The Higgs Discovery

Just got out of 8 days in the Grand Canyon which was spectacular,

Reliable rumors couldn’t wait, and they indicate that the experiments are seeing much the same thing as last year in this year’s new data: strong hints of a Higgs around 125 GeV. The main channel investigated is the gamma-gamma channel where they are each seeing about a 4 sigma signal.

More later when I reach civilization.

Update: Back in civilization, or at least New York City. The above was the first posting I’ve ever written on an iphone, late at night. Now I have a real keyboard, so I can write a bit more. The “4 sigma signal” refers to the combined 2011 and new 2012 data. To oversimplify the situation, last year both experiments were seeing roughly a 3 sigma excess in gamma-gamma around 125 GeV. This was enough to convince many people that it was highly likely that this was the Higgs. However, that size excess is not completely convincing, it is not unheard of for there to be statistical flukes of such size.

The 2012 data that is being analyzed for ICHEP is of a similar size to the 2011 data. If 2011 was a fluke, you expect to see nothing much around 125 GeV in the 2012 data. If the 2011 signal really was the Higgs you expect the signal to strengthen. What I’m hearing from both experiments is that they are seeing an excess in the new data, strengthening the significance of the signal.

Exactly how much data they’ll have analyzed by ICHEP and exactly what the significance of the signal in the gamma-gamma channel will be (as well as what other channels will show) is still to be seen. CERN will soon have to decide how to spin this: will they announce discovery of the Higgs, or will they wait for some overwhelmingly convincing standard to be met, such as 5 sigma in at least one channel of one experiment? The bottom line though is now clear: there’s something there which looks like a Higgs is supposed to look. Attention will soon move to seeing if this signal is exactly what the SM predicts (e.g. will the excesses in different channels agree with SM predictions?).

More details about this from Philip Gibbs (who is speculating about what will be announced), and from Tommaso Dorigo (who is keeping quiet about what he knows, but providing context for what the ICHEP announcements will mean).

Update: Matt Strassler has more about this here. He provides about 20 links to his own blog, no link to the source of his information (this posting). It appears that this is because I’m a “non-particle-physicist blogger” engaged in a conspiratorial plot with some of the 6000+ people who know this latest news to “subvert the scientific process” by sharing it with others.

Update: There are stories about this at Wired, New Scientist and the New York Times. The New York Times article emphasizes that the Higgs results are now “Shrouded in Secrecy”, with the spokeswoman for ATLAS pleading “Please do not believe the blogs”.

According to Matt Strassler “the experimentalists can’t possibly have their data in presentable form yet, so the rumors can’t be correct in every detail”. To clarify any confusion

“Exactly how much data they’ll have analyzed by ICHEP and exactly what the significance of the signal in the gamma-gamma channel will be (as well as what other channels will show) is still to be seen”

means that the above rumors were based on just part of the data (significantly less than half in the ATLAS case, somewhat more than half in the CMS case).

Update: I think I’m too old to ever really understand Twitter, but it seems that #HiggsRumors is a “Trending Topic”, whatever that means. More explanation available from Jennifer Ouellette, and sensible commentary from Chad Orzel.

This entry was posted in Experimental HEP News, Favorite Old Posts. Bookmark the permalink.

88 Responses to The Higgs Discovery

  1. N. says:

    Good for Higgs, finally.

    Not so good for SUSY, as I understand.

    Congrats on the Grand Vacations!


  2. Sven says:

    Hi N,

    125 < 135, so a 125 GeV Higgs is perfectly compatible with the MSSM (not talking about more extended models).

    Cheers, Sven

  3. Gennaro says:

    Four sigma? Is that higher evidence than in 2011? Or less evidence? Please let us know…

  4. Chris says:

    4 sigma is significantly higher than last year. Still not enough to really claim discovery (5 sigma is the standard), but if they really are at 4 each then the combination would almost certainly do it.

  5. Susan Mones says:

    Any plot supporting those rummors ?

  6. Truth says:

    The Higgs was discovered in the 70’s when the standard model was put together; this is just the latest confirmation of the standard model. It once again disputes all the naysayers, who think the heirarchy problem is fake. No. The higgs is real. The standard model is real. Its properties, such as the heirarchy problem, are real. Nature has spoken, we have listened.

  7. OMF says:

    No. The higgs is real. The standard model is real. Its properties, such as the heirarchy problem, are real. Nature has spoken, we have listened.

    Be careful what you wish for! A 100% confirmation of the standard model by the LHC would in fact be the worst possible outcome for theoretical particle physics as a whole. Everyone would basically have to pack up their things and go home(and weep presumably).

    Better to hope for a multichromatic Higgs or the like. Fingers crossed!

  8. Henry Bolden says:

    Where’s the picture of Peter standing in front of the Grand Canyon?

  9. Pingback: ICHEP Higgs Rumours = Discovery ? « viXra log

  10. Pingback: El Segundo Luz » Blog Archive » ¿Se ha detectado el bosón de Higgs?

  11. Brian says:

    @ Truth that is completely laughable viewpoint. You can put together a 1000 wrong theories that make predictions. Just because you have a theory that predicts something not yet been seen does not mean you have discovered something. The Higgs was predicted in the 70s and that is all. Theory only supplies a framework to predict outcomes in reality. Reality can turn a theory to ashes or lend it creditability, but that is about it.

  12. Pingback: El Higgs “fermiofóbico” y los rumores para San Fermín sobre el Higgs en el ICHEP 2012 « Francis (th)E mule Science's News

  13. David Nataf says:


    The same standard model that predicted the Higgs predicted many other properties and particles that have been borne out by experiment.

  14. Pingback: Do you like to spread rumors? « Collider Blog

  15. anon says:

    Peter, when will you reach to civilization?

  16. Roger says:

    I agree with Truth. The Higgs mechanism is an essential part of the Standard Model and has been quantitatively confirmed since the 1970s. The value of the mass was unknown, and there is always the possibility of some other explanation for the data, but the Higgs was discovered in the 1970s.

  17. Noah Smith says:

    The Higgs was discovered in the 70′s when the standard model was put together

    the Higgs was discovered in the 1970s

    Scientific method FAIL

  18. Dan D. says:

    Sorry, but I have to agree with Brian and Noah. The idea that the Higgs boson was *discovered* in the 1970s when the *theory* was fleshed out is ridiculous. It’s enough of a triumph of the scientific method that a theory *predicted* the existence of such a particle 50+ years before it was finally (potentially) discovered, then to go overboard and claim a *prediction* is the same thing as a discovery. One may have high confidence in a prediction, but no matter which way you slice it, that’s no substitute for experimental discovery. Lots of people in cosmology had high confidence that the Universe was slowing in its expansion, too, and we know how that turned out…

  19. Beelzebud says:

    I’m no scientist, but wouldn’t it be correct to say that the Higgs was theorized in the 70’s and is still awaiting discovery? Does experimental validation matter at all to some of you?

    I think this is the sort of mindset Not Even Wrong and The Trouble with Physics was calling out. For something to be discovered, you must confirm it with experiments.

  20. Eric says:

    I am amused about these comments that infer the ‘strength’ of the standard model. If the model were so strong, it would have needed dozens of refinements over the years.

    And please pardon my skepticism, but an sd of 4 sigmas is hardly a solid confirmation. There are 5 sigma+ level events that occur on earth regularly by random circumstance.

    Why is the higgs pegged at 124 GeV in one very recent study, and 126 in another? If the resolution of the detectors is < 2 GeV, which I believe it is, then how can two ~3-sigma level findings 2 GeV apart mean the existence of a 125 GeV particle? The error exists on both the low-side of 124, and the high-side of 126. And the probability DROPS OFF in between. If anything, attention needs to be paid to possibilities including: 2 particles 124/126, 3 particles 124/125/126, or more likely, zero particles anywhere near 125 because this is due to fluctuation.

  21. Eric says:

    The model “WOUDN’T” have needed dozens of refinements over the years, is what I meant to say.

    And furthermore, I don’t see much evidence that people are thinking about what other possibilities could exist at 125 GeV OTHER than a higgs particle. The existence is completely inferred by the detection of indirect mechanisms in the first place, and the inference is fairly weak when you consider there may be multiple explanations for the behavior in the detectors including a completely new physical mechanism.

  22. Anonyrat says:

    Civilization no longer exists. Peter may have a ETA for the urban jungle.

  23. tommaso dorigo says:

    To Eric above (on 124 vs 126):

    I think your skepticism, although healthy, is a bit misguided.

    – planes crash on almost a daily basis so 5-sigma events do occur, true – but we do not look for a new particle a million times a day. The 5-sigma bar was set in HEP by common wisdom, and it has been surpassed very rarely by things that later boiled down to be false. The 5-sigma criterion is imperfect of course, but it is a good hunch. Note that it incorporates the famous “look-elsewhere effect”, in the sense that experiments are allowed to claim observation of a new particle based on local significance, i.e. not accounting for the trials factor.

    – the issue of 124 vs 126 has been put forth many times by people with handwaving arguments. If you know a bit of Statistics it won’t take you long to realize that they are fully compatible, and that the weighted average of two signals can then be taken. The exercise of 124+-2 oplus 126+-2 gives the maximum likelihood estimate of 125+-1.4; but you need to stop there since for the significance things are not as straightforward, so you should rather look at Peter Gibbs combinations (vixra log), which show that the global effect is of the order of 4-sigma for the Winter 2012 data (which had 3-sigmaish effects in CMS and ATLAS). The trick is that while the most sensitive channels have resolutions in the 1-2 GeV ballpark, the combined results of each experiment benefits from channels with lower resolution, so that the p-values have broad distributions.


  24. anonymous says:

    Is there a vacuum instability issue with an SM Higgs at 125 GeV?

  25. Mitchell Porter says:

    Eric: the standard model “needed dozens of refinements over the years”

    Could you list just one dozen refinements that it needed?

  26. ScientistfromMars says:

    Calm down, guys. The Higgs Boson had been discovered long time ago. Why all the noise now? I know CERN has to come up with results after eating up billions of taxpayer funds. But for what: I have one Higgs Boson in my freezer and another one stored in a vault. Both are really pretty, but a bit small tough. Now, I need to walk my dogs.

  27. Eric says:

    Hi Mitchell — Isn’t every run of every experiment to find the Higgs a ‘refinement’? The model doesn’t predict the mass. I don’t know how else you can look at it.

    Hi Tommaso — Thank you for your insight. But can you further clarify what you mean by ‘[physicists] don’t look for the Higgs particle a million times a day’? As above, isn’t every run of every experiment a trial? And is it wrong to break it down further by counting every single quanta of energy hitting the detectors as a separate ‘trial’, such that you really are looking for it “a million” times a day?

    Second, please forgive me for referencing a web page instead of working through the math myself, but I like ‘handwaving’ 🙂 :). Please read what Sean Carroll has to say about probability in terms of football coin flips, of all things:

    Finally, could you please offer some advice on what the true resolution(s) of the detectors are? I would like to put together a thought experiment, if I can find the time, to show that maybe there’s a different way to look at these results. The fact that you mentioned the experiment relies on data from detectors with even lower resolutions than 2GeV is both interesting and worrisome to me. I do know a fair bit amount statistics (although not as much as you ‘particle’ guys. You guys know your stats like nobody’s business!). Anyway, have you thought about applying the error in the detectors to model an explanation to describe a new/different physical mechanism at play here? Just curious.

    – Eric

  28. Eric says:

    Tommaso et. al. Please don’t take my skepticism in the wrong way. I am not trying to be rude, for sure. I’m just very interested in the experiments, and I think you guys are indeed doing a tremendous job with tremendously complicated equipment (which you built to perfection from scratch), faced with looking for a needle in a trillion haystacks.

  29. Truth says:

    Dear anonymous,

    Thank you for asking a relevant and important question, while most others here seem to be confused about the nature of science and evidence.

    To answer your question about vacuum instability: as is the case for EVERY value of the higgs mass, something bad happens at suficiently high energies, whether it is vacuum stability for light higgs, or a landau pole for heavy higgs. For a 125 GeV higgs, there is vacuum instability at an energy around 10^12 GeV. There is also an associated timescale for our vacuum to decay, but this turns out to be longer then the present age of the universe. So the standard model with 125 GeV higgs is META-stable. This is probably a bad thing that is fixed by new physics entering before 10^12 GeV, although perhaps meta-stability is okay, we don’t know.

    So the standard model is yet again holding true. It’s heirarchy problem is, in my opinion, its strangest feature, that most people think imdicate that the new physics will enter far low than 10^12 GeV. Or the heirarchy is anthropic. Those are the only possibilities.

  30. emile says:

    Truth: you wrote: “The Higgs was discovered in the 70′s when the standard model was put together; ” . Then you add: “while most others here seem to be confused about the nature of science and evidence.”.

    You are the one who seems confused. The Higgs has not been experimentally discovered yet. Maybe you are the only one for whom a “discovery” of a particle does not involve experimental observation. Seeing neutral currents in the 70s is not a Higgs discovery. Maybe “discovery” means something different to you (and you alone, because I don’t anyone in this field who would support your statement above).

    Let me blunt: if a student in experimental HEP said something like this during a PhD examination, he/she would not walk away from my University with a PhD if I’m on the committee. This is so “Scientific Method 101” that a “fail” is the only grade one can give.

  31. Peter Woit says:


    I’m back in the office, can now try and get a grip on moderating this discussion. Unless you have something interesting to contribute about the latest news here, please refrain.


    “Those are the only possibilities”

    Actually, no. For the obvious reason that at the edge of scientific research, we don’t understand what is going on well enough to know the “only possibilities”, and in this particular case, because of the subtleties of the “hierarchy problem”. But that subject is both off-topic and already beaten to death at another posting. No more here.

  32. tommaso dorigo says:

    My answer to Eric above:

    “But can you further clarify what you mean by ‘[physicists] don’t look for the Higgs particle a million times a day’? As above, isn’t every run of every experiment a trial? And is it wrong to break it down further by counting every single quanta of energy hitting the detectors as a separate ‘trial’, such that you really are looking for it “a million” times a day? ”

    Not really. The analysis is done once per year or so. And if you broke the data down in bits, you would have no chance of saying anything in any given bit. It would be stupid, and meaningless.

    About what Sean Carroll writes in his blog on probabilities: after three years working side by side with statistics experts who have published groundbreaking papers on statistics for HEP, I believe S.C. does not qualify as a source for me. And that’s just another topic on which he doesn’t, after his showing his bigotry in the issue of my comments to Lisa Randall’s physical appearance 😉

    [Sorry for my attitude in the paragraph above, I must be late in delivering my weekly dose of venom.]

    About resolution: mass resolution depends on the objects you use to reconstruct the decaying object. Of course we use that information in all our analyses. I am not sure I understand your suggestions, but please consider that routinely we extract results from a likelihood L(m;x,θ) which is a product of probability density functions of all the observed events. The latter are written as P(m;x,θ) where m is the measured mass of an event, x is a vector of additional parameters measured in that event, and θ a vector of nuisance parameters (for which we usually have additional priors from subsidiary measurements Π(θ) which multiply L).

    Now, in θ are contained systematic effects, but x is a measured vector of event characteristics, from which the mass resolution is deducible. So really P() contains that information; for instance if your measurement of m is gaussian, you could write P as G(m,σ(x)). I hope this is clear enough.


  33. Zathras says:

    Does anyone know something about the structural correlation between last year’s experiment and this year’s? Or can we just add the sigmas because they are structural independent?

  34. Peter Woit says:


    As far as I know, for the most important channels for the Higgs search, the errors are dominated by statistics. So, to a reasonable accuracy you should be able to combine the 2011 and 2012 analyses as statistically independent.

    This is a complicated business though, and I’m no expert, so surely there are subtleties about exactly how to do this that I’m missing. The experiments definitely are both doing combinations of the data from the two different runs (which were at different energies). I don’t know the details of how they do this. Maybe someone else here does…

  35. N. says:

    @ Sven:

    I meant Sir Peter H. , not the particle 🙂

    As for SUSY, no signs so far (as I understand it)?


  36. Gennaro says:


    in Germany, your blog announcement about the Higgs made its to the internet news ( and into radio today (monday). You are now a celebrity …

  37. Pingback: Higgs boson found? | Uncommon Descent

  38. Anonyrat says:

    A relevant physics question, I think. Matt Strassler wrote at the link provided in the update above, regarding combining 2011 and 2012 results: “So in combining them you are making a theoretical assumption about how the production rate for Higgs particles changes as you change the energy from 7 TeV to 8 TeV. ”

    Actually, at 7 TeV or 5 TeV or 8 TeV, the constituents of the protons are what are colliding, and these constituents have some energy distribution with non-trivial spread, and the physicists are making a theoretical assumption that they know the production rate along this energy spectrum. So I’m not sure what great additional assumption we are making in combining 7 TeV and 8 TeV results.

  39. Peter Woit says:


    I don’t seem to have made this clear enough: pointless arguments about the scientific method, what’s a discovery, etc. don’t belong here. Stop submitting them.

  40. Pingback: Gerüchte ums Higgs-Boson: Forscher bloggen über Gottesteilchen-Durchbruch | PRAVDA TV – Live The Rebellion

  41. tommaso dorigo says:

    Hi anonyrat,

    indeed, the PDF of the proton constituents have to be used in determining the production cross section of a hard subprocess, in what is called a “factorization integral”: this is (pardon my laziness of not using latex), say for producing a H particle inclusively (“+X”, where X is “whatever else”)

    σ(pp–>H+X) = Σ_(ij) Integral [f_i(x_i) f_j(x_j) σ_point(x_i x_j s, H) dx_i dx_j]

    where x_i and x_j are the momentum fraction carried by the partons which produce the hard subprocess, and f_i, f_j are their PDF; s is the squared proton-proton CM energy. As you see in σ_point enter only the partons and their energy, so this part “factorizes” from the mess of the proton-proton collision.

    The above does not mean that one has trouble combining data at different values of sqrt(s). Indeed, one has already taken assumptions about her capability to compute the pdf at a given scale, when using the factorization integral for a given s. Adding sigmas for different s does not cause a further headache, since it only means a minor “evolution” of the PDFs from one scale of hard process energy to a contiguous one. In other words, one is not sampling the PDFs in widely different ranges of momentum fraction in changing from 7 to 8 TeV.

    There is one further point to make. LHC experiments searching for the Higgs are basing their background predictions on data (and so do not care about sqrt(s) at which data is taken) or, in specific cases, on electroweak calculations (say for instance the ZZ background in the H–>ZZ searches) which are very well known and computed to NNLO accuracy.

    Finally, even within each experiment the methodology of combining the results should be that of extracting them separately from the two datasets (7 and 8 TeV), and then combining with the same tools LHC has used since last Summer to combine ATLAS and CMS data on the Higgs search. In a sense there is no difference in combining H->ZZ, H->γγ, H->WW, H->ττ results together and combining the same set twice, one for each sqrt(s) point.


  42. SLAC claims that BaBar Data Hint at Cracks in the Standard Model. The excess decays has to be still confirmed, but they claim that data already rules out the Two Higgs Doublet Model.

  43. jon says:

    About the last part of Peter’s post (“will the excesses in different channels agree with SM predictions?”) is there enough data already (5/fb of 2011 + 5/fb of 2012) to, in any case, answer it (i.e a summer annoucement possible) ? Or is the full 2012 run needed to beat any statistical bad luck (i.e. a spring 2013 announcement) ? Thanks!

  44. Peter Woit says:


    This depends of course on whether reality disagrees with the SM predictions, and by how much. The uncertainty in the size of the signal will be large at ICHEP, smaller by early 2013. If the SM prediction is dramatically wrong, there may be evidence of this at ICHEP, although likely not statistically convincing.

    The size of the signal depends on the Higgs production cross-section x branching ratio. The SM branching ratio should be accurately known, the production cross-section much less so (I’ve seen estimates of theoretical accuracy of 10-20 percent for this). Looking at the ratio of signals in two different channels (e.g. gamma-gamma and ZZ goes to 4 leptons) gets rid of the production cross-section uncertainty, but adds a lot of statistical uncertainty as you take the ratio.

  45. Dear Peter,

    I can not comment on the rumors themselves, but I can comment on the “Update” that you posted. Indeed, leaking the information (or making up stories) from a 3000+ experiment is a problem for a scientific process. We do not hide answers from you because we are such teases. We just want to figure out what the data tells us first, and it is only people working on experiment who are qualified to making the pronouncement.

    Since your postings are followed by people who may not appreciate the difference: the official word from the experiment is the reality. Everything else is just hot air and narcissism…


  46. Peter Woit says:


    I understand that it’s in the interest of the experiments not to have public announcements coming from them until the job of getting the most reliable possible result is finished.

    At the same time, I just don’t see the supposed problem posed by an accurate characterization of preliminary results appearing on this blog. Yes, it’s in principle possible that some reader here may not understand that these results are preliminary, but there’s a huge amount of misunderstanding of particle physics by the public, with very little of it caused by accurate information being posted here.

    Let me repost here a comment I wrote on Michael Schmitt’s blog, which is relevant:

    ” The bottom line here is that over the last couple weeks, 6000+ particle physicists, the majority of the particle physics community worldwide, have seen or heard about preliminary analyses of data from their experiments which pretty conclusively confirm the existence of the Higgs. This is huge, historical news, and I don’t happen to see why it shouldn’t be shared at this point with the rest of the particle physics community. Notice that I’m not posting plots, or much in the way of detail. We all look forward to the day not too far from now when the best results possible from the new data are released, and we can all give heartfelt thanks to those like you who worked hard to get there.

    Blogs introduce a new vector for the spread of rumors, but surely you’re aware that historically the news of a big result has circulated fairly widely among physicists in the days and weeks before an announcement, often in highly inaccurate form. The current official policy that no one on an HEP experiment should breath a word about results before the public announcement doesn’t correspond to the historical reality of the field. For instance, I remember going to tea at Princeton one day as a grad student back in the 80s, where I joined a group listening to Carlo Rubbia explain to everyone, with details, that his group had the top quark “in the bag”. That’s obviously not much of a good model for how to do things either, but the model of “no one will say anything at all to anyone” is neither realistic nor grounded in past practice.”

  47. Peter,

    while I love the Rubbia comment, Rubbia was may be the only person in existence to be able to present wrong results in Nobel lecture and keep his credibility. I don’t think we have people like that at the LHC. It was also a much different time, when our funding came from the echoes of the Cold War.

    Yes, “accurate characterization of preliminary results appearing on this blog” would not be a problem. This however is NOT what was posted on this blog, since there are no preliminary results out. And if you think for a moment that what you hear from the rumor mongers is THE TRUTH, just remember ATLAS’es 115 GeV higgs signal around Easter of last year.


  48. LHC security kernel says:

    Attention Everyone!
    Scientific process has been subverted!
    All LHC data servers are now undergoing mandatory reformatting.
    Access to network has been restricted.
    Please report to your supervisors for further instructions.

  49. Anon says:

    I think the blogs carrying these rumors have given a chance for many of us to feel the same “pre-we-know-for-sure” excitement before an official announcement that those in the collaborations and their friends with whom they share unofficially in the natural course of doing good science may be feeling. As long as the rumors are clearly labelled as rumors what is the harm that is done to science? In fact sharing a rumor on the internet as opposed to by word of mouth (that anyway happens) has the advantage that we know someone with a name and a face has originated it (on the internet)…..unless of course someone like Prof Matt doesn’t reference the internet source in which case s/he becomes an independent source of rumors.

  50. Peter Woit says:

    Some more details about the Rubbia story, since I just did a little research, and many people may not know this history.

    On July 4, 1984 CERN announced the discovery of the top quark by the UA1 experiment led by Rubbia, see here

    The Princeton tea I remember must have been a couple months earlier, spring 1984, since I left Princeton summer 1984 and teas would have ended with the end of the spring semester. Rubbia wasn’t just talking to physicists about this, see this May 9 story in the Christian Science Monitor:

    Of course it turned out that Rubbia’s 40 GeV top quark was a mistake, with the real thing at around 173 GeV, and actual discovery waiting until eleven years later at the Tevatron. So, I’m not claiming Rubbia’s behavior in this case as a good model for what the leaders of ATLAS and CMS should be doing…

Comments are closed.