Waiting For December 13th
    By Tommaso Dorigo | December 7th 2011 06:41 AM | 31 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
    It is by now public that Rolf Heuer, the Director General of CERN, in announcing for December 13th two back-to-back talks of the CMS and ATLAS experiments on their Higgs search results with 2011 data, warned that the results might not be conclusive yet. Besides, nobody really could expect them to be, since the sensitivity expected by both ATLAS and CMS in the still not excluded region of the Higgs mass, with 5/fb of data per experiment and 7 TeV running conditions, ranges from 2 to 4 standard deviations in the rosiest circumstances.

    Despite that, blogs around have raised expectations on the possibility that the Higgs boson might be spotted at less-than-observation-level significances in the whereabouts of 125 GeV. I say this without fear of being crucified by my LHC colleagues, since the discussion has been raging on many high-traffic sites for a while.

    One way to look at it, while we wait for December 13th, is to go back to the results produced by the good-old Tevatron experiments this summer. This is entirely independent information, so it gives us some extra discrimination power, so to speak, on where the Higgs might be hiding.

    So let us give back a look at this plot:

    In it, you see how a particular statistic (the log-likelihood ratio LLR) extracted from the combined CDF+DZERO data (black curve) distributes along the Higgs mass hypotheses. This statistic allows to discriminate the no-Higgs from the yes-Higgs hypotheses as powerfully as any other. The red dashed curve shows the value it should take at the correct Higgs mass, if the Higgs were present in the data; the black dashed curve shows instead what its value should be should the Higgs not be there (background-only hypothesis, or null hypothesis if you are a statistics expert).

    The dashes are only expected values: they show what is the median of the result of combining CDF and DZERO results, given that much data and given the analysis methods employed. The green and yellow band ("brazil band") instead shows the 1-sigma and 2-sigma possible ranges of the value the statistic would take for no Higgs anywhere. Of course the graph should be read by first fixing a mass hypothesis, and then reading off the values along the corresponding vertical line.

    So we see that the Tevatron data does show to prefer the signal hypothesis to the no-signal hypothesis for masses between 125 and 130 GeV. Of course the discriminating power of the statistic is poor, the signal red dashed line lying just one-point-something sigma away from the dashed black one. There's more to say: the 125 to 130 GeV mass region is the one where the Tevatron's sensitivity is the worst possible, as indicated by the getting together of the two dashed curves... Still, the 1-sigmaish result is an indication, if you believe the Tevatron results to that level of detail.

    That is all for today - just a teaser, I guess. But expectations for December 13th grow, despite the caution warnings. We will see what happens... Stay tuned!


    Let me tell you why it is not possible for Higgs boson to be there, because there cannot be any fields in a realistic understanding of the natural world. Fields were devised in the times of Maxwell to comprehend pre quantum phenomena. Every event has to have a particle/wave explanation, no field would fill in the details where a postulation is weak.

    Vladimir Kalitvianski
    As soon as all this is not decisive, I am rather annoyed with this fuzz about Higgs discovery. It won't be until several months or years of the data collecting and analyzing. And my guess is: no Higgs.
    Hi Tommaso, very trivial (for you) question:
    could you explain the main reason why the sensitivity (also for ATLAS/CMS) is so bad in the low-mass region?
    I mean below ~130 GeV or so. Which kind of background is present there? Or thre are other experimental reasons?

    Hi Anon,

    very trivial answer. It is not that at low mass bad things happen. Only, at higher mass the WW and ZZ decay channels open up, making the Higgs search quite a lot easier. The cross section goes down as the mass increases, but it is not a decisive factor for the integrated luminosities we have.

    I think Mr. Heuer has been hired by a conglomerate of travel agencies.
    Flights into Europe around the season have been "slow sellers" lately.
    I mean, appart from panicking bankers.

    There is no need to travel - the seminar will be webcast:

    December 13 is a special day in Sweden, because everybody is celebrating Lucia, a napolitanian saint. If you think that is weird, consider that americans celebrate the greek saint Nicholas at Christmas.

    From Syracuse in Sicily, not from Naple.

    Very much looking forward to this! Are they preparing extra servers at CERN for that webcast? I suspect there will be lots of folks wanting to watch it live, and that would be a pity if it had technical glitches due to overcrowding...

    The slides are normally made public at the same time or a bit earlier, so this could cut down on the traffic... If even because great expectations will be deflated by looking at them ;)

    Much like radio broadcast for a webcast the amount of traffic for the sending station does not depend on the number of listeners.

    Some previous webcasts from CERN have struggled under heavy load, especially the first collisions event. Perhaps they have upgraded since then but demand is going to be very high and it would help if the stream was rebroadcast via other services such as ustream or livestream.

    is it normal that they make such a lot of fuzz about two talks (even if they are announcing some new results)?
    Or should this fact itself tell us something? :-)

    Cheers, Sven

    Hello Sven,

    well, it is and it isn't. I was indeed somewhat surprised by the announcement Heuer made. I think the reason is that everybody expected these results anyway...

    The Stand-Up Physicist
    Let's say there is a signal around 125 GeV. What kind of work is then required to show that the signal has a particular spin? The Higgs must be spin 0, but if this was some really odd process that was not in the selection code, and it had a spin different from 0, then one see a real signal that was not the Higgs boson. The analysis of this huge amount of data must presume our understanding of what can happen is almost perfectly complete.
    "What kind of work is then required to show that the signal has a particular spin? "

    A simple way is the measurement of the angular distribution of the Higgs particles. Higgs boson is a scalar, and so we have to expect a spherical angular distribution.

    If it decays into two photons you are almost done. There is no angular momentum of the two particles in the centre of mass frame so the spin of the original particle is the sum of the spins on the particles it decayed into. A photon has spin +1 or -1 so you already know the source was spin 0 or 2. If they can get the polarization of the two photons they would know which.

    Thank you Nick and PhilG for your very helpful explanations on this score.

    Good answer. Pi0 (0-) decays into two gamma, and in very rare cases 3 gamma. In my answer, I used partial wave decomposition ;)

    Hello Nick,

    the decay of pizeroes to three photons has never been observed. It violates charge-conjugation invariance, so it cannot be an electromagnetic process...
    Hi Tommaso,

    Yes, there are just limits for three and four gamma channels. Thanks.


    Very nice post!

    Hi Tommaso,

    I think it should be terrible to know the truth and to be impeded to cry it out. But it is just a matter of days...



    Let's put sigma into some perspective.

    I would like to comment on the probability of a scientific discovery tuesday.

    Scientists measure their experiments by sigma.

    There is no scientific discovery about a sigma 2.5 or 3.00 as probably will be the case on tuesday. Maybe even lower. Orr maybe higher, but it's thereabouts.

    If Newton were sitting under his appletree and made 100 observations, and in one instance the apple didn't hit him in the head, it is a sigma 2.5. A sigma 5.o is EVERY time that the apple falls down, and that includes doing it a million times and more.

    THAT is a scientific discovery.

    At sigma 3 the apple fals wrong one out of every 370 times you do the experiment. That is not a scientific discovery either.

    My prediction is that the Higgs particle never will reach close to a sigma 5.

    Behind the prediction there is a theory, if you are interested.

    Google crestroyer theory and find it or visit directly at

    If GUT has a higher-than-expected cut off, then LHC will struggle to see only the "tail" of the new physics.

    Hi Tommaso,

    Nice article - but the full black curve seems to be under the dashed black curve all the way from 110-160 GeV.

    Maybe the dip at 130 is just a fluctuation on a background signal that has not been estimated correctly ?

    Paul Wells

    Hi Paul,

    I don't think that's evidence of systematic underestimation of backgrounds. The correlation length of the curve, given the searches combined in the result shown and their relative strength, is large for masses above 140 GeV, where the H->WW searches totally dominate the result. And also for lower masses it is sizable.

    In other words, if you have an excess of data over backgrounds at 150 GeV, it is going to affect a wide region, maybe 30 GeV wide.

    Tommaso, do the preferred masses at ATLAS and CMS differ by more than 2%?

    Hi Rick,
    let's wait for 15 more hours, would you ?
    Bonny Bonobo alias Brat
    Its already December 13th in Australia and has been now for 13 hours, I guess that we just have to wait even longer!
    My latest forum article 'Australian Researchers Discover Potential Blue Green Algae Cause & Treatment of Motor Neuron Disease (MND)&(ALS)' Parkinsons's and Alzheimer's can be found at
    55 minutes to go