TWiki> Main Web>UPG>FTR18017Review (2018-10-15, MartinFlechl) EditAttachPDF

Questions and answers for FTR-18-017

General

Cadi line : link

Color code

For the answers, the following color code is used:
* green - comment answered
* blue - comment requires further work to be addressed
* red - comment acknowledged and arguments are given why we do it differently.

Comments from Roger Wolf:

  • [28] confirm that the discriminating variable is the transverse di-tau mass. We call it mTtot with corresponding definition in HIG-17-020.
Thanks - indeed this was not clear, changed to "total transverse mass" and added reference to HIG-17-020.

  • [29] States 3.2 TeV as upper bound for the limits. This might be true for the model independent plots (can not fully judge from the log scale on the x axis). But for the model dependent limtis it's definitely not true. Why do the model dependent limits only go up to 2 TeV?
Simple answer: The LHC HXS WG numbers only go up to 2 TeV. We could add "and show MSSM limits up to mA=2 TeV" but I think that would raise more questions than it answers... Technically the sentence is still correct, all masses are considered (the H boson could even be much heavier than 2 TeV, though we know in practice this does not happen) but then the plot stops at mA=2 TeV because we lack inputs...

  • [68-71] This had been raised by Andreas Meyer already. I think the paragraph is still not OK. It's clear what you want to say and what the issue is. If you include h in the signal the whole mA-tanb is be excluded due to discrepancies in the description of the h couplings. But the way this is phrased ("exclusion would be artificial...") is by far too negative and even misleading. While what you want to say is that it were wrong to include the h in the projection, it reads like it were wrong to include the h in our searches. I think a) there is no need to state that the whole mA-tanb plane would be excluded when including the h in the signal; and b) as of now it sounds like it were per se wrong to include the h in the search, which it is not -- in contrary it is correct! I think this is a subtle and non-trivial point that mist be phrased clearly. You should phrase it somehow like this:
""" We exclude the h from the benchmark scenarios for the following reason: With increasing luminosity the search will become sensitive to the h. But the current benchmark scenarios do not incorporate the properties of the h boson within the required accuracy. We anticipate that the benchmark scenarios will evolve with time in this respect. We therefore restrict the signal to the heavy A and H bosons, to demonstrate the search potential only for these. """ Indeed if the compatibility w ng. So CMS would give inputs to the theory community that ATLAS will not be able any more to provide.

Very good point, thanks, implemented as suggested!

  • [88] vac-uum --> is this correct?
It actually is, but it is indeed awkward so we suppressed the hyphenation for "vacuum" now.

  • [91/92] "try to observe.." --> is this phrasing correct?
Changed to "Generally, MSSM scenarios incorporate the $125\,\GeV$ Higgs boson as the lighter scalar $\textrm{h}$ and are compatible with the current experimental constraints on its mass for at least a significant portion of the $m_{\PA}-\tan\beta$ space."

  • [95] this sentence is misleading. Do you translate th results of the model independent limits into the ma-tanb plane, or does he do a consistent likelihood analysis?
Of course the latter, sorry for the sloppy formulation. Changed to "The results are interpreted in terms of these benchmark scenarios based on the profile likelihood ratio of the background-only and the tested signal-plus-background hypotheses."

  • [99] and Conclusions: the numbers given are not consistent with/the same as given in the abstract.
Fixed (avoided rounding in the abstract as well).

Comments from Isobel Ojalvo on v2 ( 2018-10-08):

Show Hide

  • Please change 'heavy Higgs' to MSSM Higgs everywhere
Done, in all places where this is not too ambiguous (as we show results both on model-independent heavy Higgs searches and on MSSM interpretation of these results it is necessary to be precise and call it "heavy Higgs" in some places.)

  • L1 In 2012, a Higgs boson was discovered by the ATLAS and CMS collaborations [1-3]. Since then, a large number of precise measurements of the properties of this particle as well as searches 4 for additional Higgs bosons have been performed with LHC Run-I and Run-II data. -> The discovery of a Higgs boson was announced by the ATLAS and CMS collaborations in 2012. Since then a large number of precision measurements have been made which have shown the new particle to be Standard Model like within statistical and systematic uncertainties. However, there are many arguments in favor of theories which predict the existence of supersymmetric particles [cite]. A minimal extension of the SM consists of 2 Higgs doublets which predicts the existence of five higgs bosons, H, h, A, H^{\pm}. (Define each). Analyses which search for the SM Higgs boson have been performed using the 2016 dataset with 36.1 fb-1 of data and have shown no excess beyond the SM background. However, so far the LHC has only collected a small fraction of its lifetime target integrated luminosity.
Done, with slight modifications in order to incorporate also the suggestions by Keti and Jean-Baptiste.

  • p.3: L39 "concerning" -> to project. Try to avoid the use of the word "concerning" as it is ill defined.
Done.

  • p.4: L53 Uncertainties on the estimate of the background due jets misreconstructed as Ï"h which are not of a statistical nature are scaled to a minimum of 50% of the 2016 value. The SM processes where a jet is reconstructed as a tau are modeled as a single background using a jet to tau fake factor estimation method. (Citation) in the primary analysis this uncertainty is not limited by statistical uncertainties and therefore the systematic uncertainty is scaled to ... (Please check this is reasonable by contacting Roger)
Done. This has been made clearer, following comments also from Keti. The FF method has both uncertainties which are "real" systematics and which are statistics-like (stat. unc. in control regions). Only for the former we set a minimum value, while the stat. unc. are treated like all other stat unc.

  • p.4: L57 for limits where the Higgs boson mass is above 1 TeV
Done.

  • p.4: L71 production mechanisms are shown in...
Done.

Done.

  • p.6: L97 Even at low mass improvements -> Even at low mass, improvements
Done.

  • p.6: L101 have been shown using the 2016 dataset corresponding to a total integrated luminosity of _ fb-1.
Done.

  • p.6: L104 roughly -> approximately
Done.

  • L105-L106 quote the corresponding tan beta value for 2 TeV for each scenario. Instead of saying e.g. and moderately
Done.

Comments from Ketino Kaadze on v2 ( 2018-10-04):

Show Hide

  • Abstract: L2: results presented à results are presented
Done.

  • L5: “standard model” à “the standard model”
Done.

  • L13: “taken” à “collected”
Done.

  • L14-15: “In this publication” ambiguous à “In HIG-17-020”
Done.

  • L28: “expected amount collected” -- can you change it somehow?
Done (as suggested by JPS).

  • L29: Sentence “The upgrade of …” à “The upgrade and expected performance of the CMS detector is described …”
Done.

  • L34: “model independent limits of” à “model independent limits on”
Done.

  • L37-39: These two sentences are repetitions of sentences in the introduction. Can be dropped.
Done.

  • L46-47: This sentence needs rephrasing. Currently it is not very clearly structured.
Done.

  • L53-55: You state that uncertainty on jet->tau fake background is not of statistical nature in 2016 analysis. Statistics in control region from where fake rate and/or estimated fake background is not propagated as uncertainty on backgrounds?
These uncertainties have both a statistical and a systematic component. The text states that those components which are not of systematic nature are assigned a minimum value of 50%. This has now been made clearer in the text.

  • L46-59: The amount how you scale different uncertainties – is it dictated from some other documentation about upgraded CMS detector performance or different object performance? If so, should not you cite that documentation here?
This is a general prescription for the HL-LHC yellow report. Currently it is only documented in internal twikis and talks. For the section in the yellow report, we can then refer to a common section.

  • L60-68: These are a bit vaguely stated. Could you add some more clarity, perhaps extending the discussion a bit would help as somehow it is cut too short.
We have extended the explanation, hope it is better now. If not, it would be great to get feedback on which part of the explanation needs further information.

  • L24-26 says that signal hypothesis between 90 GeV -- 3.2 TeV are considered. L88-89 says that 125 GeV Higgs is considered a slight scalar Higgs (h). Are not these two statements contradictory?
Added "additional", i.e. "of additional Higgs bosons in the mass range of 90 GeV to 3.2 TeV", to avoid any apparent contradiction.

Comments from Jean-Baptiste Sauvan on v2 ( 2018-10-04):

Show Hide

  • abstract: since it is written "the reach is extended", it could be useful to mention the current mass reach in the 2016 analysis
Done.

  • l. 6: "so far" is used twice in a row which sounds a bit redundant. The second one could eventually be removed.
Done.

  • l. 15,16: "all the details ... are described in detail." sounds also redundant. The "in detail" at the end of the sentence could be removed
Done.

  • l. 16: "systematic uncertainties and different interpretations" -> "systematic uncertainties, and different interpretations"
Done.

  • l. 22: "This is in order to target..." could be rephrased. Something like "The goal of this categorization is to increase the sensitivity to the dominant production modes..."
Done.

  • l. 28: "luminosity of 3000fb-1, the expected amount collected" -> "luminosity of 3000fb-1, expected to be collected"
Done.

  • l. 29-31: There should be references to the more recent TDRs rather than the Technical Proposal.
Done.

  • l. 49-53: The sentence "Concerning uncertainties ... " is too long and hard to read. It could maybe be split in bullet points listing each case.
Done (bullet list).

  • l. 53: "Uncertainties of the estimate" -> "Uncertainties on the estimate"
Done.

  • l. 54: "due jets" -> "due to jets"
Done.

  • l. 57: "independently of luminosity" -> "independently of the luminosity"
Done.

  • Fig. 1: The uncertainty scenario used here should be written on the plots and caption.
Done.

  • l. 96: "intermediate tan beta": it would be useful to add a typical tan beta value or range, in order to better define "intermediate"
Done - added "and $\tan \beta$ values of about 30, depending on the scenario."

  • l. 100: "HL-LHC Projections" -> "HL-LHC projections"
Done.

  • Fig. 3: Here also the uncertainty scenario used should be written on the plots and caption.
Done.

  • Ref. 1,2,6: Use capital letters for GeV, CMS, LHC, HL-LHC, and ECFA
Done.

Comments from Andreas Meyer ( 2018-09-03):

Show Hide

  • My main question: Can we include in the PAS results (i.e. limits) for all three configurations? scenario 1 (Run-II uncertainties), scenario 2 (scaled with "floor values"), and stat only ?
Thanks for the feedback. We do not have strong feelings about whether or not to include these further plots. The main reason against is that they will look quite similar, anyway, as we are statistically limited. But of course there is also merit in bringing exactly this message across. We will bring the question of including these plots up during the pre-approval.

  • line 6: his -> its
Done.

  • line 6: suggest: target INTEGRATED luminosity
Done.

  • line 36: The phrasing "the statistical interpretation is extended in two ways" seems somewhat misleading. Better write something like: "The projection is performed assuming two different scenarios for the improvement of experimental and theoretical uncertainties with increased integrated luminosity"
Done.

  • line 44 and following: suggest to also give absolute values for the uncertainties. If I understand correctly, tau-ID unc. will be 2.5% ?
Done for most uncertainties, except for those which differ strongly for different channels and kinematic regions.

  • line 59: I did not get the connection between the foreseen continuous improvement of the H(125) measurements, and the statement that this ensures that the exclusion power for a heavy H comes from the high-mass region.
Rephrased, and hopefully now clearer: "At high integrated luminosities the MSSM limits become sensitive to the small differences predicted in properties of the already discovered Higgs boson with a mass of 125 GeV. Thus if the lightest Higgs boson, h, is included in the SM versus MSSM hypothesis test, this would result in a complete exclusion of the $m_{\PA}$-$\tan\beta$ parameter space. This exclusion would be artificial since as integrated luminosity increases and the properties of the 125$\,\GeV$ Higgs boson are measured to increasingly high precision, the MSSM benchmarks are expected to be updated to retain consistency with the experimental observation. To avoid this, only the presence of the two additional neutral Higgs bosons H and A is tested. This procedure ensures that all exclusion power for a high-mass Higgs boson comes entirely from the high-mass region."

Comments from UPSG and HF conveners during UPSG status report (2018-08-29):

Show Hide

  • Please try to evaluate the maximum effect the b-tagging NP could have on the measurement uncertainties if it is unconstrained

When removing the impact of the b-tagging NP from the uncertainty, then assuming the impact to be 5 times larger (i.e. unconstrained, as it is constrained to 20%) and adding it to the total uncertainty, the uncertainty on the signal strength increases by 4.5 per cent. This value is very conservative as the calculation assumes no correlations and that there is no real constraining power in the measurement. Even so, the effect of "unconstraining" this NP is barely visible on any result plots and certainly smaller than already the intrinsic uncertainties of extrapolating by two orders of magnitude.

-- MartinFlechl - 2018-09-04

Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2018-10-15 - MartinFlechl
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2024 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback