There’s a new paper out this evening from a large collaboration entitled Implications of Initial LHC Searches for Supersymmetry. Instead of just adding it to the bottom of my recent posting, I thought it would be a good idea to start a new one, and add a bit more explanation of what is going on.
For a good news story from today by Kate McAlpine, see this at Physics World. For excellent more technical explanations, see the latest blog postings at Tommaso Dorigo’s blog (today) and at Resonaances (yesterday). Physics World, Tommaso and the new arXiv preprint discuss published results from CMS and ATLAS, while Resonaances discusses even more stringent preliminary limits on SUSY from ATLAS made public last week at Aspen.
Tommaso also refers to a 2008 guest blog posting by Ben Allanach explaining how statistical predictions for SUSY masses were being made, adopting various simplifying assumptions (CMSSM) and assuming supersymmetry solves the problems it is advertised as solving (muon g-2 anomaly, dark matter, etc.). Allanach discusses the 2008 version of this kind of calculation by the same group that has just put out a new, 2/22/2011 version this evening.
The usual model for how science is done is that theorists make predictions before an experiment is done, then when the experimental results come in, they get compared to the predictions. That’s not quite what is going on here, where as far as I can tell, the new paper doesn’t directly compare the 2008 predictions to the new experimental results. Instead, the new experimental results are used to make new predictions. Since a large part of the parameter space favored in the 2008 predictions has now been ruled out, the new ones move the favored part of parameter space up to higher particle masses. The authors do make clear what is going on, showing on their plots a “snowflake” where the 2008 best-fit value was, and “stars” for where the new best-fit values are based on data from the two experiments. Note that the paper does not include the latest, stronger results from ATLAS announced last week, which presumably would move the “stars” up to even higher mass.
While the question this paper addresses about where supersymmetry might be given that it hasn’t been seen yet is interesting, it leaves unaddressed the more conventional question: do the LHC experimental results show that the theoretical predictions about supersymmetry made in 2008 before the machine was turned on were wrong? This is a statistical question, so should have a statistical answer. Assuming that the LHC continues to not see supersymmetry as it collects more data, I’m interested in the question of how the experimental data will falsify the theory. Will its proponents just keep calculating statistical predictions of higher and higher masses as lower ones get ruled out? Most will undoubtedly at some point throw in the towel, although there will be some who will never, never, never, never give up (see here):
SUSY may still be there even if it remains invisible to the LHC, indeed. And yes, I don’t hide that I will be convinced that SUSY is there even if the LHC doesn’t find it. The LHC will only confirm or exclude effects at particular regimes – usually low energy but it’s not quite accurate a description of the regime that may be excluded.
What I have been scared for several years is the pseudoscientific propaganda of your kind trying to claim – without any justification – that not seeing SUSY at the LHC should imply that physicists shouldn’t be allowed to work on SUSY or believe that it is a key feature of our Universe. There are many reasons to think it’s the case and theorists whom I consider any good will continue to treat SUSY as an essential feature whether or not it shows up at the LHC.
Update: See figure 1 of this evening’s What if the LHC does not find supersymmetry in the sqrt(s)=7 TeV run? to see how how much of the predicted region of superpartner masses was ruled out by initial LHC results, and how much of the rest is likely to be ruled out during by the 2011-2 7 TeV run.
Update: There’s a very new up-to-the-minute survey of LHC results concentrating on supersymmetry by John Ellis here. Unfortunately no figures that superimpose CMS/ATLAS exclusion regions on the statistically favored regions for supersymmetry that are discussed (based on assuming supersymmetry explains dark matter and the muon g-2 anomaly). It does look like this year’s data should be able to convincingly rule out the idea that supersymmetry explains both of these phenomena.
Update: The ATLAS results providing the strongest limits so far on SUSY are now out, see the paper here.
Sven,
Interesting to see that you’re abandoning the naturalness argument for supersymmetry. Arguing that it’s remarkable that by doubling the number of degrees of freedom and adding 120 new parameters one can fit some conjectured new physics isn’t exactly convincing.
Will your group put out a new analysis soon using the latest ATLAS paper, and showing how their results compare to your 2008 CMSSM predictions? It seems to me that your argument that CMSSM explains muon g-2 AND dark matter has likely already been blown out of the water by ATLAS. But maybe more data is needed, I’d love to see an appropriate plot.
Arkani-Hamed slides on “Approaches to the Hierarchy Problem” (PDF file, best I can tell, it is from July 2004)
http://www.sns.ias.edu/~arkani/pdfs/HierarchyAppr.pdf
Peter,
our simple realizations of the MSSM have 3 or four new parameters, not more.
Consequently, the chi^2/dof is excellent.
Our latest paper (the one that you discussed) includes already the latest official
ATLAS results (although not the one that will appear only next week 😉
Sven,
The latest ATLAS paper is out this weekend, see the link I added to the posting. Get to work….
Sven,
You have a very selective way of arguing.
The test of a good theory is whether it can predicts. The only semi convincing point you gave in the earlier post was the prediction of the top-quark mass, but this is a half merit since the particle itself was predicted by the SM.
The cherry on top is abandoning naturalness and at the same time brag that the model can “easily produce a Higgs mass value high enough”.
Peter and Bernhard,
The current constraint on the Higgs mass in three-generation SUSY models is that it should be less than 130 GeV. A Higgs mass of around 120 GeV is well within this range, and so I believe it is completely incorrect to claim that this somehow makes a SUSY solution to the hierarchy problem unnatural. If the experimental Higgs mass limit ends up being pushed above 130 GeV, then perhaps this will be the case, but until then I think it would be a good idea for you to stop exaggerating the issue.
As for Arkani-Hamed, I believe that he was also exaggerating the problem in the talk you mentioned in order to sell his own alternative models. Other theorists such as Ellis would disagree with him, as do I.
I am a financial statistician. Just want to point out a problem in Sven’s latest comment here.
You seem to be saying that your latest model has only 3 or 4 parameters and therefore can produce quite decent statistical significance. This is actually a common folly in the trading business. People spend months looking for the “right” model that gives high R^2. They forget that their own efforts at excluding unsuccessful models should be taken into account. This is called the data-mining problem. Even in a theoretical framework with zero predictive power, careful structuring of the model can always generate nominal statistical significance to any degree you want, as long as you look hard enough.
Sven,
You might not want to go too far down the path of arguing against naturalness considerations. The fact that (as you yourself argued above) even the MSSM has an upper limit on the Higgs mass is only because someone like you has chosen to apply an upper limit to the stop masses.
Make the stop masses the GUT scale, and you can get essentially any Higgs mass you like (within reason of a perturbative quartic). Then your favorite “prediction” of the MSSM fades away.
SUSY has no reliable and clear physical motive. It has arisen only from desire some mad Russian mathematicians
http://ufn.ru/en/articles/2001/9/f/
to expand a class of renormalizable models of canonical QFT. But mathematic does not work if basic physics idea Not Even Wrong.
Pingback: Comparing the ATLAS and CMS SUSY Searches « Collider Blog
There are some comments in this blog asking to see the effect of the lastest ATLAS 0-lep susy exclusion data on the CMSSM fit … if you are still interested, see the first paper on that topic here:
http://arxiv.org/abs/1103.0969
(btw – I should declare a self interest as an author in the above)
Thanks Christopher,
If I’m not mistaken, figure 3a in your paper answers the question I’ve been asking about comparing the ATLAS exclusion results to earlier estimates of SUSY parameters. It looks like very roughly half of the probability-weighted parameter space is now excluded (although a much smaller fraction of the size of parameter space). Presumably over the next year or so we’ll see results that cover much of the rest…
Thanks for taking the time to discuss and share this with us
Physicists joke again:
Extra dimensions of space, for instance, which are predicted in many forms of string theory (a variant called M theory requires 10 spatial dimensions rather than the familiar three), could be accessible at the Large Hadron Collider (LHC), said Barnard College physicist Janna Levin. In the LHC’s unprecedentedly high-energy experiments, some products of particle collisions could go missing, having vanished into those extra dimensions.
http://www.scientificamerican.com/ar…SA_DD_20110310