Monday, 12 December 2011

Tinker Taylor Soldier Higgs

Tomorrow the ATLAS and CMS experiments will reveal new results of the Higgs search based on the full 2011 dataset. The date of the presentation was not chosen accidentally: December 13 is the anniversary of the capture of Saddam Hussein. Is this suggesting that tomorrow we'll see another notorious fugitive dragged out of its hole?

Two things are 100% certain because they appeared in an official statement from CERN:
  1. Neither experiment will announce the discovery of the Higgs, in the sense of a signal with a significance of 5 sigma.
  2. Neither experiment will exclude the Standard Model Higgs over the whole low-mass range.
The rest of the story can be to a good approximation reconstructed from blogs, facebook status updates, whispers in the corridors and coffee rooms, drafts mistakenly left on printers, microfilms smuggled in diplomatic mail, etc. The story that emerges can be summarized as follows:
  • The Standard Model Higgs boson is excluded down to approximately 130 GeV, but not below.
  • As already reported widely on blogs, both experiments have an excess of events consistent with the Higgs particle of mass around 125 GeV.
  • The excess is larger at ATLAS, where it is driven by the H→γγ channel, and supported by 3 events reconstructed in the H→ZZ*→4l channel at that mass. The combined significance is around 3 sigma, the precise number depending on statistical methods used, in particular on how one includes the look-elsewhere-effect.
  • CMS has a smaller excess at 125 GeV, mainly in the H→γγ channel. They have 3 events in H→4l as well, but they are oddly shifted to somewhat lower masses of order 119 GeV. All in all, the significance at 125 GeV in CMS is only around 2 sigma.
  • With some good faith, one could cherish other 2-sigmish bumps in the γγ channel, notably around 140 GeV. Those definitely cannot be the signal of the Standard Model Higgs, but could well be due to Higgs-like particles in various extensions of the Standard Model.
Stay tuned, in about 24 hours we'll know everything there is to know. More in-depth interpretations of the ATLAS and CMS findings depend on the quantitative details of the results, in particular, to what extent the signals in various channels are consistent between the 2 experiments and with the predictions based on the Standard Model Higgs. There is a good chance we're finally looking at the real thing, I'd say 50% based on the data alone and 80% adding our sincere convictions that Higgs must really be in that mass range. However, one should not forget that last summer the combination of ATLAS and CMS 1 fb-1 results was showing an excess corresponding to a ~145 GeV Higgs with the significance as large as 4 sigma, which was blown away when more data arrived.

You're of course welcome to fill in more details or paste an excerpt of the ATLAS or CMS draft into the comment section ;-)

20 comments:

Tony Smith said...

Is the 140 GeV bump ruled out as Standard Model Higgs only because its cross section is less than 100 per cent of that expected for the Standard Model ?

If there is a 3-state Higgs in which the Standard Model cross section is split up among the 3 states, one being at 140 GeV,
the other two in the range up to 240 GeV,
so that all of them have cross sections substantially less than 100 per cent of that expected for the Standard Model,
then
would that be consistent with the 5/fb data analysis
(which only excludes above 130 GeV for full single Standard Model Higgs) ?

Do either ATLAS or CMS see bumps (less than full Standard Model Higgs cross section) in the Higgs to ZZ to 4l channel in the range 140 GeV to 240 GeV ?

Tony

cliff said...

Noticed this very interesting tidbit from a danish article linked pn Google+:

"Denmark participates in ATLAS experiment, and the Danish side is Professor Jørn Dines Hansen, Niels Bohr Institute. 'In our experiment, we have a smaller peak at 2.5 sigma, and if the second experiment, after rumors have a corresponding peak, also has top at the same particle mass, it begins to resemble something,' "

[google-translated text]
http://ing.dk/artikel/124885-tampen-braender-for-higgs-partiklen-snart-slut-med-gemmesteder

Alejandro Rivero said...

The point of "drafts mistakenly left on printers", or in the printer queue, is always good to remember, as security goes. Also, a new one is "do not enter wikileaks and search xxx GeV from the institute IP address", and "if you search for a sensitive word and you see a hit in a blog, do not navigate to it, they will notice the search!"

Anonymous said...

Once upon a time, when I was still in physics and attending a phenomenology seminar from time to time, people were saying that a Higgs above 119 GeV (really, 116 or so, but they kept pushing it upwards) would start making things really hard for susy (mostly because of the little hierarchy problem). What's the status of that line of thought these days?

Anonymous said...

@Anonymous - that's true for the minimal supersymmetric standard model (MSSM). But experimental exclusions of standard model Higgs don't amount to an experimental exclusions of MSSM Higgs.

See the supersymmetry section on this post:
http://profmattstrassler.com/articles-and-posts/the-higgs-particle/implications-of-higgs-searches-as-of-92011/

Ervin Goldfain said...

Jester, on the likelihood of finding the SM Higgs in this mass range, you say:

"...50% based on the data alone and 80% adding our sincere convictions that Higgs must really be in that mass range."

Isn't a bit premature to assign chances at this time?

Kea said...

If all they have is a 2 sigma non result, I'll be extremely pissed off.

Anonymous said...

2-sigma is not a non-result. 2 sigma=95% confidence level. This certainly carries weight, assuming that all sources of error and look-elsewhere effects have been properly accounted for. Strictly speaking, if your standard is high, only 1 in 20 2-sigma results will go away. It's a shame that this doesn't seem to be what's happening in practice.

Anonymous said...

Why pissed off?

Anonymous said...

Ok, so let's now move on to supersymmetry !

Jester said...

They have definitely much more than 2 sigma at 125, hence the excitement. Still things can flip, like last summer.

Tony, 140 GeV in the SM is excluded by ZZ and WW channels, still you can have a Higgs-like particle with a reduced coupling to the gauge bosons, like e.g. a singlet scalar mixing with the Higgs.

Anon, Higgs at 125 in the MSSM means very heavy stops (above TeV) which means large fine-tuning. But if you go to the length of introducing 100 new particles then it's easy to introduce one more, like e.g. in the NMSSM, in which case the problem disappears.

Anonymous said...

17:17 anonymous here:

As far away as I always was from phenomenology, the MSSM at least had some aesthetic appeal (the weak susy breaking lagrangian not so much, but that's just parametrizing our ignorance). The NMSSM just seems icky, but maybe there was something nice about it I never picked up on.

PeteHH said...
This comment has been removed by the author.
cliff said...

Just to point out, while its true that generically much heavier scalar superpartners (stops, etc) would now require significant fine tuning, its apparently not the case of Gordon Kane's G2-MSSM models "due to an automatic cancellation of scalars and trilinears which are both close to m3/2 in this setup". Itss the same work that predicts a Higgs mass between 122 and 129 GeV.
http://arxiv.org/abs/1112.1059

While I realize its certainly early to draw serious lines between the Higgs mass rumors and this phenomenology, that doesn't stop me from being extremely interested in that suggestive correlation...

Jester said...

He doesn't say the fine-tuning is small, he says it's *reduced* wrt the generic case with 10 TeV superpartners. So it's reduced from humongous down to very large :-)

What I meant above is that for ~115 GeV Higgs one could still have the MSSM with stops around 300-500 GeV and large stop mixing. That would allow one to get away with ~10% fine-tuning of electroweak symmetry breaking. If Higgs is 125 GeV then that hope is gone...

cliff said...

Actually the exact phrase is "significantly reduced"! ;]

For details on this point he refers to http://arxiv.org/abs/1105.3765 which I haven't read in any detail yet.

Though if there is some good reason that the fine tuning is quantitatively worse than he suggests, Im definitely interested about it.

Anonymous said...

I wouldn't say its humongous finetuning. With stop masses greater than a TeV but less than say 10 TeV, you are looking at finetuning on the Mu parameter at the percent level, which isn't that bad. (it starts getting more problematic if the Higgs had been past 130)

Also, the analysis is a little foggy to begin with. Its typically done at the 1 loop level and you could argue that another order could improve things a little bit.

Jester said...

To be more precise, my fine-tuning dictionary: 10%=moderate, 1%=large, 0.1%=very large, 0.01%=humongous. In Kane's model the scalar superpartner masses are above 10 TeV so you'd expect FT worse than 0.01%. They argue, however, that due to some correlations between scalar masses and A-terms the fine-tuning is reduced by two orders of magnitude, so from humongous to large/very large. This may or may not be true, I haven't checked them on that, and there's few details in 1105.3765 (although the fact that A0 in fig. 2 is specified to the 3rd significant digit rings the bell...)

Anonymous said...

@12 December 2011 21:21 Anonymus:
95% CL does not mean 95% confidence that it is a real signal. It means that you get a 2sigma effect at 5% of your studies. If there is no new effect, _all_ 2sigma deviations will go away. And if you look at the right place, the chance that the signal will go away is 0.
At least if you don't do something wrong.

Paolo Valtancoli said...

It is intriguing that this hypothetical Higgs mass is related to the other W and Z masses of the
Standard model. Maybe this fact need to be understood.