http://cms.cern.ch/cds/SUS-14-022

Search for electroweak production of charginos in final states with two tau leptons in pp collisions at 8 TeV

Marco M

General comments on the draft

The aim of this article is a careful description of a search for SUSY charginos in the tau-tau final state by using a sample of pp collisions at 8 TeV collected by the CMS detector in 2012. The first part of the paper, that introduces the search variables and the definition of signal regions and the corresponding event selection criteria for different channels, is quite clear. Slightly harder is the section 7, describing the several background contributions, the methods for their estimation and validation, and their uncertainties. A huge amount of work is summarized here, so we can understand that it is not easy an easy explanation of so many details. However, the obtained results are interesting, the analysis methods are thorougly described, and overall the paper is very well written.

A few comments on specific parts of the writeup are given in the following:

Abstract (second line) 18.1 to 19.6 fb^-1 Why not change to: 18.1 and 19.6 fb^-1 in different channels as (better) expressed in line 27 of the Introduction?

Introduction line 29 two massive particles –> sparticles (I THINK…) (since you are defining the “stranverse mass” variable M_T2) line 118 at NLO+NLL —> at NLO+NLL accuracy line 131 can be written as —> can be defined as line 138 So, the mass of a visible final state lepton is set to zero also for a final state tau? line 139 also set neutralino mass=0 (however along the text sometimes you use also different values of neutralino mass in SUSY models, as 1 or even 40 GeV in Fig.2, 4 and 7. line 145 The distribution of M_T2 reflects… and is much higher —> has a longer tail line 167 models with small (Deltam) —>small Deltam. Table 1, pag. 6: for sake of clarity and consistency with the text of sect. 5 and 6 maybe it would be better to move below the pTmiss>30 GeV requirements, close to the DeltaPhi and MT2>40 GeV requirements. Furthermore, an additional line in the table framework could separate the pre-selection requirements from the additional requirements to select the signal regions (last three lines). section 7.1 Here there is apparently some confusion about the level of tightness of the tau_h isolation requirement. At line 210 you quote a “loose” selection, as a “less stringent” isolation requirement with respect to a somehow tighter selection criterium. However, within the body text in Fig. 3, when defining the control and signal region, you only quote a “medium” requirement: is it the tightest one or not? Also the sentence used to define Control Region 2 and 3 is not clear: “one leg passing loose but not medium isolation and other leg passing loose isolation” Does it mean both legs pass only loose isolation or, maybe: at least one leg is not passing medium (but only loose) isolation? line 225 distribution as the function —> as a function section 7.2 Second line (not numbered) after the section title: the sentence “The simulation is validated using a data control sample” seems redundant and in a wrong position, since it is not directly connected with the following formula and text lines, but rather with the last paragraph of the section: “The W+jets simulated sample is validated in data sing a same-sign mu-tau←h control sample, ” Section 9 line 399 … are shown with the red solid and dashed lines, respectively. There are no solid lines here, but only dashed: thicker for the expected limits and narrower for the +/- 1sigma contours. line 401 The signal cross sections in NLO+NLL order —> at NLO+NLL order line 407-408 Here it is not clear why the lepton+tau_h channels “do not improve the results”. Can you clarify?

Slight inconsistency: line 416 The observed (expected) upper limit….. is almost TWO times larger than the theoretical NLO cross section. vs line 467 … the limits are more than THREE times larger than the theoretical NLO cross sections, even for a massless neutralino. Eventually can we add a reference to the theoretical NLO cross section calculation? line 436 Table 9 shows efficiency in different channels —> in all channels line 436 … to pass the pTmiss>30 GeV as a function —> to pass the pTmiss>30 GeV requirement as a function….. Table 9, pag.17 (text of the caption): different —>all line 452 SMS plane: explain the acronyme SMS for not-expert readers.

Linda

line 25: dataset –> data set (PubCom recommends two words) lines 81-85: It is not clear to me what “measure of isolation” means. How is it defined? Why is it measured in GeV? line 90: and di-boson are –> and diboson, are line 96: a pair of charginos () are produced and decay –> a pair of charginos () is produced and decays line 104: there were on average–> there are on average line 107: dataset –> data set line 117: di-boson –> diboson line 132-134: What does “correct neutralino mass” mean? The one predicted by the theory? line 150: loosely identified, isolated –> loosely identified and isolated line 158-159: To reject Z→tautau… are required –> Moreover, to reject Z→tautau…. are also required line 163: pre-selection –> preselection line 177: were collected –> are collected line 208: a set of .. are defined –> a set of .. is defined lines 241 -243: it is not clear to me what is the subject of this sentence… line 392: it is not clear to me what “better” is referring: is it “better” with respect to the previous analysis? line 418: a simplified SUSY model was used –> a simplified SUSY model is used line 421: The number of the remaining signal events: “remaining” from what? line 452: the efficiencies were used/ the results were –> the efficiencies are used/ the results are

Roberto

[Congratulazioni, vere perchè sia la stima dei fondi che l'interpretazione mi sembrano fatte davvero bene]

MAIN COMMENTS ON PHYSICS AND STYLE: - The main physics comment is about the meaning of small/large m(chi0) and small/large Dm thorughout the text. In specific places: 1) Introduction: it should be mentioned that this search is particularly targeted for very low neutralino masses (and/or) large Deltam. Deltam is currently mentioned only at line 165. Also it should be clarified what “small” means. 2) Samples: L96: please give information on the range of chargino and neutralino masses used in the scan. 3) L139: “We set mchi0 = 0”. However, neutralinos with large masses are also considered in the scan (up to 350 GeV). In this case this hypothesis gives a MT2 very large from the true one. Is there any impact on this variable? - Visibility of lines in plots: while the plots are done in a very accurate manner, the choice of line styles in a few cases affects readability: 1) Figure 4: Signal lines are grey and very finely dashed, which makes them very hard to see even in color printing. Suggestion: use red for SUSY lines and a less outstanding color for Higgs (e.g. grey) 2) Figures 5 and 6: Text and caption refer to “THE diagonal line”. Aside from the fact that there are two diagonal lines, these are very hard to see in color and absolutely impossible in b/w printing.

MINOR COMMENTS L112: “the most accurate calculations”: however, the recent MATRIX tools by Grazzini et al. provide NNLO cross-sections for dibosons. We suggest removing this clause and just state the QCD orders used in each case. Eq. 2: How is the minimum taken? A finite value of pT1 is scanned, computing pT2 and the value of mTi? L145-147: They seem to repeat (and in a less precise way) what is stated already in the paragraph above. They would be most suited for the introduction, e.g. in line 30. L166-167: “distribution can have A long tail” + remove brackets around the 2nd occurrence of DeltaM. L247: It is said that N(beforr final selection) is taken from a data control sample, but then they are referred to as “W+jet events”. Please clarify. L273: “Significantly low”. How much is significantly? Seems not the right word. Table 6: Too much use of shortcuts/jargon for a publication: “iso”, “id”, “MonteCarlo statisticS or statisticAL” Figure 4: If SR1 targets models with large Dm and SR2 with small Dm, why do you plot the same signal? Is this considered small or large? L363: How was 50% determined? Given that these are (N)NLO predictions, it seems far too large…

Valentina

The paper presents the results of an interesting and detailed search, revealing the amount of valuable work performed and providing deep information to allow interpreting the results with models other than The ones explored in the paper. However, the quality of the analysis exposure is poor, more suitable for an Analisys Note, rather than for a CMS paper. In general, a deep revision of the text is suggested.

General comments: - There are many places where the definite article 'the' is missing - Too much information is often used to describe the analysis technique, with no valuable benefit to the text exposure - Two different luminosities are used for the analysis, but no mention is made to which analysis cannel each luminosity value corresponds to - Words as 'should' or 'could' are used in the text, better to substitute them with more assertive expressions - The word 'kinematical' is employed through the text, but it's wrong. Use 'kinematic' instead - The active voice is widely used. Even if this choice is allowed by the CMS guidelines, the suggestion is to reduce the usage of 'we' and 'our' as much as possible - Tables always occur in the middle of paragraphs. Please, avoid.

In-line comments:

= Introduction L.28: stranverse = The CMS detector and event reconstruction L.55: around 100 kHz […] around 400 Hz → drop 'around' L.80: Additional discriminators are used to separate tau_h from electrons and muons. → Which 'additional discriminators'? Expand or drop L.85-87: A similar isolation algorithm is used in this analysis to separate leptons (e or mu) from tau decays from those arising from hadron decays within jets. → Similar? Expand or drop

= Monte Carlo samples L.91: In the following figures and tables, → Drop 'figures and tables' Bring to this Section the description of 'VV' acronym and remove it from the label of Table 7

= Backgrounds Figure 3: not clear what 'medium isolation' refers to, put a reference to the definition in Section 2 (L.85) L.220-222: a transfer factor, which is the y-intercept of a horizontal line fitted to the ratio of the number of events in CR1 to the number in CR2 in different bins of the low values of the search variables. → unclear, please rephrase the sentence L.228-230: The systematic uncertainties on the background estimates include the uncertainty on the validity of the assumption that isolation and MT2 or SUM(m_T^tau_i) are not correlated and the uncertainties on the residual SM backgrounds which are subtracted based on Monte Carlo expectations. → unclear, please rephrase the sentence L.239: extrapolation → extrapolating? L.241-243: Fitting the whole range of the low values of the search variables by a horizontal line or a line with a constant slope or using the value of the last bin before entering the signal region are examined. → unclear, please rephrase and better expand L.244: the uncertainty shown as “fit”� in the table. → the “fit” uncertainty. Section 7.2: it is difficult to get the values of Table 3 following the prescriptions given in the text, please better explain, in particular L.264-266 L.267: the systematic uncertainty and referred to as “shape”�. → the “shape” systematic uncertainty. Table 3, label: The “shape” takes […] → The “shape” uncertainty takes […] L.308: statistical uncertainties associated with the sidebands. → statistical uncertainties associated to the sidebands. Table 5, label: rFake (rReal) is shorthand for fake (real) rate. → drop the sentence L.312: uncertainties are considered fully → uncertainties are fully

= Systematic uncertainties In general this section has to be rewritten. L.321-322: They amount to 10-15% for backgrounds and 2-15% in different parts of the signal phase space. → please rephrase the sentence L.326-327: move the ')' after 4.5% L.338: The uncertainty ON → IN is used for all the other uncertainties in the list L.339-340: because for the backgrounds either the data- driven methods are used or the normalization is found from data. → please rephrase the sentence L.348-349: The uncertainties related to p_T^miss can arise from different sources e.g. the energy scales of lepton, tau_h, jet objects and unclustered energy. → please rephrase the sentence Table 6: label: 'next to their names' → drop id → ID iso → ISO sys. →drop Statistic → statistics

= Results and interpretation In general this section has to be rewritten. Table 7: please, rewrite the label Figure 4, label: For more detail, see the text. → redundant, drop L. 384: There is no excess of events over the SM expectation. → No excess is OBSERVED […] L.390-391: This should be compared to the ATLAS limit of 345 GeV [22]. It should be noted that the ATLAS results are based on the tau_h tau_h channel. → The results can be compared to the ATLAS limit set to 345 GeV [22], based on the tau_h tau_h channel only. L. 397 and Fig.5, label → m_tau or m_stau? L.399: 'red solid line' → where is the 'red solid line'? L.400 → 1 standard deviation (remove s) L.412: vs. → not acceptable, expand L.385-386/404-405: Why the Monte Carlo sample used to simulate the first production channel (Fig.1, left) is described in Section 3, while the sample used to simulate the second production channel (Fig.1, right) is not?

= Information to test new models This section is very important and useful for external physicists who want to interpret the results of the paper with models other that the ones presented in the paper, but the description is very poor and confusing. It has to be improved.

= Conclusion L.460-461: there is a mixture of experimental observation and interpretation in the first sentence. Please, decouple it. First describe the results and then their interpretation. L.463: Where DeltaM is defined? Please, recall L.467 stau pair → is the first time the stau symbol is not used (except of Fig.1, label) L.467-468: but the limits are more than three times larger than the theoretical NLO cross sections, even for a massless neutralino. → unclear, please rephrase the sentence