Sadly CERN decided not to show any full LHC Higgs combinations today but we can always do an unofficial version again using the new ATLAS and CMS plots with more data. From 200 GeV to about 500 GeV everything is excluded and above that there is not enough data to say much so we just look below 200 GeV now. This is what we get

The previous Combo after EPS was consistent with a standard model Higgs somewhere between 125 GeV and 145 GeV, or a more complex mixture of bosons over a wider range. The conclusion has now swung back away from the standard model with masses above 135 GeV all but eliminated. There is still a signal for something but it is much less strong than before. The 3-sigma “observation” that CERN could have claimed has gone.

Technically there is still a chance for a boson at around 140 GeV, and a standard model Higgs boson is not excluded around 130 GeV but in that case the vacuum would be unstable or metastable unless there is something else such as superpartners. The Higgsless models have also been resurrected with an outside chance that the excess could fade away completely.

The case for a lighter Higgs at around 120 GeV is still wide open.

If you are wondering what it looks like with the Tevatron data added, the only difference is at the low mass end. Conclusions don’t change.

Like this:

LikeLoading...

Related

This entry was posted on Monday, August 22nd, 2011 at 7:26 pm and is filed under Higgs Hunting. You can follow any responses to this entry through the RSS 2.0 feed.
Both comments and pings are currently closed.

Just to make sure I’m reading these CL graphs correctly:

If a ~140 GeV Higgs with the SM production cross section exists, and you collected infinite data, the value of the solid black (observed) line on these graphs at 140 GeV should be exactly 1, correct?

(Not that you’d expect to ever get exactly 1 with a finite data set, of course.)

Yes, a perfect Higgs signal would have the black (observed) line at one plus the dotted (expected) line, at the appropriate mass. The dotted line goes down with more data so ultimately the Higgs should just leave a blib above the value of one.

Indeed, there already are people arguing with no Higgs at all up to 600 GeV. Next months are crucial for: a) Higgs mechanism ( SM-like) b) Low energy SUSY. I am surprised there are not theorists assuming that low-energy SUSY can be almost excluded up to 1 TeV, excluding some tiny intervals…I was told in a recent workshop that: “LHC triggering focus specially on Higgs-like, SUSY-like events. All the remaining ( or almost) is being disfavoured by the software”. Questions:

a) If low energy SUSY is absent, it has nothing to do with EWSB, what stabilizes the WW, WZ, ZZ amplitude is something new, what could it be? Compositeness? We should have to reanalyze the data with other assumptions. Interesting.

b) Low energy Higgs are still alive, how long?

What it seems clear from data is that we are going to find some interesting stuff at TeV scale. I can not wait to see the next data.

No Higgs seems to be raising its odds…And even very light Higgs and very heavy ( strongly coupled) Higgs is being pursed in mass parameter space: not too much space for risky odds! LOL

PS: What it also fun is the mediatic sphere, we are all excited with models or ghostly low statistics “bunches”, but experimental people are advertising us with our “primitive” models and cooking…Feynman, Böhr, even Einstein would be smiling from the other side! LOL

Yeah, but…Is the Higgs itself among that fairies? LOL LHC is not only being a SUSY naive models killer, it can be the Higgs killer…
Higgs=Ether-like concept! If it exists it would be a really interesting ether to discuss. Tempus fugit, Higgs!

Can you explain the meaning of the vertical axes in these plots? I know it is “95% confidence limit on sigma/sigma_SM”, but I’m not sure how to interpret that. Is the comparison made only for a Higgsless SM, or how to think about it?

If the black line is at X on the vertical axis for a given mass point on the horizontal axis, then they exclude an excess cross section of X times what is expected for a standard model Higgs at that mass with 95% confidence. “excess cross section” means anything above what is expected from the standard model background if there were no Higgs at that mass.

To me, scaling the Higgs cross section in a quantum theory isn’t really well defined, because there are other channels which interfere coherently with the higgs channel, so there isn’t really “the Higgs cross section”. In particular next to the pole, there could be strong destructive interference which would lead to a deficit in neighboring masses. In doing these plots, is one ignoring interference with background, or is one scaling the amplitude and then squares it, or is the question stupid and I miss the obvious ;)

I don’t understand why they would not be able to model such interference if they had to, assuming that measured values for masses of Z,W,t etc are accurate enough. Perhaps I am not understanding your question.

In any case I don’t think they need to calculate what would happen at neighboring masses to make these plots. It may be useful to do that to see better what a signal should look like, but for the plot they just need to know the cross section at the Higgs mass and far away from it.

If there is another process with the same initial and final state with scattering amplitude A, and our higgs process has a scattering amplitude B, we get our cross-section by calculating

sigma ~ |A+B|^2,

basically.

What has one actually considered theoretically when “the higgs cross section” is experimentally determined to be smaller than x times the SM “higgs cross section” – where does the factor x come in exactly?

Strictly speaking there is no such thing as a “Higgs cross section” because we do not observe a Higgs final state, only a cross-section to which Higgs production (B) contributes strongly among other stuff (A).

My question now is what is meant by for example the factor sigma/sigma_sm=10. Do we ignore the other contributions (A) on the higgs pole and simply scale |B|^2 -> x |B|^2 and compare that to data? or do we scale the amplitude, |A+ sqrt(x) B|^2 – or are those interference terms unimportant for a narrow higgs?

I think I see what you are saying. The cross-section is only defined for one channel, so my explanation for what the axis shows only makes sense when there is just one channel in the analysis.

I’m really only talking about one channel, such as gg->WW, which can have a quark box diagram and a higgs contribution, for example. Those contribute coherently to what you would call one channel I think.

I notice that for 110 GeV < m_h < 185 GeV, the observed data (black line) lies above the expected data (dotted line). Is this statistically significant to falsify the standard model on this basis alone?

It is impossible to put an objective figure on it because it depends on your personal theory dependent prior probabilities for the standard model and how big you think the trial error is. I.e. how many other plots did you select from before you decided that this one was interesting.

If it excludes anything it would exclude that there is no Higgs-like boson in this mass region. The standard model Higgs is not excluded.

Dear Bob, if the logic of your proposed falsification of the Standard Model is that “it couldn’t happen by chance that for this whole 110-180 GeV interval, one gets a uniform sign – excess”, then your argument is a little bit too fast.

Actually some of the very recent work by the CERN folks has focused on the question how many times the black curve (observed limit) crosses the dotted line (expected) and for this case, it is not really much greater than what you observe. So there’s nothing too anomalous about having an excess over a long interval. Note that there are really 3 crossing in the 100-200 GeV interval if you’re strict – and this is actually not too few.

Obviously, such wide excesses are due to some events that seem to result from particles whose mass can’t be measured too accurately – so they produces an excess for a wide interval of Higgs masses. This is surely the case of the WW channel that decays to particles including two neutrinos – that are invisible to the detectors.

I won’t ask you whether your method of combination correctly reflects the weights. CMS probably had more data than ATLAS in the yesterday’s final press releases and both CMS, ATLAS should already be giving more info than the Tevatron.

At least, I like that the 119 GeV bump from the CMS isn’t quite away even in the global combination. It’s expected to be smaller than at higher masses…

The combination is weighted with inverse square of the expected value line so it should automatically take into account the different amounts of data used by each experiment.

“At least, I like that the 119 GeV bump from the CMS isn’t quite away even in the global combination. It’s expected to be smaller than at higher masses…”

Sure, for the SM Higgs, but wouldn’t a 119GeV Higgs be an indication of SUSY, which means the cross section is unlikely to be anywhere near the SM one? It would be interesting to know what to expect for a “typical” MSSM or NMSSM Higgs.

[…] is traditional to present the results of searches such as Higgs hunting as Brazil plots that show us where a signal can be excluded at 95% confidence, but when the data starts to show a […]

[…] no podía ser menos, Philip Gibbs acaba de publicar su “New Unofficial Higgs Combo,” 22 aug. 2011, utilizando los datos publicados por ATLAS y CMS ayer lunes en Mumbai, India. […]

has a perfectly sensible theory why CERN hasn’t shown the ATLAS+CMS combination in Bombay: the reason may be that the combination turned out to be identical to your, Phil’s, graph. ;-)

If they dont deliver the promised combination plot within a reasonable time Phil is legally authorized to fully explore his famous public treasure trove, ha ha … :-P

I like it too that the 119 GeV peak still looks as expected by the wise people :-)

[…] no CMS+ATLAS combination has appeared. Philip Gibbs has hacked together an unofficial version (see here and here). Comparing the EPS data to the latest, one sees clearly that a marginally significant […]

[…] acaba de publicar su “New Unofficial Higgs Combo,” 22 aug. 2011, utilizando los datos publicados por ATLAS y CMS ayer lunes en Mumbai, India. La […]

Phil, a simple way to check your formulas is possible, it seems. Since you seem to have to graphs for teh single channels, you can add them youself and see if you recover the CMS or ATLAS plots. If the two “sums” for each experiment work, the your combination plot should also be ok.

It is hard to read the numbers accurately from some of the channels plots because they are steep, narrow and low resolution. Sadly they are unlikely to give the numerical data to an outsider such as myself.

Just to make sure I’m reading these CL graphs correctly:

If a ~140 GeV Higgs with the SM production cross section exists, and you collected infinite data, the value of the solid black (observed) line on these graphs at 140 GeV should be exactly 1, correct?

(Not that you’d expect to ever get exactly 1 with a finite data set, of course.)

Yes, a perfect Higgs signal would have the black (observed) line at one plus the dotted (expected) line, at the appropriate mass. The dotted line goes down with more data so ultimately the Higgs should just leave a blib above the value of one.

Indeed, there already are people arguing with no Higgs at all up to 600 GeV. Next months are crucial for: a) Higgs mechanism ( SM-like) b) Low energy SUSY. I am surprised there are not theorists assuming that low-energy SUSY can be almost excluded up to 1 TeV, excluding some tiny intervals…I was told in a recent workshop that: “LHC triggering focus specially on Higgs-like, SUSY-like events. All the remaining ( or almost) is being disfavoured by the software”. Questions:

a) If low energy SUSY is absent, it has nothing to do with EWSB, what stabilizes the WW, WZ, ZZ amplitude is something new, what could it be? Compositeness? We should have to reanalyze the data with other assumptions. Interesting.

b) Low energy Higgs are still alive, how long?

What it seems clear from data is that we are going to find some interesting stuff at TeV scale. I can not wait to see the next data.

No Higgs seems to be raising its odds…And even very light Higgs and very heavy ( strongly coupled) Higgs is being pursed in mass parameter space: not too much space for risky odds! LOL

PS: What it also fun is the mediatic sphere, we are all excited with models or ghostly low statistics “bunches”, but experimental people are advertising us with our “primitive” models and cooking…Feynman, Böhr, even Einstein would be smiling from the other side! LOL

Fairies are fairies, you know.

Yeah, but…Is the Higgs itself among that fairies? LOL LHC is not only being a SUSY naive models killer, it can be the Higgs killer…

Higgs=Ether-like concept! If it exists it would be a really interesting ether to discuss. Tempus fugit, Higgs!

Juan, I have been calling the Higgs a fairy for many years. Susy stuff, I call zombies.

I prefer neutrinos, no doubt about that! They are more ( ghostly, though) real. Boo! :)

PS: No one should put all the eggs in the same basket!

Can you explain the meaning of the vertical axes in these plots? I know it is “95% confidence limit on sigma/sigma_SM”, but I’m not sure how to interpret that. Is the comparison made only for a Higgsless SM, or how to think about it?

Thanks

If the black line is at X on the vertical axis for a given mass point on the horizontal axis, then they exclude an excess cross section of X times what is expected for a standard model Higgs at that mass with 95% confidence. “excess cross section” means anything above what is expected from the standard model background if there were no Higgs at that mass.

I have a question about that, Philip –

To me, scaling the Higgs cross section in a quantum theory isn’t really well defined, because there are other channels which interfere coherently with the higgs channel, so there isn’t really “the Higgs cross section”. In particular next to the pole, there could be strong destructive interference which would lead to a deficit in neighboring masses. In doing these plots, is one ignoring interference with background, or is one scaling the amplitude and then squares it, or is the question stupid and I miss the obvious ;)

I don’t understand why they would not be able to model such interference if they had to, assuming that measured values for masses of Z,W,t etc are accurate enough. Perhaps I am not understanding your question.

In any case I don’t think they need to calculate what would happen at neighboring masses to make these plots. It may be useful to do that to see better what a signal should look like, but for the plot they just need to know the cross section at the Higgs mass and far away from it.

I apologise if I my question was confusing :)

If there is another process with the same initial and final state with scattering amplitude A, and our higgs process has a scattering amplitude B, we get our cross-section by calculating

sigma ~ |A+B|^2,

basically.

What has one actually considered theoretically when “the higgs cross section” is experimentally determined to be smaller than x times the SM “higgs cross section” – where does the factor x come in exactly?

Strictly speaking there is no such thing as a “Higgs cross section” because we do not observe a Higgs final state, only a cross-section to which Higgs production (B) contributes strongly among other stuff (A).

My question now is what is meant by for example the factor sigma/sigma_sm=10. Do we ignore the other contributions (A) on the higgs pole and simply scale |B|^2 -> x |B|^2 and compare that to data? or do we scale the amplitude, |A+ sqrt(x) B|^2 – or are those interference terms unimportant for a narrow higgs?

I hope this is a little clearer…

I think I see what you are saying. The cross-section is only defined for one channel, so my explanation for what the axis shows only makes sense when there is just one channel in the analysis.

For the combined channel plots you need a more complex explanation using log-likelihood I think http://en.wikipedia.org/wiki/Likelihood-ratio_test

I’m really only talking about one channel, such as gg->WW, which can have a quark box diagram and a higgs contribution, for example. Those contribute coherently to what you would call one channel I think.

I notice that for 110 GeV < m_h < 185 GeV, the observed data (black line) lies above the expected data (dotted line). Is this statistically significant to falsify the standard model on this basis alone?

It is impossible to put an objective figure on it because it depends on your personal theory dependent prior probabilities for the standard model and how big you think the trial error is. I.e. how many other plots did you select from before you decided that this one was interesting.

If it excludes anything it would exclude that there is no Higgs-like boson in this mass region. The standard model Higgs is not excluded.

Dear Bob, if the logic of your proposed falsification of the Standard Model is that “it couldn’t happen by chance that for this whole 110-180 GeV interval, one gets a uniform sign – excess”, then your argument is a little bit too fast.

Actually some of the very recent work by the CERN folks has focused on the question how many times the black curve (observed limit) crosses the dotted line (expected) and for this case, it is not really much greater than what you observe. So there’s nothing too anomalous about having an excess over a long interval. Note that there are really 3 crossing in the 100-200 GeV interval if you’re strict – and this is actually not too few.

Obviously, such wide excesses are due to some events that seem to result from particles whose mass can’t be measured too accurately – so they produces an excess for a wide interval of Higgs masses. This is surely the case of the WW channel that decays to particles including two neutrinos – that are invisible to the detectors.

Nice graphs.

I won’t ask you whether your method of combination correctly reflects the weights. CMS probably had more data than ATLAS in the yesterday’s final press releases and both CMS, ATLAS should already be giving more info than the Tevatron.

At least, I like that the 119 GeV bump from the CMS isn’t quite away even in the global combination. It’s expected to be smaller than at higher masses…

The combination is weighted with inverse square of the expected value line so it should automatically take into account the different amounts of data used by each experiment.

“At least, I like that the 119 GeV bump from the CMS isn’t quite away even in the global combination. It’s expected to be smaller than at higher masses…”

Sure, for the SM Higgs, but wouldn’t a 119GeV Higgs be an indication of SUSY, which means the cross section is unlikely to be anywhere near the SM one? It would be interesting to know what to expect for a “typical” MSSM or NMSSM Higgs.

[…] is traditional to present the results of searches such as Higgs hunting as Brazil plots that show us where a signal can be excluded at 95% confidence, but when the data starts to show a […]

[…] no podía ser menos, Philip Gibbs acaba de publicar su “New Unofficial Higgs Combo,” 22 aug. 2011, utilizando los datos publicados por ATLAS y CMS ayer lunes en Mumbai, India. […]

Jester here

http://resonaances.blogspot.com/2011/08/higgs-wont-come-out-of-closet-part-ii.html

has a perfectly sensible theory why CERN hasn’t shown the ATLAS+CMS combination in Bombay: the reason may be that the combination turned out to be identical to your, Phil’s, graph. ;-)

haha

If they dont deliver the promised combination plot within a reasonable time Phil is legally authorized to fully explore his famous public treasure trove, ha ha … :-P

I like it too that the 119 GeV peak still looks as expected by the wise people :-)

[…] no CMS+ATLAS combination has appeared. Philip Gibbs has hacked together an unofficial version (see here and here). Comparing the EPS data to the latest, one sees clearly that a marginally significant […]

[…] acaba de publicar su “New Unofficial Higgs Combo,” 22 aug. 2011, utilizando los datos publicados por ATLAS y CMS ayer lunes en Mumbai, India. La […]

Phil, a simple way to check your formulas is possible, it seems. Since you seem to have to graphs for teh single channels, you can add them youself and see if you recover the CMS or ATLAS plots. If the two “sums” for each experiment work, the your combination plot should also be ok.

It is hard to read the numbers accurately from some of the channels plots because they are steep, narrow and low resolution. Sadly they are unlikely to give the numerical data to an outsider such as myself.

[…] excludes the standard model Higgs between 130 GeV and 480 GeV. This is slightly stromger than the unofficial version I put together last […]