots of people have connected.

  3. andy says:

    Nothing! (except for a weak anomaly at the lowest energy which they’re trying to hype.)

  4. John says:

    Lubos is getting a little flak in his comments I see. ha.

  5. Peter Woit says:


    Don’t worry, he has dealt with that problem in the usual way.

  6. Bob Jones says:


    I guess I am one of the less serious readers of your blog. And apparently you have REALLY pissed Lubos off. I only read his blog by linking from yours…and now he is blocking that with a hilarious message that reads:

    Sorry but I really don’t intend to share readers with that particular aggressive liar, parasite, and crackpot. If you find it appropriate to read Not Even Wrong, you’re just not welcome on The Reference Frame. This type of direct links won’t work. I’ve had huge problems with the scum pumped onto my blog by the Not Even Wrong website. I apologize if you’re not a part of it but there’s no way to distinguish.

    You can open if you wish.


    Wow!! Thanks as usual for the useful information and sometimes entertainment!!!

  7. gunpowder&noodles says:

    Still, I’d rather read LM’s [current] summary than Heather Ray’s hilariously anticlimactic and infinitely tedious explanation of the fact that they didn’t find *anything*…..

  8. Heinrich says:

    This blog is not totally correct. True, miniBooNE ruled out standard neutrino oscillations as a solution for the LSND anomaly, but that possibility was already at the verge of being ruled out by combinations of results from solar, atmospheric, reactor and collider neutrino experiments. On the other hand we proposed a spectacular solution to the LSND anomaly involving neutrino shortcuts in extra dimensions, published in Phys.Rev.D72:095017,2005 [hep-ph/0504096]. If you look at Fig. 5 in this paper, you will see that for a choice of the resonant energy in the region 200-300 MeV we not only predicted the small counting rate for electron neutrino events above 475 MeV, but also the large rates in the 300-475 MeV region!
    While the anomalous effect seen by miniBooNE might have a conventional explanation, it might well be the first hint for extra dimensions of spacetime!

  9. mclaren says:

    Like most people, as soon as I encounter the word “Lubos” or “motl” I stop reading.

    Clearly something odd is going on with neutrinos. Does anyone have any take on whether this new result rules out neutrinos as the putative source of the missing dark matter? I.e., as I understand it, neutrino oscillation of the standard kind requires that neutrinos have mass, which in turn implies that they might be responsible for some or all of the missing matter.

    Does this new result, in short, imply anything about neutrino mass?

  10. Kea says:


    No, it doesn’t rule out DM sterile neutrinos that appear in weird CP sectors, as Dharam points out.

  11. The neutrino energies of LSND and MiniBoone are 60-200 MeV and
    300-1500 MeV and oscillations in LSND mass range are found to be absent above 500 MeV. Evidence for oscillations is found below 500 MeV.

    Hence LSND and MiniBooNE are consistent if one accepts that neutrino mass scale depends on its energy as TGD strongly suggests. For details see the posting at
    my blog page.

    Matti Pitkanen

  12. Kris Krogh says:

    We’ll also hear the preliminary results from another major experiment this week. Those from Gravity Probe B will be announced Saturday at the meeting of the American Physical Society in Florida, preceded by a NASA press/media event. (My prediction: general relativity fails.)

  13. woit says:

    mclaren (and others),

    Remember, this is a null-result. All that happened is that this experiment was looking for something specific seen by LSND inconsistent with the standard picture of three massive neutrinos (the masses and abundances are considered too small to explain dark matter), and they showed it’s not there. You can’t claim this as evidence for your favorite exotic theory. The only anomaly they see is the low energy one, and this is based on data they just first looked at two weeks ago, and have not had the time to properly evaluate. They are in no way claiming this as a reliable measurement, their only claim is to have ruled out the supposed oscillations.

  14. Thomas Larsson says:

    OT: A review of NEW and TWP by Martin Gardner can be found here. He ends by quoting Glashow’s poem.

  15. tomj says:

    So if it rules out the inconsistent LSDN result, that is pretty important science, isn’t it?

    Does this mean that the LSDN result is no longer considered a valid result, like oops, never mind?

  16. Peter Woit says:


    Yes, the importance here is ruling out LSND, specifically the idea that it was seeing neutrino oscillations. The experiments were not the same, so you can still try and come up with something consistent with both LSND and MiniBoone, but that is very hard.

  17. Thomas Love says:

    The two experiments, LSND and MiniBoone, yielded inconsistent results. Doesn’t that mean we need a third experiment to see which was correct? It doesn’t seem right to accept MiniBoone uncritically just because it is more recent.

  18. Bee says:

    but didn’t LSND use anti-neutrinos where MiniBoone has neutrinos? I mean, yes, MiniBoone rules out the sterile neutrino scenario (which, as Heinrich pointed out was pretty much ruled out by the SNO neutral current data anyhow), but if one argues there might be an asymmetry between nu/anti-nu this does neither confirm nor refute LSND?

  19. Bee says:

    Thomas: the point is in the 3-flavor scenario LSND is inconsistent with the other neutrino data (lots of: solar, atmospheric, reactor). You need to come up with some ad hoc additional explanation to incorporate LSND (like e.g. sterile neutrinos) if you believe their data analysis is correct (LSND data is commonly excluded for global fits of neutrino osci). MiniBoone has re-checked the LSND energy sector and did not find any evidence against the standard 3-flavor scenario and its parameters (mix. angles, \Delta m^2).

  20. Peter Woit says:


    Yes, in principle anti-neutrinos could behave differently, and I gather MiniBoone is now running with anti-neutrinos, but I don’t know if they will have the statistics to directly test LSND. Also, I don’t know how hard it is to construct models where antineutrinos would oscillate but not neutrinos, without violating CPT or running into contradiction with other experiments.

  21. tomj says:

    What is the importance of the results being collected in this tamper proof box? Is this just a metaphor for computer data that nobody can see while the experiment is being run?

  22. Peter Woit says:


    It’s not that it’s tamper-proof. The idea of a “blind” analysis like this is to try and do the data analysis while not knowing how the choices one makes affect the final result, to keep oneself from subconsciously skewing one’s choices to get a certain result. The problem with a blind analysis is that you may miss problems with the data that only show up in the final result, not in the partial results you allowed yourself to look at. This may be the cause of the low-energy anomalies they found in their final result.

  23. Brett says:

    For neutrinos and antineutrinos to oscillate differently, you need CPT violation, and you can’t break CPT very easily. To have CPT violation, you need to give up Lorentz symmetry, or locality, or unitarity, or energy positivity, or something equally basic. You can’t just resolve the discrepancy between LSND and MiniBooNE by introducing a “CPT-violating mass”; no such operator exists. You instead have to introduce a neutrino mixing interaction that violates one of the conditions that I mentioned, and–very importantly–these kinds of exotic operators will not give oscillations with the same L/E dependence that is seen in mass oscillations. For example, with Lorentz violation, it would be quite possible to have oscillations only at lower energies (where they were seen by LSND) without actually violating CPT, and thus have the oscillations in both neutrinos and antineutrinos.

  24. M says:

    even LSND interpretations based on sterile neutrino(s) were already significantly incompatible with bounds from other neutrino experiments: this is why viable models for LSND already involved ingredients such as extra dimensional sterile neutrinos with CPT-violating varying masses. Using these tools, it is probably possible to write down models that fit both the anomalies in MiniBoone and LSND.

    About the part of the post that serious readers should not read, I signalled to Lubos that what he wrote in the comments to his post contains mistakes in both neutrino physics and string theory. But instead of censoring his comments, he preferred to censor mine.

  25. H-bar says:

    I get forwarded to this site from the link to Lubos’ posting:

  26. Wolf-gang says:

    by introducing a “CPT-violating mass”; no such operator exists.

    I think Gabriela Barenboim has found one though I don’t have the arXiv reference handy now.

  27. Bee says:

    now how did that happen? I certainly haven’t said I want to break CPT to explain LSND, it doesn’t even seem to me remotely appealing. I was just wondering why an experiment set up to clarify LSND wasn’t as close to the LSND experiment as possible to begin with.

  28. Ari Heikkinen says:

    Heh, that bit in your “entertainment” section nicely proves that people who think are experts at everything (no one is) will sooner or later end up making themselves look like fools.

  29. M says:

    Bee, because LSND and MiniBoone have many persons in common, so they cannot confirm LSND by repeating it. This is why they did a blind analysis and moved from neutrinos to higher energy anti-neutrinos, that are detected in a different way.

    The present unclear situations arises because MiniBoone was planned to confirm or exclude LSND oscillations, but in the meantime the oscillations suggested by LSND had been indirectly disfavored by other experiments, and more exotic interpretations have been proposed.

    For example, one proposal was oscillations among active neutrinos (without any sterile neutrino) with CPT-violating masses, but this was indirectly excluded by KamLAND. At this point both CPT-violating masses and sterile neutrinos were considered socially acceptable, so that desperate theorists could move to sterile neutrinos with CPT-violating masses…

  30. Thomas Love says:

    Brett Says:

    “For neutrinos and antineutrinos to oscillate differently, you need CPT violation…”

    But we live in a sea of neutrinos. We do not live in a sea of antineutrinos. It would be very possible “For neutrinos and antineutrinos to oscillate differently” if that oscillation is due to the interaction with the background sea.

    We do not need CPT violation.

  31. Brett says:

    Barenboim’s model is indeed CPT-violating, with a term that can be cast in a form that looks like a mass. It gets around the CPT theorem by being nonlocal, something which tends to cause a lot of problems when nontrivial interactions come into play. (I don’t claim that this particular model necessarily suffers from those problems though; I myself have worked on similar Hilbert-transform-type nonlocal models that can circumvent some of the usual difficulties.)

    I don’t think this theory has a mass term in the usual sense though. The theory has non-canonical commutation relations for the field, which is problematic for a particle interpretation. More generally, in theories that violate one or more of the conditions that ensure CPT, different notions of the mass (which are equivalent in ordinary physics) may diverge. The inertial mass, gravitational mass, rest energy, and location of the pole in the propagator will not generally coincide.

  32. Bee says:

    Bee, because LSND and MiniBoone have many persons in common, so they cannot confirm LSND by repeating it.

    This seems to me like a rather weird reason. If its a scientifically performed experiment it shouldn’t depend on the persons.

    so that desperate theorists could move to sterile neutrinos with CPT-violating masses…

    That’s what I am afraid going to happen. Everybody whose model is in danger now will jump and add some extra parameters, try to get published before also this window closes …

  33. I would still emphasize that the neutrino energy ranges in LSND and MiniBoone 60-200 MeV and 300-1500 MeV. Low energy in MiniBoone means something still above LSND and at low energies the effect was found. For some reason this point has not been emphasized explicitly in blog discussions and has not prevented people from drawing quite strong conclusions in most blog discussions.

    To me the result says that the two experiments are consistent if one accepts the possibility that neutrino mass depends on neutrino energy measured with respect to rest frame defined by laboratory. This combined with relativistic invariance of course means that neutrinos must have rather delicatate kind of interaction with matter. For my own proposal see my blog.

  34. Marco says:

    I think the point in not repeating the exact same experiment is that you want to have different systematics. The important quantity that governs oscillations is L/E (= baseline of the neutrino beam / energy of the neutrinos). The strategy for MiniBooNE was to increase both L and E with respect to LSND, while keeping a similar L/E. At a different Energy, one has different cross sections, different backgrounds etc etc: in short, different possible sources of systematic errors. But since the L/E was similar, if the LSND result had been due to (oscillation) physics, it should have popped up. Which it did not.

    As for more “extreme” theory models: well, what M was saying in the comment above is that CPT violation + sterile neutrinos has already been put on the market (hep-ph/0308299), because CPT violation alone or sterile neutrinos alone were already in trouble before MiniBooNE. It’s up to your taste to like it or not.

    Now let’s see what these “desperate theorists” (as you said) will find to explain that little excess at low energies in MiniBooNE. For the moment, hats off to Heinrich’s model (see his comment above)! It’s certainly a bit of an extreme model, but they had predicted something like this sort of excess. I think he is celebrating with some good Alabama whiskey…

  35. Bee says:

    Hi Marco,

    🙂 I’ve played around some years with extra dimensional neutrino models, and believe me, some of them were far more extreme. These models with mass varying neutrinos – well, interesting, but I admittedly find it very implausible. But yes, I see. I guess the intention was to check several points in which LSND was different but not mingling them all together?

    Hi Matti,

    thanks for pointing out that the energy range was different as well, I must have missed that. I would suspect though that if there really is an effect in this case it’s less a neutrino effect, as a data analysis problem. As Marco mentions, for analyzing the detected neutrinos you need 3 factors: production (how many start), oscillation (where do they go), and detection (cross-section). I can’t recall all the details but at some point I tried to figure out how the cross section for neutrinos are modeled in the monte-carlo analysis and found some are based on really poor data. Best,


  36. Bee says:

    I mean, look at the available data points in the

  37. Bee says:

    sorry, I always forget that HTML doesn’t like ‘smaller than’ …

    What I meant to say, look at the available data points in the region smaller than .5 GeV

    And this is for neutrinos. For anti-neutrinos it’s even worse.

    (Peter, feel free to delete the previous comment)

  38. Marco says:

    Hi Bee,
    first of all, I have also played with neutrinos in extradimensions, but my models were perfectly natural and incredibly compelling. 😉

    But since we are on the topic, here is my partial list of things proposed for LSND, and we could rate them on a scale of “extreme” or “plausible”: it could be an interesting exercise in the sociology of science. I am not sure it is very scientific though. (I don’t cite references but of course I can provide them. And if I’m forgetting some category I’d like to know.)

    – the well known sterile neutrinos with about 1 eV mass (but they were disfavored by a number of things even before MiniBooNE, among which cosmology)
    – Heinrich’s model: sterile neutrinos + one extradimension with hawaiian waves (how does this score??), still it predicted the right thing for MiniBooNE.
    – models with CPT violation in the usual 3 neutrinos: disfavored, so people moved to CPTv+sterile
    – on another side: models with Mass Varying neutrinos (their mass varies according to the matter around them, or according to the density of other neutrinos), both for LSND and for a connection to Dark Energy
    – and now: can you guess what’s coming? yes, people also proposed Mass Varying Sterile Neutrinos

    So you see you definitely need to work as soon as possible on a model of Mass Varying Sterile CPTviolating ExtraDimensional neutrinos, but maybe our friend M above is already on it.

    More seriously, I think this shows that the LSND anomaly was really hard to reconcile with a number of things. So it’s good that MiniBooNE settled the issue. Unless, of course, we have now something anomalous again from MiniBooNE itself…

  39. Bee says:

    Hi Marco,

    but my models were perfectly natural and incredibly compelling yeah, yeah, that’s what I say in my talks as well 😉 . Regarding your partial list, I would add the above mentioned uncertainties for the cross-sections. I mean, one could go so far to ask if they have actually measured the oscillation with knowing the cross-section or vice versa. I am generally sceptic about having only one extra dimension, not to mention that this most often doesn’t really explain anything. It just gives you another parameter. For all points the question is: why would such an effect only be important for the neutrinos? Best,


  40. Marco says:

    Hi Bee,
    I tend to agree that neutrino cross sections (and their energy dependance) are a delicate point. I think however that the MiniBooNE people have been very careful on this, using data from other experiments etc. Still, if you find some flaw let us know.

    If I understood correctly, your other comment is that it looks a bit artificial that people invoke all those exotic things always for neutrinos and not for other particles. I.e.: why should neutrinos always be the odd ones in town?

    Well, yes, but I think that in reality there is a different reason for any one of the oddities invoked. Some are more convincing, some are less.
    For example, when one says “sterile neutrino” one really is taking a shortcut to say: one of all the possible “spin 1/2 fermions that are neutral under the SM SU(2)L x U(1)Y gauge group”; and since there are many of these on the market (from SuSy, from additional simmetries, even from strings!…) a sterile neutrino is not toooo exotic. If in addition it is light for some reason, the mixing with the ordinary neutrinos is likely, so here we are.
    For the case of extra-dimensions, you know, once you have bought a large extra-dimension you should make the most of it and put in it everything that can fit, so the graviton and also a sterile neutrino, because they are not tied to our brane by SM forces. So here we are again having neutrinos play the exotic.
    For the case of Mass Varying Neutrino models, I think the motivation was the observation that the density of Dark Energy in the Universe today is similar to the energy density of massive neutrinos today, so DE and neutrinos could be coupled. So again neutrinos end up being singled out.
    And so on.

    But of course the underlying reason in all this might be (more mundanely) that neutrinos are the least tested of the Standard Model particles, so we still have freedom to play.
    Sorry if I talked too much.
    Best, Marco

  41. Bee says:

    Sorry if I talked too much. No problem. You can’t beat me. I managed to give a talk today which I though would take one hour, but I wasn’t finished after two hours.

    Regarding extra dimensions and neutrinos, what I found intriguing about this approach is that it explains why the neutrino IS the odd one out. The reason being that a right handed neutrino (if there is one) does not carry any charges of the standard model (electrically neutral, electroweak singlet), and therefore – like the graviton – is not bound to the brane. One doesn’t have to invent another sterile (left handed) neutrino (though one can of course, one can always make things more difficult).

    But yes, it always seems to me kind of odd to include a new feature but then only include it for part of the particles. It would seem more natural to assume it exists for all particles (but it might be only observable for some).

    Regarding the cross-section, I am sure the MiniBoone people have been careful! It is just not clear to me what the errorbar on that cross-section is (its a line! can it be there is no errorbar to it?). The point I was trying to make is if I look at the data points it seems to me there aren’t so really many (in fact, none) in the relevant energy region. If I look at the paper from which the curve comes, it is not clear to me how large the theoretical uncertainty is (there is a lot of talking about Method I and II, and I admit I didn’t look into to details), and whether this uncertainty can be fixed with the data points in the higher energy region? I’m not saying this is a flaw, this is just something I don’t understand.



  42. CG says:

    Sorry to cut in, but I have a rather naive question which seems somewhat tangential to your discussion, though theoretically based, rather than experimental. I was really wondering what the basic arguments were in favour of only having 3 flavours to each of the particle families (eg neutrinos) in the Standard Model. From a lay-scientist’s point of view (mine), it does not seem at all obvious why this should be the case. i.e, why not 4? 5? 6? infinity?


  43. woit says:


    In the standard model, particles come in generations, with one neutrino in each generation. We’ve seen all the particles in three generations, no evidence at all for a 4th generation, although in principle there could be one (or more). In addition, the observed decay width of the Z particle implies that there cannot be more than three low mass neutrinos. If there is a 4th generation, it has to be unlike the others, having a neutrino of a much much higher mass than the other 3.

  44. Poor Lubos,

    I worry that he is losing touch with the real world. You would think Harvard could afford to find medical help for whatever it is that’s really bugging him. I fear that a rather brilliant mind is going largely to waste.

  45. The L/E for the MiniBoone at 475 MeV is nearly coincident where KARMEN data starts, while LSND’s L/E range has an overlap with that very line for 2nu oscillation analysis (as chosen by MIniBoone).

    The KARMEN, LSND, and MiniBoone data are hard to understand in the context of any oscillation analysis (irrespective of any mismatch between the CP properties of the SM and the sterile mass eigenstates). Indeed, neutrino sterility should come from some new physics; and one has a big playing field for that. But, in the absence of a constraining principle, or strong experimental data, such a game is at best a good mathematical science fiction.

  46. Thomas Larsson says:

    The MiniBooNE paper is now available on the arxiv.

  47. island says:

    CIP said:
    I worry that he is losing touch with the real world. You would think Harvard could afford to find medical help for whatever it is that’s really bugging him. I fear that a rather brilliant mind is going largely to waste.

    Lumo’s behavior is highly typical of extremists and fanatics who grasp at any straw that they can reach in order to rationalize their belief system in the face of reality. He initially attempted to do exactly that when he first heard them talking about the low energy descrepancy and “other exciting physics”.

    I would expect to find lubos trying to vindicate string theory long after it’s dead and gone to the rest of the world.

  48. CG:

    As far as the math is concerned, the Standard Model can accomodate any number of particle families, and hence any number of neutrinos. (Note: only one neutrino per family.)

    There are two independent ways we know there are only 3 neutrinos in nature. One is the decay rate of the Z0 particle. Since the Z0 can decay into a neutrino and an antineutrino, the more families there are, the faster the Z0 must decay. The decay rate has been measured and shows there are 3 families.

    Secondly, in the first millisecond after the Big Bang, extra neutrinos would affect the amount of hydrogen, deuterium, and helium that gets produced. Astrophysicists have measured the relative amounts of these in intergalactic space, so they can constrain the number of neutrino species out there. Their results also give 3 families.

Comments are closed.