Banner
    Higgs Expectations
    By Tommaso Dorigo | December 2nd 2011 04:38 PM | 48 comments | Print | E-mail | Track Comments
    About Tommaso

    I am an experimental particle physicist working with the CMS experiment at CERN. In my spare time I play chess, abuse the piano, and aim my dobson...

    View Tommaso's Profile
    Update: I am keeping out of this, but you may well be interested in reading what Gibbs, Woit, and Motl have to say about recent leaks on the ATLAS and CMS results. So I hope I won't be crucified for three general links now!

    Update 2: And it is now public that a seminar at CERN will be given by ATLAS and CMS on December 13th. So the wait is almost over, officially...

    I have not written in a while - a full week. This is uncharacteristic enough that I owe you some sort of explanation.

    I spent the week in Geneva, where I attended a very intense and tremendously interesting CMS week. CMS weeks are periodic gatherings of the CMS members (they occur every three months). In a long plethora of parallel and plenary meetings one can learn everything about the ongoing activities, the organizational aspects, and the most recent advances in analyses and searches. It is also quite useful for somebody who spends most of his research time off-site to hang around during CMS weeks, because it is the best time to get a hold, interact, and discuss with other members who also live and work off-site.

    I of course cannot tell you anything at all about what I learned. I am not even sure I want to tell you what I listened to in very broad terms, since I have grown tired of living "on the edge", when I believe I am on the right side of the rules while some collaborators think I am in fact on the wrong side, disclosing internal information or writing anything that bylaws and specific guidelines explicitly forbid me to write. The limit of what one can and cannot say is fuzzy enough that it is dangerous to get close to it, as I've got to learn the hard way.

    What I am pretty sure I can tell you, however, is the following: of course CMS and ATLAS are working night and day to produce results on their Higgs boson searches. And of course I know what CMS has in its hands. I do not, however, know what ATLAS instead has gotten in their 5/fb dataset; so I only know half of the story, and half of the story is not enough to understand what the ending is. Yet, for sure we can have a look at a couple of graphs produced some time ago, to figure out where we might stand when, in very little more time, the two collaborations will release their results.



    The first one above, by CMS, shows you what kind of "significance" of a Higgs boson signal CMS predicted in 2010 they could extract by combining all the relevant search channels where the particle is looked for, as a function of the Higgs mass. Here the way to read the graph is to assume a Higgs mass value -a point on the horizontal axis- and read off the median significance that the experiment expects to obtain. There are several curves, but you should only look at the full blue line, which corresponds to the actual data collected in 2011: five inverse femtobarns of 7-TeV proton-proton collisions.

    What we learn is that if the Higgs boson has a mass of 119 GeV (a value I have been favouring for a while now, as even my TEDx conference testifies) CMS alone should typically obtain a significance of just about three standard deviations. Beware, this is a "local significance", that is a number that does not take into account the famous "look-elsewhere effect", the effect of looking for the particle in a wide mass spectrum, whereby a statistical fluctuation has a multiplied probability of occurring somewhere. But if you had bet on a particular mass hypothesis since last year, you could well ignore the "look-elsewhere correction" to the significance, and be happy with its local value at the mass of your choice.

    And then, let us look at an ATLAS projection, again done one year ago, which describes instead the other side of the Higgs search: the job of excluding the particle's masses which are the wrong ones. In other terms, the absence of a signal at a particular mass may imply that the particle does not exist with that mass, if the sensitivity of the experiment is sufficient.


    In the figure what is shown is the "95% Confidence-Level upper limit" that ATLAS expects to set on the particle's production rate, as a function of the particle's hypothetical mass. Here we pick a place in the horizontal line and we can read off, for the corresponding Higgs mass, what is the upper limit that a search can set on the production rate of the particle (in units of the predicted rate, such that 1.0 is the SM prediction), if the particle does NOT have that mass. Only look at the green curve (the one below all others) since it is the one corresponding to the data in ATLAS hands right now.

    If the figure confuses you, focus on one simple thing. What matters is if the green curve is below the horizontal line at 1.0: since the production rate is computed in units equivalent to the predicted Standard Model rate, a upper limit below 1.0 means that the particle does not exist at that mass value. The figure is a projection of data still to come, but it shows what kind of situation we might be in once the ATLAS results are released.

    I should also explain that the purple and yellow zones are the excluded production rates by the Tevatron and LEP II experiments (the Tevatron ones have since improved, but they are not qualitatively much different). What is important but missing in the plot is instead the LHC limit, which now excludes all masses above 135 GeV or so.

    And then let me clarify something in addition, before I forget: a upper limit at 95% confidence level is not enough to assert that the particle does not exist: the particle is then only "disfavoured", because only once in twenty cases would the experiment find that result if the particle did have that mass and existed and behaved as the Standard Model predicts. It is, so to speak, an "indication" of the likely and unlikely places where the Higgs may be hiding.

    Now, from Summer conferences and following combination of results by ATLAS and CMS we know that the Higgs boson probably has a mass between 115 and 135 GeV, if it exists. Now, the graph shows that ATLAS has a chance to exclude the particle in that full remaining interval, if it indeed does not exist. In other words, we might be able to have a first indication that the Higgs mechanism and the whole model is wrong, if that is the case. But beware, it would only be an indication!

    There is reason for concern that such a result would be mishandled by the press. For we would be saying "There is a one in twenty or less chance to get data smelling as little of Higgs bosons as the data we actually got, if the Higgs does exist". And the press would say: "The CERN experiments prove that there is no Higgs boson!", with all the consequent easy-to-guess comments about failures, moneys spent in vain, etcetera.

    No: in order to really prove that our understanding of electroweak symmetry breaking is flawed and that there is no Higgs boson we would need a much, much more solid evidence than a mere "95% exclusion". I would not be satisfied with anything less than a 99.9% exclusion (over three sigma) across the full mass range.

    But I do not honestly believe that we will ever get into such a situation. I do believe, in fact, that the particle is there, and that it will be found very soon! So stay tuned and place your bets if you haven't already. Time is running short.

    Comments

    Why is the time running short? How long would it take to exclude the higgs across the full mass range?

    there will be a public seminar dec 13th on Higgs searches of ATLAS-CMS.
    We will see if these projections were "reliable"...

    D

    dorigo
    I don't understand this question Anon. Anyway, as I explained, in ten days the LHC results with 5/fb per exp will be out, and a 95% CL exclusion across the full mass range is in principle possible. However, to really rule out the Higgs it would take much more.

    Cheers,
    T.
    The UFOs are from the sub quantum world which the SM had been trying to write off for a century. It is the inertial locking of space which needs to be understood and STR needs to be modified before LHC folks can read what they see. I had been conducting experiments and some information could be found on my site anadish.com . You may not take me seriously for the time being; however, with 'time' things will change, as the need to explain the anomalies at LHC grow!

    Hfarmer
    And the press would say: "The CERN experiments prove that there is no Higgs boson!", with all the consequent easy-to-guess comments about failures, moneys spent in vain, etcetera.


    They certainly would write just that.  You can see how they misconstrued the whole neutrino Vs light speed issue.  With some saying "science" could be wrong about everything from evolution to climate change and so forth.    


    Heavens forbid the "unfair", "recipe for disaster" it would be for you to be the first to divulge and calmly explain what  that sort of news would mean.  Hopefully some of the press will take note that you said. 

    No: in order to really prove that our understanding of electroweak symmetry breaking is flawed and that there is no Higgs boson we would need a much, much more solid evidence than a mere "95% exclusion". I would not be satisfied with anything less than a 99.9% exclusion (over three sigma) across the full mass range. 
    No matter what is reported in the way of not detecting the Higgs particle the search for it is just beginning. 
     
    Science advances as much by mistakes as by plans.
    "No matter what is reported in the way of not detecting the Higgs particle the search for it is just beginning."

    I've been reading about the hunt for the last few years. Now that we may finally getting near headlines I have to confess I have no idea -- and I suspect much of the press and public have no idea , of what the TENTATIVE finding or not finding "means." Or, what a specific energy value means in terms of theory.

    Can anyone provide links to such information.

    PS: I do know the Higgs boson is supposed to provide mass to other particles. So, were there no Higgs, and particles do have mass, what are the alternate explanation(s) of their having mass?

    Steve

    Hfarmer
    Can anyone provide links to such information.
    Links no.  I can try to provide a visualization.


    A very non technical way to think of the Higgs mechanism is to imagine the whole universe filled with molasses.  It would be hard to move through such a universe.  There would be allot of drag.  The Higgs field, is that molasses and the "drag" it causes is what we know as mass.


    The energy needed to produce the particle is like a measurement of just how thick that molasses is.  Is this molasses in (the cold of) January or in (the heat of) July?  It would make a very big difference depending on which it is.  The difference between Higgs particle masses would predict totally different physical interactions. 


    I hope that visualization helps a bit.  


    Science advances as much by mistakes as by plans.
    Thank you Hontas. Makes good visual sense although I wonder if the high energy Higgs is the thick or thin molasses.

    But I'm lead to ask, since we "know: particles have mass -- what/which theory says where mass comes from were there to be no Higgs?

    And, what theory predicts a high E Higgs verses a low E Higgs?

    In other words i'm sure we care about the Higgs not because it tells us to expect mass, but because it settles a theoretical question.

    Steve

    Hfarmer
    Thank you Hontas. Makes good visual sense although I wonder if the high energy Higgs is the thick or thin molasses.

    Lets go with high energy being more like thick molasses.
    But I'm lead to ask, since we "know: particles have mass -- what/which theory says where mass comes from were there to be no Higgs? etc.

    Well, there are various theories.  One is that if there is a higgs like mechanism it could be the result  two fields interacting to the same effect as Higgs.   There are also various theories of whats called "quantum gravity" which could explain mass as an emergent effect of energy interacting with the discreteness of space-time.  (Matter being basically a specific state of energy as Einstein told us, but then lately even details of that theory are open to question.  Special relativity is baked into all of these theories and if it needs to be tweaked so does every theory that uses it.)

    What I would like to see is an announcement that they have observed something really surprising.  With Special Relativity and the standard model of particle physics being questioned it opens up the world for new ideas and approaches in a way it hasn't been for a long time.
    Science advances as much by mistakes as by plans.
    whitepanther65
    Well Tommaso, it is a masterpiece you have written here! You don't say anything specifically, yet I understand perfectly what you mean. And, believe me, a value around 120 GeV for the Higgs mass is also my favorite value!

       Cheers Paolo
    "And the press would say: "The CERN experiments prove that there is no Higgs boson!", with all the consequent easy-to-guess comments about failures, moneys spent in vain, etcetera."

    I think this is a very likely scenario on Tuesday the 13th. Given what's already been excluded my bet is on SUSY and a discovery of at least one low cross section Higgs next summer. The combination of 5+5/fb should be enough to exclude the entire range of the SM Higgs now. Even the rumored 2,5 sigma signal at 125 GeV should mean a two sigma exclusion at that mass.

    Ruling out the SM Higgs is obviously a result that will have physicists jumping with joy. Communicating that to the press will be hard. There will likely be headlines of "Higgs ruled out!" "Scientists waste billions!".

    dorigo
    I do not understand how a 2.5 sigma signal corresponds to 2 sigma exclusion. Where do you get this correspondence ? I assure you it is faulty.

    In any case, the rumours around, if true, mean that the scenario that worries many of us will probably not materialize.

    Cheers,
    T.
    My logic was that a 2.5 sigma signal, when 4.5 would be expected for an SM Higgs at 125 GeV with 10/fb, would mean a two sigma discrepancy. On a normal brazil band plot, the observed curve would be 2.5 sigma above the expected with no Higgs curve and two sigma below the expected with cross section equal to SM.

    Dear JollyJoker, let me join Dorigo in his criticism so that you're criticized both from the left and from the right. ;-)

    First of all, 2.5 sigma excess is rumored to occur for one of the detectors only, so it corresponds to 5/fb only, not to 10/fb, in an acceptable agreement with 4.5/sqrt(2)=3.2-sigma excess that your numbers seem to predict for 5/fb.

    Second of all, even if you predict mean value 4.5-sigma excess, 2.5 sigma isn't quite 2-sigma (counted in the new relevant sigma) away from 4.5 sigma. The new sigma is probably about sqrt(4.5)=2.1 and you would need to calculate it differently, anyway.

    dorigo
    Hello,

    well, please be all careful about the meaning of these "sigmas". Although Gaussian distributions have nice properties (that's why we assume that errors behave that way, btw), we cannot so nonchalantly add them or subtract them. And we cannot make inferences on a limit based on the expected excess and discrepancies from that.

    To make a very dumb example, anyway: suppose you expect 100 bgr events and 44 signal events at a given mass point. That would be eyeballable as a 4.4 sigma excess, if there were no systematics: you would be 4.4*sqrt(100) events away from no-signal predictions. Instead, let's say that you see only 124 events. Can you exclude the 44-event signal based on that ? The hypothesis you want to exclude is 144+-12 now, and you are now only 1.65 st. devs away from that.

    Systematics may significantly mess up this kind of picture; plus, if we talk about very small event counts, we cannot use Gaussian approximation for the Poissons, and we lose the ability to make those kind of "one cow more, one cow less" estimates.

    Cheers,
    T.
    I completely agree with your disclaimers, Tommaso, they were really the point of my comment, too. Still, there exist Gaussian limits and approximations in a wide variety of contexts where JJ's considerations become legitimate, because of the central limit theorem and other things.

    this make me think that in fact, these searches and limits are very model dependent.
    The problem is if our 44 signal prediction is reliable. It is not only a problem of background here.
    It would be possible to exculde a SM-higgs like particle (with a given c-s production ) but still to have an excess.
    I.e. there is another particle or the SM-Higgs production rate prediction is wrong.
    is that correct?

    D

    dorigo
    Stop there a moment. It is not wrong, it is just a test for something else.
    Cheers,
    T.
    yes exactly. Wrong is not the correct word. So the plot says something only in case of a SM like higgs.
    But who knows what else is hiding there?
    should One make the same plots for each model/parameter combination of a theory to have the understanding of what is going on? probaly it is not possible and there are other methods which are model independent.

    D

    Thanks for the explanation. I realize one can't reliably add / subtract exclusions like that, but I'm still of the opinion my estimate was more "a little inaccurate" than completely wrong :)

    I'm curious: why is the "no Higgs" scenario a "worry"? Why the "fear" of making an earthshattering discovery?

    JollyJoker:

    I would look at these plots with a grain of salt. Most of the results are based on MC simulations which prove to be not accurate in describing backgrounds in most Higgs channels. Important analysis can not only use MC. They have to get the background from data-driven estimations. This means the current picture could be quite different from the projection almost one year ago.

    Thanks, that's good to know. Of course, speculating based on uncertain backgrounds is far more fun than waiting for the real results :)

    The latest version of the rumor at NEW has Atlas at 3.5 sigma and CMS at 2.5, adding up to something like 4.3, so that seems compatible with an SM cross section anyway.

    JollyJoker:

    I would look at these plots with a grain of salt. Most of the results are based on MC simulations which prove to be not accurate in describing backgrounds in most Higgs channels. Important analysis can not only use MC. They have to get the background from data-driven estimations. This means the current picture could be quite different from the projection almost one year ago.

    So, there is less room for Higgs. One last question. What about the point from the green line, which is close to 110 GeV mass? What is the error bar there? In addition, what are the error bar of the point from the green line near 120 GeV?

    Thank you Tommaso and sorry to hear about all the issues blogging brings you.

    There is one thing I have wondered about several times regarding Higgs exclusion graphs:
    The reference cross-section is the standard model's but how stable are those graphs and the resulting exclusion regions under perturbation of the standard model?

    To explain myself:
    We start from a standard theory (SM) which has clear parameters (couplings, etc.), then we may consider perturbations of usual parameters, like masses of other particles, but there are other possibilities like adding particles, e.g. axions, or a 4th neutrino, or supersymmetric extensions, or GUTs. (Some of those perturbations can only be made "discretely" I think, though I could not give an example so I am not too sure.)
    For each perturbation parameter we may predict cross sections in a given channel. Then we have as many dimensions as parameters, for which to plot observed cross section upper bounds relative to prediction, \sigma/\sigma_T (\sigma_T denotes the theoretically predicted cross section for theory T, that is a parameter).
    The typical plot shows 1 dimension = 1 parameter, the Higgs mass, assuming the SM for all other parameters (with best fits provided by the Particle Data Group's review I guess).

    So how do our hypothesis testing vary in function of those other parameters?
    Some theories will be totally unlikely but there are plenty of small changes to the standard model which are very reasonable given current experimental data, of all origins.

    My background in physics is weak so I'd appreciate any explanation or opinion.

    p.s. And one more question. Should we expect that sigma/sigma_sm saturates with respect to fb^-1? If the answer is "yes", then we might expect, that when we go from 5 to 8 fb^-1, the line doesn't go down so much (SM Higgs hope is still here). If the answer is "no", then at 8 fb^-1, the line "sinks" even more than what is at 7 fb^-1, i.e. no hope for SM Higgs.

    p.p.s plot that shows sigma/sigma_sm vs total luminosity.

    dorigo
    Interesting questions Nick. Our model of reference has parameters specified to the best of our knowledge, and we account for the uncertainty when we set a limit; however, significant deviations from what we currently know (e.g. say we messed up one measurement) might alter the prediction. But we are talking of minor effects. Nothing that an additional 20% of LHC luminosity can't wash away.

    Regarding the last question, I am not sure what you refer to. More luminosity means more events, so more significance of a signal, or more certainty that a signal isn't there. Significances, if statistics-dominated, scale with the square root of the data size; when you increase the CM things vary depending on the parton distribution functions and the likelihood that they can produce a collision hard enough to produce a Higgs.

    Cheers,
    T.
    That's right. When luminosity increases, we open a bigger phase space (f(Q)).

    I was curious what happens with sigma/sigma_SM, when we reach 8 fb^-1. According to your second plot, sigma/sigma_SM < 1 at 5 fb^-1 for masses > 115 GeV. Then, could we expect that sigma/sigma_SM << 1, when we reach 8 fb^-1 ?

    ...in order to really prove that our understanding of electroweak symmetry breaking is flawed and that there is no Higgs boson we would need a much, much more solid evidence than a mere "95% exclusion". I would not be satisfied with anything less than a 99.9% exclusion (over three sigma) across the full mass range.

    But I do not honestly believe that we will ever get into such a situation. I do believe, in fact, that the particle is there, and that it will be found very soon! So stay tuned and place your bets if you haven't already. Time is running short.

    Hi Tommaso, you sound slightly pro-Higgs prejudiced. I have a draft paper at http://rxiv.org/abs/1111.0111 which shows a way of avoiding the electroweak symmetry breaking problem, by changing electromagnetism from U(1) to a massless SU(2) gauge theory (which works out correctly, yielding Maxwell's equations from the Yang-Mills, because charged massless vector bosons can't propagate asymmetrically), so that SU(2) becomes a complete electroweak theory. (This is fine for the weak bosons, while the apparent discrepancy between weak isospin charges and fractional quark electric charges disappears with a vacuum polarization model, which predicts that 1/3 or 2/3 of the electric charge energy of quarks is present as strong colour charge.) U(1) is not abandoned altogether; it is dark energy, which also predicts gravity. The mass of SU(2) weak bosons is then produced by the Glashow-Weinberg mixing of U(1) gravity with SU(2) electromagnetism. Instead of a electroweak symmetry being broken to yield Nambu-Goldstone "Higgs" bosons, instead the weak interaction emerges from a simple mixing of SU(2) electromagnetism with U(1) gravity. I'll try to get a briefer paper done, ready to replace the Higgs boson.

    Tommaso
    I want to listen to your opinion.
    If the sigal at ATLAS and CMS remains at almost 3.5~4 sigma until the end of next year, what will CERN conclude?
    As you know, the rumors are that ATLAS and CMS see excess 2.5~3.5 at 124 or 1 26 GeV.
    If the level of sigal doesn't come up to 5 sigma and continues to remain at 2,5~3.5 sigma, they will Higgs exists or doesn't exist,
    How do you think of my question?

    dorigo
    Dear Younghun,

    to me, a 4-sigma signal is enough for observation, because I am quite sure that the Higgs is there (high prejudice about the signal hypothesis). Even a 3-sigma (honest 3-sigma) signal would be enough for me. But then we want to dig deeper, see the signal individually in each decay channel, etcetera, so the story is not over anyway.

    Cheers,
    T.
    In my theory, the Higgs particle will not be found.

    My take on dark energy and dark matter is, that they are miscalculations, stemming from our presumption that the speed of light is constant.

    What if the speed of light varies through time and space?

    That creates some interesting theory, at least I think so.

    Antimatter is the mind and consciousness of all living entities.

    You are your own universe.

    Reality is where the minds (antimatter) meets the physical universe.

    Interested? Then read my philosophical multiverse theory.

    Google crestroyer theory, and find it instantly.

    http://crestroyertheory.com/the-theory/

    Hey look! Is it the Higgs there at 125GeV?

    ....NOPE! It's Chuck Testa.

    Hi Tomasso,

    I have now repeatedly read that a 125 GeV Higgs is an indication that SUSY is correct.
    What is your oppinion?

    Cheers,
    Martin

    Hi Tomasso,
    I have now repeatedly read that a 125 GeV Higgs is an indication that SUSY is correct.

    Sorry,
    I meant ToMMaSo.

    dorigo
    Hi Anonymous Martin,

    my opinion is that 125 GeV is an indication that the Standard Model is correct.
    All the rest is speculative.

    Cheers,
    T.
    Hi Tommaso,

    Is there a leak here?

    Ciao,

    Marco

    I just read this in the guardian:

    Martinus Veltman, Universities of Michigan and Utrecht. Nobel prize in physics, 1999
    "You are mistaken about the Higgs search at Cern. The machine runs at half energy so far, and no one expects relevant (for the Higgs particle) results. After the shutdown [in 2013] the machine will gradually go up in energy, and if all goes well (this is non-trivial) then in about half a year the machine energy might reach design value and there might be Higgs-relevant results. So if you are thinking next week then you are mistaken. Of course, we never know what surprises nature has in store for us … It is my opinion that there is no Higgs."

    do you have an idea of how he gets to this conclusion?

    In reply to various comments above:

    1. Don't pay any attention to what Tini Veltmann has to say nowadays.
    2. Mh=125 GeV means the SM is correct? I don't think so: it runs into an instability at some high scale. High, but well below the Planck mass. Some new physics should kick in well below that scale and SUSY is a very good candidate for that, with low Higgs masses preferred
    3. What is gonna Kea do after the fairy field is discovered? Will that keep her away from polluting the blogosphere?

    whitepanther65
    I agree that Mh=125 GeV would be good news for susy (Mh=140 GeV would be bad news instead). But who is Kea? On Wikipedia I read "
    The Kea (pronounced /ˈkeɪ.ə/; Māori: [kɛ.a]) (Nestor notabilis) is a large species of parrot
    "
    Uh?

      Ciao Paolo

    dorigo
    Hi Paolo,

    Kea is the pseudonym of Marni Dee Sheppeard, a theorist who works on category theory and M-theory, plus a friend of mine.

    And I take the opportunity to ask commenters here to avoid insulting her, at least in this blog, or you run the risk of being censored.

    Cheers,
    T.
    whitepanther65
    Thank you Tommaso; Kea has a very nice blog indeed! It's already in my bookmarks.

       Ciao Paolo
    Greetings from a fellow Quantum Diarist! Great blog, Tommaso. I am really looking forward to Tuesday. I expect there wouldn't be such a rush to give the result and such a buzz if the result were negative... :)

    dorigo
    Hello David,

    well, depends what you call negative or positive. I am certainly going to believe that the Higgs boson exist (and tell my students too -I am teaching a subnuclear physics course and have held on the LHC Higgs results to be able to discuss the latest ones with them after Dec 13th) if two experiments both claim a 3-sigma excess in the same place. F*** the 5-sigma observation-level discovery threshold in this case - we would know that is it. But the experiments will keep a lower profile.

    Cheers,
    T.
    Neutrinoes and Higgs both cannot coexist -- either one has to be wrong. It's DCE research and supeluminal speed which has the potential of breaking current scientific barriers, rather than finding a nebulous statistical dual peak for a Higgs, which well could be due to many other anomalies, one that LHC could not decipher is that of the UFOs.

    Ladies & Gentlemen,

    The book is open. I will say 133.2 GeV thingies.

    Of course there is a consequence of such a low mass, and that relates back to that so called "first" milliseconds of existence. It could not have been a BANG, a definite rush, yes, but no bang. Think about it a bit.

    Remember you got it from Uncle Indardoohda first, , ,

    Indardoohda