Friday, 22 July 2011

Higgs won't come out of the closet

Today we had a true fireworks display in Grenoble: at the EPS conference the LHC experiments presented their Higgs search results based on 1 inverse femtobarn of analyzed data. The cream of the cream is these 2 plots:












Here is just a few fleeting remarks.
  • There was a significant chance to catch him yet the bastard escaped once again. However it's becoming increasingly clear that this year we will learn whether the Standard Model Higgs exists or not.
  • CMS excludes the Standard Model Higgs in the 149-206 GeV and 300-440 GeV windows (plus a few peeping holes here and there), while ATLAS excludes the 155-190 GeV and 295-450 GeV windows. The low mass exclusion is dominated by the search of the H→WW→2l2ν final state, while the high mass one is dominated by H→ZZ after combining different Z decay channels.
  • The exclusion range in the low mass region is smaller than expected. Indeed, there are hints of Higgs-like events in the mass range 130-140 GeV. This is nicely visualized in a plot from the ATLAS talk. The excess in the combined plot is driven by a broad excess WW → 2l+MET events. In certain mass regions the excess is amplified by γγ and ZZ→4lepton excesses, and reaches almost 3 sigma significance. CMS also has a 3-sigmish excess in that same region. This could be a fluke, a mismodeled background, or a first glimpse of the real thing. If the latter is true, we may learn it very soon!
  • We're looking forward to the ATLAS/CMS combination which should be ready for the next big conference: Lepton-Photon in Mumbai. Most of the high-mass region, up to almost 500 GeV, should be excluded by the combination, and it's not impossible that the low-mass Higgs signal will pop above the 3 sigma surface...
  • Both ATLAS and CMS presented their searches in the ZZ→4l channel. Yesterday CDF tried to launch its own firework - a statistically large excess of 4 events with two Z bosons decaying to 2 leptons each near the ZZ invariant mass of around 330 GeV. However that firework fizzled out, as none of the LHC experiments sees any ZZ → 4l excess in that mass region. Given that, there is no way the CDF result can be due to a Higgs or any other new particle; it's either a bad fluke or mismodeled background.
  • We can officially announce that Tevatron is out of the Higgs business. Both ATLAS and CMS on its own have much more powerful exclusion limits than the combined Tevatron exclusion from last summer. LHC should collect 3-5 times more luminosity by the end of the year, which will allow them to beat Tevatron's sensitivity also in the mass region near 115 GeV. Higgs hunting has moved to Geneva, for good...
On a different front, LHCb and CMS presented important limits on Bs → μμ branching fraction. Recall that CDF recently saw a 2-sigmish excess corresponding to Bs → μμ branching fraction of (1.8 ± 1) × 10−8, which, in spite of low statistical significance, prompted some excitement among theorists. However, that central value is now excluded by CMS at almost 95% and by LHCb at more than 95% confidence level. So CDF result seems to be just another fluke... bad luck.

-----
On viXra log Phil is doing a great job of keeping us updated in real time on what is going on at EPS; see this post for a royal collection of Higgs plots. Matt Strassler is blogging live from Grenoble (Et tu, Brute?). See also Tommaso's comments on CMS searches. Tomorrow more excitement guaranteed :-)

14 comments:

Kea said...

More fairy field exclusions! Yipee!

tulpoeid said...

Kea, I'm sorry to be the one to break it to you, but.
I don't know you and your work, but seeing you making the same information-free comment under every relevant blog post, not only on this particular blog, is extremely annoying and more importantly makes people who are rather glad to see these exclusions look a bit, um, let's say that it helps categorising these people in not-too-favorable categories by those who would be inclined to do so.

Kea said...

Tulpoeid, if you are too stupid to read my work, I can hardly be blamed.

Emma said...

are we 100% certain that the excess CDF showed couldn't be something?
i mean
is there so much data from the other side, that can exclude a heavy Higgs of 300+ GeV ?

ps: i'm no expert, that's why i ask!

Jester said...

Right, I don't think the CDF 4l excess has any chance of being real. Although Tevatron has analyzed more data (6:1 in this case), LHC compensates it by a larger production cross section and a better efficiency. If it were a Higgs, the cross section at the LHC would be 50 times larger than at the Tevatron, and then ATLAS/CMS should have seen a signal up to the roof. If the CDF excess were from a particle produced in q-qbar collisions, in which case the cross section is only about 4 times larger, ATLAS and CMS would still have seen it thanks to their superior lepton reconstruction efficiency. On top of everything, CDF sees an excess only in the lepton channel, and nothing in other ZZ decay channels.

Torbjörn Larsson, OM said...

@ Kea:

It seems to a disinterested outsider that tulpoeid suggested irrelevancy, not "irreadability".

[OT, but ironically that would make two information-free comments in a row, a good validation of tulpoeid's observation.]

Kea said...

Oooooo, dear me! Let's make it 3 pointless comments, just to annoy all the stupid people.

chris said...

any example of a recent CDF result that was not junk?

time to decomission that experiment

Anonymous said...

@Chris : Indeed, it seems that these guys are just polluting arXiv with disinformation these days...

Jester said...

Top forward-backward asymmetry still could be real...

Anonymous said...

I don't suppose you could tell us any more about the prospects for that before, say, 2:30 CDT?

Andrew Oh-Willeke said...

Do the high mass Higgs exclusions also exclude higher mass members of SUSY Higgs boson multiplets in those mass ranges?

If not, the exclusions from precision electroweak measurements make these results irrelevant. But, if so, then this presumably puts serious constraints in Higgs multiplet theories and might even rule out some of those before we know the fate of the SM Higgs.

Jester said...

No, the results I was referring to apply to the SM Higgs only. In general, the SUSY Higgs partners have completely different production cross sections and decay branching fractions. LHC has other results applicable for SUSY Higgses but they are less strong, and heavy Higgs partners are perfectly allowed for the time being.

Anonymous said...

I discount these claims of a 3sig signal already been observed. This is very misleading because it suggests there’s a 95% chance we have seen the Higgs. In reality the data more conclusively excludes the Higgs than vice versa. Why? Because the combined channel statistics make no sense in light of what the LHC has been seeing.

Take a look at the way the WW background behaves. Our SM expectation (this is loose because a whole pile of assumptions are needed to relate the SM predictions to what ATLAS actually SHOULD see) has about a 2sig deviation from what ATLAS actually is seeing over the 100-170GeV energy range. That’s not surprising, because of the W mass the branching ratio is very low there and so a low unaccounted sideband will have a much greater influence. Yet almost the entire of the present 3sig from the Higgs comes from these handful of events.

In this light we really have to look at the three main channels separately. Lets list the events as: expected background, expected background + SM Higgs, observed. I will sum events over the expected Higgs width in each channel (you can see the width in the ATLAS histograms). We have

In the WW channel a strong signal for a 150GeV Higgs: 43back/64 back+H/64 observed. So the data is spot on with a SM Higgs, at a 3sig CL. BUT caution, over the rest of the WW spectrum 20 events are expected versus 29 observed, a 2sig deviation, and also positive.

In the gamma_gamma channel for a 150GeV Higgs: 31oback/380back+H/320 observed. So we would have expected almost a 4sig Higgs signal, but got almost nothing. The Higgs is fairly narrow in this channel (only about 5 GeV, see ATLAS plot). So it makes sense to separate the spectrum. If we repeat the analysis at 130GeV, we have 310/370/340, that is 2sig expected vs. 1 sig observed.

In the H-bb channel at 130 GeV: 700/820/700. So a 0sig deviation vs. 4sig expected. (I neglect the bb via ZH since its much smaller than via WH)

I think the data speak more clearly for no Higgs than anything else, I would be glad if someone can show where the above is wrong.

From the ATLAS presentation, almost all luminosity came in the last month. So in 3 months we will have 4 times the integrated luminosity, and so all the sigmas above will be multiplied by 2, and if the above trend continues, this will exclude the Higgs at something like 5sig.