RS/WPrime search in the VZ channel (EXO-12-014 CWR Review)

Disclaimer: Questions read and answered: green. Answer in black.

Questions read and answered, modifications not implemented on the note yet: orange. Answer in black.

Questions read and not yet addressed/pending of discussion: red. Answer/discussion in black.

Comments in blue. Comments in a comment already discussed but still controversial, in at least one of the author's point of view.

Twiki

EXO-11-061 Results

EXO-11-081 Results

Useful links:

Table of contents

EXO-12-014 Review

Post-FR editing

Detector description: add an introductory sentence at the beginning of the paragraph, e.g. "Here a brief description is given with an emphasis on the elements mostly relevant for this analysis. A more detailed description can be found in…"

Done.

Tables 6-8: a. Change the captions of Tables 6-8 to reflect that excluded cross-sections are on sigma(W'/RS) x BR(W'/RS -> WZ/ZZ), but do not include the W/Z decays into leptons/jet. b. Add total (ie. stat + syst) uncertainties to signal efficiency column.

Done.

Introduction: Add comparison to previous experimental results at the end of Section, e.g. "Previous searches have been made", etc. Include CMS/ATLAS 3l+neu, gamma gamma, top-bottom (EXO-12-001) results. Probably have to remove discussion on 3l+neu later in the paper (e.g. line 330) if overview of previous results is going to happen in Introduction.

Done.

Figure 4 (brazilian-flag 1D exclusion plot): Two requests here: a. Interpolation-by-eye between mass points appear to give different SSM exclusion limit than what quoted in text. Please cross-check and/or fix. [Post-meeting note: confirmed that this is a ROOT "feature" when drawing lines on a log scale, ie. ROOT is plotting lines connecting two points on log scale incorrectly. Trying to find acceptable workaround.] b. Please add 3l+neu observed exclusion curve in order to show for which mass region current analysis becomes competitive.

We thank the PubCom for pursuing this. While investigating the ROOT feature, we realized that our previous linear interpolation of the theoretical curve was inadequate. We improved the interpolation by including many more points, which yields a more reliable result (albeit the exclusion limits on the mass for LO and NLO are now a bit lower).

We have added the 3l+neu observed exclusion curve, as requested, which demonstrates that this analysis becomes more sensitive for higher masses.

Add JES citation: use 2010 paper (only one available)

Done.

Long discussion on 15% uncertainty attributed to rho (propagating to background estimate). Authors believe calculation is correct but requires better explanation. Brief summary of the issue can be found at: https://twiki.cern.ch/twiki/pub/CMS/EXO12014CWR02/BackgroundEstimation_EXO_12_014.pdf PubCom would like to cross-check this claim more thoroughly before deciding on course of action.

The corresponding section (6.1.2) has been rewritten to describe better the calculation of the rho and background uncertainties, as agreed with the PubCom.

Fig. 5 (2D RS exclusion plot): remove gammagamma/ee CMS results (figure currently showing TeVatron gamma gamma/ee results, but LHC results are more competitive). Maybe add dijet results if it is easy to overlap?

We have removed the TeVatron gammagamma/ee results as requested. The (CMS) dijet results have not been added, since those are reported in a PAS, and not a paper.

Systematic on trigger efficiency: Remove "Even though". Add "However, we observe…..due to Brem"

Done.

Find recent paper with "more complete description of isolation". Copy + Paste here.

Done

L100-102: please move the sentence "the requirement that... from jets" to Sect. 4.1, as you describe a cut, not a reconstruction technique.

Change Title: "Search for exotic resonances decaying into WZ/ZZ in pp collisions at sqrt(s)=7 TeV"

Done.

Abstract: remove "for two theoretical models" altogether.

Done.

Introduction changes: Line 13: TeV -> TeV scale Line 14: Rephrase as follows: "This analysis is sensitive to searches for resonances starting at 700 GeV and above. However, there are other theories that predict light resonances (e.g. low-scale technicolor)…." etc

Done.

"This analysis is sensitive to searches for resonances starting at 700 GeV and above. However, there are other theories that predict light resonances (\eg low-scale technicolor)~\cite{TC1, TC2}."

Line 19: decayS Line 20: Keep V in roman, remove italics in other instances of V boson line 30: a data sample line 33: Sequential Standard Model

Done.

Conclusions line 352: add comma after hadrons line 357: several -> two

Done

More figure changes: Fig.2 + 3: put bar over M_pi Fig. 4: BR --> replace by B

Done.

To be decided on FR

Title (to be discussed at the FR meeting): "Search for exotic resonances decaying into WZ/ZZ in pp collisions at sqrt(s)=7 TeV" [the title of a search paper must tell the reader what you searching for and where. No need to put how you search the signal in the title]

line 45: starting the description of the detector with that sentence sounds, to me, as "we don't want to loose our time explaining you how our detector is made. If you really want to know it, read this paper and do not bore us!" :-). I prefer having it at the end (of course, is a matter of taste, this is just a suggestion).

* Tables 6-8: I suggest adding the uncertainty on the efficiency to this table. That way, one will see at a glance how the total stat + syst. uncertainty varies at each mass point for the two MC inputs (background and efficiency). If you do this, could also consider dropping Tables 4-5. The latter are very long and contain more information than would interest the average reader. They are also a bit misleading, because, for example, one sees systematic uncertainties of 72% quoted on the background in Table 4, which sounds really scary, but only later when one sees Table 6 does one realize the background at the corresponding mass point is very small, so that this big systematic doesn't do much harm. My personal preference would be to keep Tables 4-5, but greatly simply them, perhaps just showing the breakdown of systematics at a couple of illustrative mass points.

* L100-102: consider moving the sentence "the requirement that... from jets" to Sect. 4.1, as you describe a cut, not a reconstruction technique.

* L192-193: These sentences should be moved to Sect. 6.1.1 ?

title suggest "a jet and either a lepton pair or missing..."

line 97. Most papers have a much more complete description of isolation

I was surprised that there is no figure in this paper showing the jet mass for the lepton channel. I think it would be a nice addition. maybe around line 171 you could show this distribution? and mark the "NS" and "SB" regions? and maybe show the M_VZ distribution fo the SB selection?

Table 6: no error quoted for the efficiency. Quote it and write it with the appropriate number of digits.

line 111. need a citation for the JES corrections (and they use more than just dijets and gamma+jets) --> The Detector Performance Note with 2011 data is not yet public. The link to CDS is here: https://cdsweb.cern.ch/record/1454659?ln=en The paper with 2010 data is here: http://iopscience.iop.org/1748-0221/6/11/P11002/

(58) Table 5 (now Table 4 in v2), (a) in the 2nd header row, the 2nd column, to be consistent with the expression in the 1st column, the capitalization should be changed from "Mass Window" --> "Mass window"

Comments for Final Reading

Comments by Giovanni Organtini

Title (to be discussed at the FR meeting): "Search for exotic resonances decaying into WZ/ZZ in pp collisions at sqrt(s)=7 TeV" [the title of a search paper must tell the reader what you searching for and where. No need to put how you search the signal in the title]

A general comment: please excuse me if I am so pedantic about this subject, but I continue thinking that quoting limits with three significant digits, when the error bands on the expected limit are so large, is unreasonable. In the abstract, for example, I would have said that the limits are between 700 and 940 GeV in one case and 750 and 930 GeV in the other.

Done

Type A


line 16: "this": which one? maybe "in this case"?

Done.

line 18: "i.e." --> "namely" ("i.e." implies that the following are all the possible leptonic modes of the Z decay, but in fact, despite the fact that apparently nobody remember that, tau's are leptons, too)

Done.

line 22: "close the" --> "close to the"

Done.

line 35: remove "Large Hadron Collider" and use just LHC (already defined at line 11)

Done.

line 45: starting the description of the detector with that sentence sounds, to me, as "we don't want to loose our time explaining you how our detector is made. If you really want to know it, read this paper and do not bore us!" :-). I prefer having it at the end (of course, is a matter of taste, this is just a suggestion).

To discuss during FR.

line 65: "e.g." --> "i.e." (e.g. stands for "exampli gratia", so it's not exhaustive)

Done.

lines 72: "several" is used also in line 73. Use a synonym avoid repetitions.

Done.

line 75 (and following): in this paragraph all the names of the software packages must be written in small caps fon, using the right case -upper or lower-. See also caption of Table 1. Search for those names throughout the paper.

Done.

line 124: "pairs" --> "pair"

Done.

line 157: do not start a section with the title of a subsection. Put some line of introduction here.

Done.

line 175: "[" -> "(" and "]" -> ")"

Done.

line 192: comma missing between "estimation" and "are"

Done.

line 223: "these uncertainties": which ones? It seems you are talking about eliminated systematic effects. Since you eliminate them, you don't need to describe them, no?

Changed to:

In the following, we consider the remaining relevant uncertainties in detail for the dilepton and \MET channels.

line 227: "that region" --> "the corresponding region"

Done.

line 259: "and trigger" --> ", the trigger efficiency"

Changed to:

, and the trigger and reconstruction efficiencies.

line 267: "around 2-9" --> "between 2 and 9"

Done.

line 272: "effects in" --> "effects specific for"

Done.

Type B


Abstract last line: k and M_Pl are undefined (user defined) symbols. Not suitable for the abstract. The sentence may be rephrased or the ration should be "described" in words.

Changed to "coupling parameter"

lines 27-29: here you talk about isolation, but we never said before that leptons are expected (or must be, for selection purposes,) isolated. Also, in this paragraph there is a mixture of matter suited for an introduction, and matter regarding the experimental procedures. I would prefer to separate things.

Done. Moved the isolation discussion under Reconstruction and Event Selection.

line 78-79: "For the GKK analysis...and the diphoton final state" I could not even figure out what that sentence means. In particular, while I can understand up to "slicing method", the inclusion of the "diphoton final state" has no meaning for me: is there something missing? By the way, I assume that "two cutoff phase space slicing method" is a name for a method (I will check the referenced paper later to understand it). But, if so, you should quote that name in "...".

Changed to:

Mass-dependent $K$ factors are applied. For the \GKK~analysis, next-to-leading order (NLO) corrections are calculated using the ``two cutoff phase space slicing'' method~\cite{Kumar200945,Kumar200928} in the diphoton final state. For the \Wprime analysis, the next-to-next-to-leading order (NNLO) corrections are calculated with {\sc Fewz}~\cite{Gavin:2010az} in the leptonic final state. These $K$ factors are used for lack of better (N)NLO calculations for the final states considered.

Table 1: no errors given for cross sections in the table. Irrespective of the fact that those errors are useful or not, here we report about measurements (or computations, it doesn't matter), hence they deserve quoting the corresponding error.

We have followed the tradition of not including any theoretical uncertainties in the quoted signal cross sections (with the exception of the PDF, which is a rather experimental quantity). The signal cross-sections are only listed as a reference (along with the K-factors) and no assumption on the precision of these values is (should be) assumed. To the best of our knowledge, this is a common practice followed by theorists when the experimental results are interpreted.

line 92-93: "primary" and "leading" deserves some clarification. It depends on what you intend for "primary". In my understanding, primary vertex is a term reserved for the vertex considered the one from which the hard scattering process took place. In this case there is one primary vertex per bunch crossing. In fact, the term can easily be intended as any pp collision, then there could be many "primary vertices" in a bunch crossing. In most cases the difference is irrelevant, but in this case. You have then define what is a primary vertex. In any case you need to define what is a "leading" vertex. What makes a vertex the "leading" one? The sum of p_T or p_T^2? The number of tracks? Its coordinate along the beam axis?

We changed to:

Charged hadrons that are consistent with primary vertices other than the leading one (defined as the vertex with the largest sum of track $p_T^2$) are removed from the collection of particle candidates used to reconstruct the jets, to mitigate the effects of multiple proton-proton interactions within the same bunch crossing (pileup).

line 113: "energy" --> "momentum" (energy is not a vector)

Done.

line 114: "E_T^miss" --> "p_T^miss" (according to comment above)

Done.

Table 2: see comment on table 1.

See answer on the comment on table 1.

line 130: "lepton directions": which one? Are there two DeltaR (one per lepton) or one DeltaR (built around the pair momentum)? Make it clear.

There is one deltaR per lepton. Changed to:

We require the jet to be well separated from the two leptons forming the Z candidate: $\Delta R \equiv \sqrt{(\Delta \eta)^2 + (\Delta \phi)^2}>1.0$ for each lepton, where $\Delta \eta$ ($\Delta \phi$) is the pseudorapidity (azimuthal) distance [...]

equation at page 5: in the cosinus, "E_T^miss" --> "p_T^miss"

Done.

Figure 1: indicate the two figures with (a) and (b). One of them is referred later in the paper (line 245, change this accordingly) and, consistently with the other figures, you must have subfigures.

The figure does have a and b on the top. Should we move them to the bottom? Also, should we also add a and b to Fig. 2? And change the subtitles and all references in text to a, b instead of left/right?

NB: We were explicitly told in the CWR to change these figures to their current form. The rationale for this was that the figures should be self-contained. In any case, we have switched to using subfloats now, and changed the later reference in the paper.

lines 172-174: "of the total ...selections:" --> "as" (as far as I understand, how the ratio is built is described in lines 175-176, no need to repeat it).

Changed to:

The procedure is as follows: we first produce the MVZ distribution for the sideband selection. We define the ratio \alpha(MVZ) as the total number of Monte Carlo background entries in the MVZ spectrum with the nominal (65 < Mj < 120 GeV) and sideband (30 < Mj < 65 GeV) selections:

\alpha(MVZ) = NNS(MVZ) NSB(MVZ)

where NNS(MVZ) [NSB(MVZ)] is the number of events in the signal (sideband) region contained in a bin of the VZ mass distribution centered at a given value MVZ.

line 180: "function and": add "f_A(M_{VZ})" between "function" and "and" to define the symbol.

Done.

Figure 2: as far as I understand the continuous line is not the expected background, but the fit to the data, that in turn is used to derive the expected background.

The continuous line IS the expected background, as determined by the fit of the data in the sideband region (not shown in the figure) and NOT the fit to the nominal data (that is shown in the figure). The expected background in a given mass window is the integral of the continuous line in the same range.

line 210: I probably not understood how rho is computed. My computation, according to my interpretation, gives 0.43. At least, the method description must be improved. Also, the error on rho (0.06) seems much larger than expected from the errors of the quantities from which it is derived. Reading the paper I understood that this number comes from considerations in 6.1.2, but at this stage this is not clear at all.

The difference comes from rounding errors. The inputs from rho are indeed the 4th column of Table 3, but in that column they were rounded. We tried to add more significant digits to the table, but in one of the previous review rounds it was pointed to us that it would not make sense to have extra digits with the errors quoted. We settled on the format currently seen, which leads to the small rounding error observed.

To make the text more clear, changed to:

"The 𝜌 parameter is estimated from the simulated SM samples by rearranging Eq. (1) in the following way: (Equation goes here) and setting the values of $N_A,\ldots,N_D$ to the ones from the SM prediction. Using the values reported in Table~\ref{tab:yields_fourRegions} gives $\rho = 0.42 \pm 0.06$. The quoted uncertainty includes both the statistical uncertainties of the simulated SM samples reported in Table~\ref{tab:yields_fourRegions} and the systematic uncertainty discussed in Section~\ref{sec:syst_MET_channel}. The value of $\rho$ derived in then reinserted in Eq.~(\ref{eq:estimatedBackgroundCorrected}). By setting $N_B,\ldots,N_D$ to the yields observed in the data, we obtain an estimate for the remaining background $B_{\text{est}}$.

Figure 3: in the figure you use MET for transverse energy, while in the body it is called E_T^miss. Be consistent.

Fixed both figure and caption.

line 234: symbol N_{+Bgd} not defined line 235: symbol N_{-Bgd} not defined

Changed to:

The yield differences, in each mass window, between the expected background with the positive (N+Bgd) and negative (N−Bgd) jet energy scale variation with respect to the nominal selection and fit are taken as the +/-1sigma estimates for the JES systematic uncertainty.

lines 237-239: very difficult to read. Maybe (partially) using formulas make it very clear.

This has been rephrased. Hopefully it is now clearer.

result for B_{est}: my estimation, based on numbers given in the paper, gives 155 \pm 23. Can you check it? If there is any "subtle" difference in the computation, w.r.t. the details given in the text, these should be described.

Done. The value 153 is correct; again, the difference w.r.t what you found is due to rounding errors. The \pm 29 value was a typo. The correct value is \pm 24 and, again, the difference is due to rounding errors.

line 294: start the section saying that you do not observe any significant excess over the expected background.

Done.

line 299: sigma(VZ) and B not defined. I frankly do not understand why you should quote that number, since you know the B(Z->ll) and B(Z->nunu).

That was rewritten and the B to which we report the limits was modified.

line 300: see above

See above.

line 307: "\sigma \times B" --> "cross section times branching ratios"

Done.

line 314: what means "comparable"? Complementary? Similar? Not clear to me.

Changed to "complementary"

line 332: "dielectrons" --> "electrons", "dimuons" --> "muons"

Done.

Please also check that acknowledgements are consistent with those published on https://twiki.cern.ch/twiki/bin/view/CMS/Internal/PubAcknow

Changed to the last version in the mentioned Twiki.

Figure 4: the process under study cannot be inferred from the figures. You may add a caption inside them or rewrite the y-axis label.

Done.

Figure 5: move the legend a little bit higher in the body of the figure (top aligned with the integral top)

Done

ref 33, 34: you used MadGraph v5 for your analyses. Put the correct reference, here.

This was fixed in v3.

Table 4: first line of caption, remove "a" between "for" and "given". Also, in line 4, remove last "and" and substitute it with a comma. Many errors are quoted with an exaggerated number of figures (e.g. 108 --> 110, 10.2 --> 10, 11.1 --> 11, 125 --> 130). I would remove all the "%" signs in the table, putting it in the header.

Done.

Table 5: see table 4 for "%"

Done.

Table 6: no error quoted for the efficiency. Quote it and write it with the appropriate number of digits.

We changed the number of significant digits. We would like to discuss the efficiency uncertainty during the FR.

Table 7: see table 6.

See answer in table 6.

Table 8: see table 6. Also, the number of digits used to quote the limit seems to be exaggerated! You quote up to 5 significant figures, while from the figures it is clear that even two or three, at most, are meaningful

See answer in table 6. We changed the limit to 2 significant figures.

Table 9: see table 8

Done.

Comments by Ian Tomalin

Type B (Physics/Clarity) =================

* General: I find it bizarre that you don't simply quote limits on sigma(G_kk --> ZZ) and sigma(W'-->ZW). i.e. Remove the branching ratios of the W & Z bosons to jets/leptons. These W/Z branching ratios are known to a few parts per mille, so you could do this without degrading your limits. And your limits would then be easier to understand and compare with other results. For example, in Tables 6, 7, 8 and 9, you present limits on 8 different quantities ! This is confusing for the reader. And also means that if the reader wishes to know which of your channels are most powerful, they have to factorize out the W/Z branching ratios themselves. Please place limits on sigma(G_kk --> ZZ) and sigma(W'-->ZW).

This has changed: we now quote limits on sigma(G_KK->ZZ) and sigma(W'->WZ). Figure and table are both fixed.

* General: From L314-315, I understand that you limit on W' bosons is weaker than an existing published CMS limit, so your new result is simply a cross check of the old one ? Cross checks are useful things. But it is a shame that the two results could not have been published (and perhaps their limits combined) in the same paper. In any case, if CMS has a stronger existing limit in this channel, this should be clearly stated in the Introduction to the paper, and possibly in the Conclusion too. We do not wish to give people the misleading impression that this paper provides the best W' --> WZ limits from CMS. In addition, I note that in Fig. 5, you compare your graviton limits with the weaker ones from existing publications. But in Fig. 4, you do not compare your W' limits with the stronger ones in Ref [53].

You are absolutely right here, except for the fact that Fig.5 is NOT a comparison to other analyses having produced weaker limits. Fig.5 shows that our limit is complementary to those obtained with other searches, as a function of the mass and the coupling. We cannot produce a plot like this for Fig.4 since there is not such a complementarity to show in that case. In this sense, you could also imagine that our W' result is not our main result. But we think it is important to include in the paper. However, for W', at higher masses our limit becomes more competitive. Note the plots are not directly comparable because of the different Y axis (WZ-> 3lnu has sigma x BR(W'->WZ->3lnu) while the WZ->ll(nunu)+jet analysis has now the sigma x BR(W'->WZ)).

Should we add in the text something to make more clear that our result in W' is complementary to the trilepton channel, which has a better mass limit for SSM?

* L6-8: You say that `the SM is valid below a new physics scale Lambda, in which case one expects new resonances'. With this form of sentence, the reader would expect the second part of the sentence to follow in an obvious way from the first, which is does not. To me at least, it is not obvious that if the SM breaks down, this necessarily causes new resonances to appear.

The sentence starts with "In such scenarios" referring to a list of models for which one does expect the new resonances. The sentence is correct within the boundary of the given scenarios

* L15: You say "lighter resonances". Lighter than what ?

The previous sentence places the typical mass of these resonances at the TeV (basically because of EWPT). Lighter is lighter than the TeV. We propose to change the sentence to "resonances lighter than one TeV".

* L17-18: You can't talk about "vector boson V pairs", since this equals "VV", which implies that you only consider "WW" or "ZZ", but not "WZ". You need to say "decaying to vector bosons pairs: ZV, (where V=W,Z)".

We have changed it as follows:

"In this Letter we present a search for heavy resonances decaying to ZW and ZZ final states"

and then we say "V (V=W,Z)" at line 19

* L24-25: I assume that the fraction of merged jets must increase as a function of the resonance mass. So "that ~ 70%" should be replaced by "more than ~70%", since you consider resonances "heavier than 800 GeV" ?

Agreed. Changed it.

* L97-98: provide a reference for "combined relative isolation"

We could use this reference:

@article{CMS:2011aa,
      author         = "Chatrchyan, Serguei and others",
      title          = "{Measurement of the Inclusive W and Z Production Cross
                        Sections in pp Collisions at sqrt(s) = 7 TeV}",
      collaboration  = "CMS Collaboration",
      journal        = "JHEP",
      volume         = "1110",
      pages          = "132",
      doi            = "10.1007/JHEP10(2011)132",
      year           = "2011",
      eprint         = "1107.4789",
      archivePrefix  = "arXiv",
      primaryClass   = "hep-ex",
      reportNumber   = "CMS-EWK-10-005, CERN-PH-EP-2011-107",
      SLACcitation   = "%%CITATION = ARXIV:1107.4789;%%",
}

or we could copy the definition from that paper (would produce a lengthy/verbose description). Which option would the PubCom prefer?

* L106: is the jet momentum really good to 5-10 % before any pileup for jet energy corrections are applied?

Yes. These are PF jets. At high energy the calorimeters dominate the precision. At low energy the tracker does. These numbers are correct.

* Section 5.1: you need to convince the reader that the MC will obviously simulate the variable alpha(M_VZ) more accurately than the variable N_NS(M_VZ). Currently you provide no arguments to support this. If you could say that alpha(M_VZ) is almost independent of M_VZ, because most of the physics cancels in the ratio, and therefore you expect it to be better simulated, that would be convincing. (Especially if you could say "alpha varies less than 5% as a function of M_VZ" or similar). Can you provide this kind of argument in the paper ?

Added:

There are several advantages in using the ratio $\alpha(M_{V\cPZ})$ for the background modeling of the $M_{VZ}$ distributions: the background estimation becomes insensitive to effects such pile-up corrections and integrated luminosity uncertainty which cancel out in the ratio; $\alpha(M_{V\cPZ})$ is less sensitive to improper modeling of the matrix element calculation for the background and to theory systematics (\eg~normalization and factorization scale, PDFs, \etc) since the background composition is similar in the two regions.

* Fig. 2: please show the expected signal from gravitons (like figure 3). Also show the expected signal from W'.

Done.

* Fig. 2: your background curve disagrees strongly with your background histogram in the lowest mass bin. Should we be worried?

Any disagreement below m_VZ~ 700 GeV has no impact on the analysis (we only search above that region). Besides, the difference between the histogram (MC estimation) and the curve (expected background) is treated as a systematic uncertainty. This is explicitly written in the paper in the sentence:

"The fit describes well the tail of the m_VZ distribution, which is the region of interest for the new resonance search. A discrepancy is observed at low m_VZ values. Any modeling imperfections, quantified as the difference between the best-fit function and the data distribution in the m_J sideband, are taken into account by assigning a systematic uncertainty."

* L208-210: "by constraining ... SM prediction" -->"from the equation rho = N_B N_D / (N_A N_C), where all variables are determined from simulation".

This sentence was already rearranged:

"The $\rho$ parameter is estimated from the simulated SM samples by rearranging Eq.~(\ref{eq:estimatedBackgroundCorrected}) in the following way: \rho = \frac{N_D \cdot N_B}{N_A \cdot N_C} and substituing the values setting the values of $N_A,\ldots,N_D$ to the ones from the SM prediction."

* Fig. 3: please show also the expected signal from W' bosons. No-where in the paper is there any indication of how this compares with the graviton signal (e.g. resonance width ?).

Done

* Sect. 6.1.1: If you are searching for a resonance in a particular mass bin, I would have thought that you could have used the rest of the WZ mass spectrum to measure the normalisation of the background. (If I remember correctly, this is what other resonance searches, such as EXO-11-019 did). That would have avoid the need to rely on simulation to predict it. Just a thought. I don't expect you to do it now ...

Thanks for the suggestion. We could have indeed done that. We will add this to the 2012 update.

* Sect. 6.1.2: I don't understand why you are using a toy MC to determine the statistical uncertainty on rho. The variable rho is given by a simple formula ` rho = N_B N_D / (N_A N_C)', where the values and uncertainties on the numbers N_i are given in Table 3. So one can calculate the relative statistical uncertainty on rho. It is `sigma_rho / rho = sqrt[ Sum_i ( sigma_N_i / N_i )^2 ] = 4.0%. Also, why do you get 15% when this calculation gives 4% ? Also, what is your 15% the statistical uncertainty on ? Is it the uncertainty you get with your 100k toy MC events ? If so, it is not relevant. We care about the statistical uncertainty on the CMSSW simulated events, of which there were presumably a different number to 100k. Furthermore, the CMSSW simulated events are presumably weighted (according to cross-section * lumi), so N_i will not be Poission distributed, as assumed at L246. Also, if the statistical uncertainty on rho is really 15%, how did you establish that its variation with the sideband cuts is correctly behaved to the 5% level ?

The term "statistical" here is really a misnomer, which we have being carrying since the beginning of the analysis. We changed the language in the text. The first check is just the first systematic uncertainty. What we do is to take it into account the correlation on the generation of the toys, and neglect in the computation of rho (which is what the ABCD method does by construction). We systematically see that rho deviates from one, and that is the distribution we see in the analysis note (AN2011_22g, Fig. 12) . We take the width of that distribution as a measure of the error in rho. This is where the number 0.064 (which we round to 0.06) comes from, which we quote as 15% of 0.42. As an analogy, the fact that rho is different from one is equivalent to a "Jet Energy Scale", while the fact that it has a distribution around 0.42 is equivalent to a "Jet Energy Resolution".

* Sect. 6.1.2: When using the ABCD method, one usually tries to find two variables, which on physical grounds, one expects to be reasonably uncorrelated. Since your correlation coefficient is 0.42, you clearly did not manage to find such variables. You therefore need to do a lot of work to convince the reader that this rho parameter is predicted accurately by MC. I find your choice of the two variables strange. Since your dominant background is Z+jets, and the Z and jet will be approximately back-to-back in the transverse plane, your y-axis in Fig. 1 is to good approximation equal to "2 * jet Et". It is not at all surprising that this is correlated with your x-axis "jet mass". And it is clear that to simulate "rho" correctly, your MC must accurately describe how jet mass depends on jet energy. This could depend on alpha_s, parton shower, fragmentation etc. Given how crucial knowledge of "rho" is to your result, I am surprised that you don't have stronger evidence that you know its value accurately. (e.g. Checking if you can predict the background for M_jet > 130, which is presumably also a background dominated region). At the very least, you need to provide more information in the paper about how the 5% systematic was obtained (L253). How much did you vary the sideband regions by ? Also, I understand from you AN-NOTE, that what you cared about is that the predicted number of background events in the A region in data was independent of rho, despite rho varying significantly with the sideband region. So please say so.

First, we should state that the correlation coefficient is not 0.42. The correction factor used for the ABCD method is 0.42. The correlation coefficient calculated from Figure 1a is around 30%, which is on the verge of small-medium correlation. In AN-2011-226, we find the effect of changing the jet mass threshold on the jet-MET transverse axis (Figure 11b). We agree that additional studies could be done as a cross check on the dependence of the jet mass on the jet energy. However, we believe the two tests described in the paper already account for these effects. We will also take the pointer on the cross-check (checking if you can predict the background for M_jet > 130 GeV region) in the 2012 update. We have changed the text with more information about the variation of the sideband regions and explicitly stated that what we care about is the independence of the predicted number of background events in the A region in data.

* L253-254: I think B_est is the estimation of the mean expected background. Please say so.

Done

* L278: What is a "flat" efficiency ?

A flat efficiency is an efficiency which doesn't vary with the pT/eT, except in the high-pT/eT asymptotic region. Rewritten as:

"assume that the efficiency does not vary with \pt (\ET),"

* L279: Please clarify how you see the "small decrease in the efficiency" ? Do you see this using signal MC events ?

This is a well documented inefficiency for high-pt muons.

* L281-284: The mass selection efficiency presumably depends on the W/Z mass resolution. For your signal, the W/Z decay to a single merged jet, whereas in your t-tbar control sample, they decay to two distinct jets. Is it obvious that the mass resolutions should be the same in these two cases ?

This is not true. We use a sample of ttbar with a boosted selection, so the two jets coming from the W from the hadronic t leg are also overlapping. You can see more details in: https://twiki.cern.ch/twiki/bin/viewauth/CMS/EXO11081QA#Questions_from_ARC_Sal_Mar_9th_2

* L286: What does "Propagating the jet energy scale effects ... E_T^miss" mean ? You don't use jets to calculate E_T^miss.

Indirectly, we do use jets to calculate E_T^miss. MET - particularly PFMET (which is what we're using) - is the sum over jets, electrons, muons, taus, photons, etc and the "unclustered energy". The Jet Energy Scale introduces an overall factor to the jets reconstructed in the event, which in principle must be backtracked to the constituents of the jets themselves, i.e., the particles. This factor should them be propagated to the calculation of the MET. Naturally, uncertainties in the JES will reflect into uncertainties in the MET, which is what we mean by that sentence. We used the procedure described in https://twiki.cern.ch/twiki/bin/view/CMS/MissingETUncertaintyPrescription.

* L298: Your counting experiment is performed in the region "M_T > 900 GeV & M_jet > 70 GeV".

Done.

* L316: You say that to LO, you have an 845 GeV limit on gravitons. But looking at Fig. 4b, it looks more like your limit is ~780 GeV.

The expected limit is ~780 GeV - where the dashed line crosses the LO cross section line. For calculating the observed limit, a linear interpolation is made between the points, corresponding to Monte Carlo signal samples. The linear interpolation crosses the LO cross section line ~840 GeV. This value is calculated by calculating the intersection between the linear interpolation and the LO line, it's not taken by eye.

* Sect. 7: If you persist in quoting limits on quantities that include the W/Z branching ratios, please make clearer that you combined limits are on a quantity that depending on BR(Z--> ee + mumu + nunu).

We changed the limits to quote in terms of sigma x BR(RS->ZZ) and sigma x BR(W'->WZ).

* L319-320: You "assume the signal efficiency remains the same". Do you mean that you assume that the signal efficiency is independent of the signal cross section (which is obvious) ? Or does changing "k" also change the signal kinematics ?

Changing the k/MPl changes not only the cross section, but also the width of the resonance. However, we have tested that for k/MPl in the range [0.01 - 0.3] the resonance width is narrow enough that the experimental resolution is the dominant factor.

* Fig. 4: Please change y-axis in plots to specify which process you are placing limits on. e.g. "sigma[W' --> WZ --> (q qbar) (l+l- + nu nubar)]".

Done.

* Fig. 5: You caption says ATLAS searched for ZZ --> 2l2j/4j in ref. [23]. But that is their dilepton resonance search. I think it puts limits on graviton --> ll, not graviton --> ZZ.

This was a typo in the citation name, and is now fixed.

* Tables 4-9: You use different mass binning for the graviton and W'. I assume this is because these resonances differ in width ? Please say so in the text of your paper.

Yes, the width and efficiencies are not the same for W' and RS. We added a line in the results section:

The mass windows optimization has been carried out separately for the \Wprime and RS graviton hypotheses to account for differences in the width and efficiencies.

* Tables 6-8: I assume that the uncertainties you quote include both the statistical and systematic uncertainties. Please say so. Also, I assume B_bgd is the mean expected background ? (i.e. The average one would get if we built a million identical CMS detectors), so the quoted uncertainty on it doesn't include the statistical variation one would get amongst these millions experiments. If so, please say that it is the mean expected background.

Done (in post-V4 version of the paper).

* Tables 6-8: I suggest adding the uncertainty on the efficiency to this table. That way, one will see at a glance how the total stat + syst. uncertainty varies at each mass point for the two MC inputs (background and efficiency). If you do this, could also consider dropping Tables 4-5. The latter are very long and contain more information than would interest the average reader. They are also a bit misleading, because, for example, one sees systematic uncertainties of 72% quoted on the background in Table 4, which sounds really scary, but only later when one sees Table 6 does one realize the background at the corresponding mass point is very small, so that this big systematic doesn't do much harm. My personal preference would be to keep Tables 4-5, but greatly simply them, perhaps just showing the breakdown of systematics at a couple of illustrative mass points.

The statistical uncertainty (determined by the finite statistics of the MC samples) is negligible, at the per-mil level, significantly lower than the main uncertainties (e,g. PDF for the dilepton channels, and a combination of PDF, JES and MET induced uncertainties for the MET channel). We do not think that adding explicitly a "<1%" in the table adds any value.

For the rest of suggestions, re: simplifying and/or merging some of the tables. We would like to discuss this during FR to make sure there is agreement among the members of the PubCom.

* Table 4 caption: What is a "flat uncertainty" ? Also "reconstruction" --> "lepton reconstruction" ?

Changed "flat" to "constant", and "reconstruction" to "lepton reconstruction"

* Table 6: This has a bin 1390-1610 in both the W' and RS models. Why is the statistical uncertainty on the background different in this bin for the two models ?

This was a typo, in one case we rounded the number and in the other we didn't.

Type A (English/Trivia) ================

* L2: "a spontaneous" --> "the spontaneous" ?

Done.

* L16: "region in" --> "region to" but

Done.

* L40-41: add commas either side of "which acts as the coupling constant in the RS model"

Done.

* L41: "experimental" -->"RS graviton"

Done.

* L100-102: consider moving the sentence "the requirement that... from jets" to Sect. 4.1, as you describe a cut, not a reconstruction technique.

To discuss in FR.

* L103: "vertex" --> "primary vertex"

Done.

* L148: "distance" --> "angle"

Done.

* L148: "2.8" -->"2.8 radians"

Done.

* L179: "fit the" --> "fit this"

Done.

* L179-180: "an analytic function ... distribution" --> "the following analytic function, and the fit result is used to parametrize the expected SM background distribution:"

Done.

* L192-193: These sentences should be moved to Sect. 6.1.1 ?

Discuss in FR.

* L197-198: "remaining SM background after the selection procedure" -->" SM background"

Done.

* Caption Fig. 3: "Region A and data" --> "and data in Region A"

Done.

* L223: Please rephrase. "these uncertainties" suggests that you are about to describe the systematic uncertainties which were "eliminated".

Changed to:

In the following, we consider the remaining relevant uncertainties in detail for the dilepton and Emiss channels

* L226: "number of expected" -->" expected number of"

Done.

* L232-233: "scaling up ... as follows, Pt' = Pt +- sigma_JES" --> "varying the jet Pt to Pt +- sigma_JES(Pt,eta), where sigma_JES(Pt,eta) is the uncertainty [ref]". (And similar correction at L261)

Changed to:

varying the jet \PT to $\PT \pm \sigma_{JES}(\PT,\eta)$, where $\sigma_{JES}(\PT,\eta)$ is the total jet uncertainty, and applying the full fitting procedure.

* L266: "For resonance" --> "However, for resonance"

Done.

* L313 & L316: Almost always, when CMS quotes mass limits, they are lower limits on a mass. Your limits are therefore highly unusual in that you rule out a range of masses. So we need to make this completely clear. e.g. Replace "between 700 and 938 (889)" by "in the range 700 - 938 (839)" (where the hyphen is double).

Done.

* L318: "extended in" --> "translated into" (and following hyphen should be double)

Done.

* L319: "a quadratic" --> "the quadratic"

Done.

Comments by Sarah Eno

figure 4 and limits. given the scatter in the observed limits, and that you linearly interpolate between then, can we really quote mass limits to 3 significant figures? If you want to quote so many figures, you need more reference mass points.

We rounded the limits to only two figures.

for the RS gravitons, why is the lower limit 750 and not 700?

Because although the analysis strategy has sensitivity above 700 GeV, the lowest mass point for RS gravitons is 750 GeV. So we use this lower limit to quote the limits. This was already brought up during the CWR.

when I look at the RS plot, the limit at LO to me looks like 750-800, not 750 to 845. it also looks like 800 at LO in fig 5. Is this a mistake in the text? (line 316)

The expected limit is ~780 GeV - where the dashed line crosses the LO cross section line. For calculating the observed limit, a linear interpolation is made between the points, corresponding to Monte Carlo signal samples. The linear interpolation crosses the LO cross section line ~840 GeV. This value is calculated by calculating the intersection between the linear interpolation and the LO line, it's not taken by eye.

figure 5. for your curves, you say if they are LO or NLO. for the earlier/other experiment results, though, I don't know if they used LO or NLO. Can this be indicated?

Done

figure 5. looking at figure 5, and depending on whether the previous experiments used LO or NLO cross sections, it looks liek we set the best limit on k for masses above either 1 TeV or 1.3 TeV? perhaps we should mention this in the conclusions?

Fig.5 is NOT a comparison with weaker limits. Fig.5 shows that our limit is complementary to that obtained with other searches, as a function of the mass and the coupling.

The 700 GeV mass for k of 0.05 is already excluded by d0 ditto for the 800 GeV limit. It looks like the D0 limit at 0.05 is 950 GeV? so better than our limit? but our limit becomes the best on lambda for masses above 1000 GeV, so why not emphasize this?

The D0 limit is derived in the ee/gamma gamma channel. In these channels, CMS and ATLAS both have much stronger limits than those by Tevatron and this analysis. The point that is made in the paper is that the limit derived in the VV channel is interesting in its own merit. The paper comments on more recent models on RS gravitons (Agashe et al) in which the leptonic and gamma gamma channels would be strongly suppressed, while the VV and ttbar channels strongly favored.

WRITING

title suggest "a jet and either a lepton pair or missing..."

To be discussed on FR.

abstract do we really need the "in several theoretical models"? (at the very least, it should be "for several theoretical models"

Changed to "for two theoretical models."

lines 6,7,8. Is the existence of new resonances the consequence of the new strong dynamics? If so, then the current construction of this sentence does not say this. It almost says that if the SM if an effective low-energy theory of any sort, then it predicts these resonances. I think you want. ... smaller than a new-physics scale \Lambda. In these theories, one expects...

Done.

line 14. Can we mention the strongest limit on the mass for the parameters mentioned in your abstract for these earlier results? (so that when the reader hits line 39 and learns that you start your searches at 700 GeV, they will know if there is an uncovered search region or not)

line 9 should "in relation to" be "for"

Done.

line 16 "in which" -> "for which" or "to which"

Done.

line 22 close -> close to

Done.

line 23. QCD jet -> quark or gluon jet

Done.

line 30. It would be better to structure this sentence so the "threeness" is clearer: consider three final states: one heavy jet and either Z-ee, Z->mumu, or Z+met, where MET is the characterstic signature of neutrino production. (Emission is also not a good work. sounds like to should be a brem process)

Done.

lines 45-59. You forgot the HF, which is used in MET. Please add "CMS also has extensive forward calorimetry."

Done.

line 61 "is done separately" -> "is different"

Done.

line 63 "have been" -> "were"

Done.

line 63. do you mean "with double-electron or single-muon triggers" or "a double-electron trigger or single-muon triggers"?

Changed to: "with double-electron or single-muon triggers."

line 65. if 40 and 17 really are the tightest, then you don't need the "e.g.". You only need that if these are examples and there are other thresholds that are also the tightest.

Changed to i.e. due to a previous FR comment.

line 69 "The trigger here" -> "These triggers have efficiencies"

Done.

line 70. What do you mean by "and corrected offline"?

We mean both the offline reconstruction and corrections were used to calculate the trigger efficiency. Changed to:

"with a leading jet of transverse momentum $\pt > 160 \GeV$ and $\MET > 300 \GeV$ after offline reconstruction and corrections,"

line 72 "We consider several possible" sounds like maybe you don't consider al possible backgrounds. Maybe "We consider the SM back..."

Done.

line 77 this is normally phrased "with a full simulation of the detector based on GEANT", as this is more technically correct than what you have written.

Done.

line 87. What is a "point window"? do you mean within a window of +-24 cm of the nominal interaction point?

The word window was removed. "where the vertex is reconstructed within $\pm$24 cm of the nominal interaction point along the beam axis,"

Table 1 caption. Do the K factors just include these corrections or are they THE NLO (NNLO) corrections? (i.e. do they include anything else beyond the NLO (NNLO) corrections?) Maybe comprise would be a better word?

Done.

97 components -> energy deposits?

Done.

line 97. Most papers have a much more complete description of isolation

To be discussed in FR.

line 106 this -> the

Done.

line 111. need a citation for the JES corrections (and they use more than just dijets and gamma+jets)

The Detector Performance Note with 2011 data is not yet public. The link to CDS is here: https://cdsweb.cern.ch/record/1454659?ln=en The paper with 2010 data is here: http://iopscience.iop.org/1748-0221/6/11/P11002/

line 118. for electrons, you don't exclude 1.57 to 2.1?

The gap is not explicitly removed, but we treat electrons in this region like endcap electrons. This was consistent with the official Egamma recommendation (ie. to not explicitly remove them). Trying to find the reference for that.

line 123 requiring their invariant mass to be in the range 70<M_Z<110

Done.

line 124 If there are multiple Z candidates, the one with mass

Done.

line 127. I don't understand "computing tfrom the vectorial sum of the four-momenta". Mass isn't a vector. And you already defined momentum of a jet. maybe you mean calculated from the jet energy and momentum?

Changed to: "computed from the jet energy and momentum calculated as the vectorial sum of the four-momenta of the constituent PF particles"

line 134 mother particle -> parent particle

Done.

line 138. In case people skip reading the header, I would start this section "For the ETmiss channel, ..."

Done.

line 142. combine two sentences for "Events containing a track with its isolation parameter smaller than 0.1 are discarded".

Done.

line 145. need a comma after 2.4

Done.

I was surprised that there is no figure in this paper showing the jet mass for the lepton channel. I think it would be a nice addition. maybe around line 171 you could show this distribution? and mark the "NS" and "SB" regions? and maybe show the M_VZ distribution fo the SB selection?

We could add Fig.3 or Fig. 6 from AN-11-414_v13. However, we did not think that they are terribly exciting. To be discussed in FR.

line 174. Can you give a typical value of alpha?

Alpha is not flat, it has a ~linear dependence on the M_VZ. We think showing the plot is too much detail for a paper, and it is hard to quote a typical number. Please see Fig. 7 (middle plot) in AN-11-414_v13 and let us know if you agree.

line 179. which M_VZ distribution is fit? the data? the sideband? the MC? I guess the MC as you then say "the result is used...expected SM background" If it is instead the data, do you exclude the signal region when doing the fit?

For clarity, changed to: "we fit the $\alpha$-corrected sideband data $M_{V\cPZ}$ distribution to the following analytic function"

line 212 how big is the scaling factor? can you quote it?

The scaling factor is approximately 11%.

line 244 aimes at -> is used to estimate

Done.

line 272. I'd delete this line

Done.

figure 4. Why no limits from other experiments or other channels for the W' figure in figure 4?

In the brazilian-style 1D exclusion limit plots, (we think that) it is customary to only show results from the analysis discussed in the paper, and quote in the text limits from other experiments/analyses in a particular model (e.g. SSM). This is what we have done here, by quoting the W'->WZ->3lnu limit from the EXO-11-041 analysis. Fig. 5 is somewhat different since the RS models are characterised by two parameters (k/M_pi and mass) that require these 2D exclusion plots. Here it is interesting to show results from various experiments because different analyses tend to cover different regions of the 2D space. It is therefore a convenient/useful way to visualize the complementarity of the different experimental results.

Comments by Robert Brown

Type B

Abstract: Add 'Model-dependent' before 'upper limits'

Done.

L12: Change the order of the last 2 sentences in this paragraph. Remove the word 'Indeed'.

Done.

L74: It is not appropriate to use parentheses as a shortcut when referring to a Table. Spell the text out in full.

Done.

L93: What is meant by 'leading vertex'? This needs to be explained.

Done.

L96: I do not understand what is meant by: 'the ratio of the hadronic to the electromagnetic components.' The implication from the start of the sentence is that this is achieved using the ECAL alone, whereas I assume that it uses information from the HCAL as well. This needs explaining more clearly.

This was rewritten as: "Charged hadrons that are consistent with primary vertices other than the vertex with the biggest sum of p2T (leading vertex) are removed from the collection of particle candidates"

L103: 'to the vertex'. Which vertex? The 'leading' vertex?

Changed to "leading vertex"

The columns in Table 2 should be given appropriate headings: 'Samples' -> 'Channel'; 'Notes' -> 'Simulation details'

Done.

L149: 'aims at' is colloquial, but in any case the sentence is poorly written. Allowing some 2-jet events may increase the signal efficiency, but 'this condition' (that they are not back to back) must inevitably 'reduce' the efficiency. I have a feeling that I am not expressing myself clearly, but the point is that a new class of events is allowed, and a 'condition' is imposed on that new class. There is confusion in the wording concerning which of the two actions 'increases' the signal efficiency and which 'reduces' the background.

Done.

Fig.2. Legend. Do not use the ROOT option with a horizontal line to identify 'data'. Ideally it should be a point with a vertical line, but I believe that this option is not available - however, there is one with a horizontal and a vertical line - use that.

Done.

Fig 3. Legend Ditto. A plain point is used here, which is better, but I prefer the option requested above.

Done.

L223: I simply do not understand what is meant here. Should 'these uncertainties' be 'the remaining uncertainties'?

Done.

L237: Again, I am not clear what is meant. In particular, 'the largest value between' could mean a difference between two values, or a choice of one or the other. I think that the sentence should start: 'The largest difference between the values obtained with the various fitting procedures.....', but I am unable to guess what should come after that. Please express more clearly.

Changed to:

These variations are compared to the difference in the number of expected background events in the mass window as estimated from data and with MC simulation. The largest of the two is used as the systematic uncertainty in the background determination.

L257: 'which' -> 'that' since the last clause is essential to the understanding.

Done.

L259: The implication is that the value for the Lumi uncertainty is an assumption, whereas it was obtained by careful analysis. -> 'A value of 2.2% was taken for the uncertainty in the measurement of the integrated luminosity [45].'

Done.

L278: 'Even though these numbers assume a flat efficiency..' I do not understand what I am supposed to gather from this. Please spell out the message more clearly, for example: how might the discrepancy arise?

Conclusions -> Summary. This is the CMS style if no additional information is given in this section.

Done.

'We' should not be used in the Summary.

L344: A new wording has been released for the Acknowledgements'. Please update from the Pub Com twiki.

Done.

Type A

Abstract: Here and throughout, I do like 'decaying hadronically' (or 'leptonically') since this implies something about the decay mechanism rather than the final state. In this instance, 'hadronically' is not needed at all, since 'into two overlapping jets' is sufficient.

Done.

Delete 'the' before 'mass' (4 lines up from bottom)

Done.

Here and throughout: 'Randall-Sundrum' needs a long dash, not a hyphen (two different people).

Done.

L4: 'can' -> 'may'

Done.

L18: -> 'decay to leptons'

Done.

L19: -> 'decay to hadrons'

Done.

L27: Replace 'vector boson' with 'decay'

Done.

L30: 'In summary' seems odd part way through the introduction, I suggest 'Thus' On the same line, I had trouble counting the '*three*' final states until I realised that 'or' appears ambiguously here. Despite the repetition, I think it would be clearer to spell the list out in full: "...one heavy jet plus Z -> e+e-, one heavy jet plus Z u+u-, and one heavy jet plus missing transverse energy..."

Done the "Thus", the remaining sentence was changed due to another FR comment.

L38: replace the comma after 'pair' with a semicolon

Done.

L40: a comma is needed before 'which'.

Done.

L57: delete 'in turn'

Done.

L71: 'allows to' is incorrect. Either 'allows one to' or: 'allows resonances.....to be probed...'

Done.

L87: delete 'window'

Done.

L94: insert a comma after 'jets'

Done.

L96: 'with dedicated requirements' -> 'which satisfy requirements'

Done.

L113: -> 'missing-transvers-energy vector'

Done.

L124: remove 's' from 'pair'

Done.

L126: 'hadronic' - see above. In any case it is not needed here since the jet requirement is spelled out.

Done.

L127: 'vectorial' -. 'vector'

Done.

L138: delete 'leptonic'

Done.

L150+: 'signal region' -> 'signal sample' 'Figure' should be written in full at the start of a sentence. 'vs' should not have full stop.

Done.

L154: 'Unlike' -> 'In contrast to the approach used for'

Done.

L156: Move 'region' to appear before 'Mj'.

Done.

L166: The wording is clumsy. ('for which...and which' doesn't work) I suggest: 'The background is modeled using a control region consisting of a sideband in Mj (30 < Mj < 65 GeV). The remaining selections are applied unmodified to these events, providing a sample that is kinematically equivalent to the nominal selection.'

Done.

L173: Put 'NS' and 'SB' before the respective brackets, so that they are adjacent to the defining words 'nominal' and 'sideband'.

The lines were reformulated because of another FR comment.

L175: The use of brackets is totally confusing. I do not like this linguistic shortcut at the best of times, and mapping square brackets on to round brackets later in the sentence is a step too far. Please write the sentence out in full.

Done.

L178: move 'ratio' to come before 'alpha' on the previous line.

Done.

L193: Add 'in the limit calculation' after 'account'.

Done.

L197: insert 'charged' before 'lepton' (neutrinos are leptons!)

Done.

L198: delete 'similar', insert 'similar to that described above' after 'technique'.

Done,

Fi2. 2 caption: delete 'as described in the text'

Done.

L206: Add 's' to 'number' (it does have a plural form).

Done.

L212: in general we avoid qualifying 'agreement with subjective words such as 'good'. 'Agreement' is sufficient here.

Done.

L215: I don't understand the motivation for changing tense at this point: replace 'were' with 'are'.

Done.

L258: 'These' must be followed by a noun. -> 'These uncertainties'

Done.

L265,266: Move the terms in square brackets to appear after 'range' (three instances).

Done.

L267: 'due to' is used incorrectly. Either 'owing to' or 'because of'

Done.

L272: -> 'Systematic effects specific to the dilepton or EmissT channels are discussed below.

This sentence was deleted because of another FR comment.

L277: -> 'apply them *to*'

Done.

L288: insert 'the lowest,' before MG = 750. 'due to' -> 'because of'

Done.

L290: 'quote' -> 'arrive at', then continue the sentence: ', where a value of 10% is obtained for the final systematic....'

Done.

L299: (and L308, 314) The CMS style is 'CL'.

Done.

L300: move 'products' to appear after 'combined' in the previous line.

Done.

L316: 'is for' -> 'rules out'

Done.

L318: there should be a space on either side of the '-'

Done.

L323: Move ',proposed in recent studies [54],' to come after 'models' in line 321.

Done.

You can't say 'opens ..... the interest' -> '...pairs, and motivates the investigation of large values...'

Done.

L325: 'released' -> 'published' (I think that this is OK, even if it is only on ArXiv, since published means 'announced publicly')

Done.

L329: 'of' -> 'for'

Done.

L332: 'hadronically' -> 'to hadrons'

Done.

L337: Move 'in several theoretical models' to appear after 'set' in the previous line.

Done.

L340: 'Randall-Sundrum' needs a long dash not a hyphen.

Done.

Comments by Sijin QIAN

Page 1 (1) This is a previous comment for pas-v5, i.e. (23) L2, some words may should be swapped, i.e. "In many extensions of the standard model of particle physics (SM) [1-3]" --> "In many extensions of the standard model (SM) of particle physics [1-3]" This item has not been touched yet, please consider.

Done.

Page 2 (2) L83, the "PDF" should be explained at its 1st appearance in text here instead of L259, i.e. "The PDF used is CTEQ6L1 [35]." --> "The Parton Distribution Function (PDF) used is CTEQ6L1 [35]." Correspondingly, L258-259 may be shortened from "the parton distribution functions (PDF) and trigger and reconstruction efficiencies." --> "the PDF and trigger and reconstruction efficiencies."

Done.

Page 12, Fig.4 (3) At the top lines above each plots, the square-root signs miss the horizontal bars above the letters of "s". They should be added.

Done.

(4) The location of Fig.4 may should be moved to Page 11 since its first citation in the text is on L311 in Page 10, so that its 1st citation would not be separated from the figure itself by a full page.

We will not change the location of figures and tables while a final text is not made. Latex is not the greatest thing customizing this sort of stuff.

Pages 15-17, in the References Section, (5) L398, in [16], to be consistent with the capitalization in the title, I'm not sure whether the word of "Minimal" should be capitalized or not, i.e. "The Minimal composite Higgs model" --> "The minimal composite Higgs model"

Done.

(6) L417, in [23], obviously "[23] ATLAS Collaboration Collaboration," --> "[23] ATLAS Collaboration," Other ones which need to be similarly modified are [24], [44] and [49].

Done.

(7) L437, in [30], to be consistent with other PLB Refs. in this Section (e.g. [21], etc.), it should be shortened from "Physics Letters B 672 (2009), no. 1, 45," --> "Phys. Lett. B 672 (2009) 45,"

Done.

(8) L439, in [31], to be consistent with other NPB Refs. in this Section (e.g. [16], etc.), it should be shortened from "Nucl. Phys. B 818 (2009), no. 1, 28," --> "Nucl. Phys. B 818 (2009) 28,"

Done.

(9) L445, in [33], to be consistent with other multi-author Refs. in this Section (e.g. [31], [32], [35] and [36], etc.), two more names of author should be added, i.e. "[33] J. Alwall et al.," --> "[33] J. Alwall, x. xxxx, y. yyyy et al.," Another one which needs to be similarly modified is [47].

Done.

(10) L451-452, in [35], to be consistent with other JHEP Refs. in this Section (e.g. [33] and [36], etc.), it can be shortened from "Journal of High Energy Physics 2002 (2002), no. 07, 012." --> "JHEP 07 (2002) 012."

Done.

(11) L472, in [44], to be consistent with other Refs. (e.g. [40]-[42], etc.), a space should be added before the volume index, i.e. "Phys. Lett. B704 (2011) 123," --> "Phys. Lett. B 704 (2011) 123,"

Done.

(12) The "year" number should be given for Ref.[47]. If there would be problems to display the year number with the default bib file, it may be fixed by changing from "article" to "unpublished" in the bib file.

Done.

(13) L486, in [50], to be consistent with other PAS Refs. (e.g. [48] and [58], etc.) in this Section, it should be changed from "CMS PAS EWK-10-002 (2010)." --> "CMS Physics Analysis Summary CMS-PAS-EWK-10-002 (2010)."

Done.

(14) L488, in [50], one space should be added after the symbol "=", i.e. "sqrt(s) =7 TeV" --> "sqrt(s) = 7 TeV"

Done.

Page 18

(15) This is a previous comment for pas-v5, i.e. (58) Table 5 (now Table 4 in v2), (a) in the 2nd header row, the 2nd column, to be consistent with the expression in the 1st column, the capitalization should be changed from "Mass Window" --> "Mass window"

To be discussed in FR.

(b) The location of Table 5 (now Table 4 in v2) may should be moved to Page 9 since its first citation in the text is on L196 in Page 7 (now L217 in Page 8 in v2), so that the 1st citation would not be apart from the Table itself by a full 3 pages (now a full 10 pages in v2). This item has not been touched yet, please consider. Especially for the item (b), the physical location of Tables 4 and 5 should be moved to before Figs. 4 and 5, since the 1st citations of Tables 4 and 5 are on L217 and L218 while the 1st citations of Figs.4 and 5 are on L311 and L320, i.e. Tables 4 and 5 are earlier.

We will not change the location of figures and tables while a final text is not made. Latex is not the greatest thing customizing this sort of stuff.

Pages 19-21 (16) This is a previous comment for pas-v5, i.e. (68) For all 4 Tables (i.e. Tables 8-11, now Tables 6-9 in v2) (a) In the header row, the right-most 2 columns, similar as the items (58a) and (61a) above, the capitalization should be changed from "Obs. Limit (pb) | Exp. Limit (pb)" --> "Obs. limit (pb) | Exp. limit (pb)" (b) ... (c) All 4 Tables should be moved to the positions before Fig.4 since the 1st citation of Fig.4 in text is on L287 (now L311 in v2), which is later than the 1st citation of Table 6 on L286 (now Table 9 on L310 in v2). This item has not been touched yet, please do consider, especially for the item (c).

To be discussed in FR - to be fixed after final version to avoid multiple headaches sorting position of floats on Latex.

English/Style comments from Language Editor (Paul Padley) %Completo_5_%

Line 5: Please write ``or a composite Higgs boson'' You can drop all the other worlds

Done

Line 6: ``In this respect..'' is a bad way to start this sentence. Also, I think this can be conjoined with the following sentence to produce something tighter: ``In such scenarios the SM is an effective low-energy theory, valid for energies smaller than a new-physics scale �> in which case one expects the existence of new resonances coupling to pairs of vector bosons (ZZ, ZW, and WW).''

Done

Now in line 8-9 you also refer to ``This effective Lagrangian''. What effective Lagrangian would that be?

Indeed we should replace effective Lagrangian with effective theory. Since any QFT is expressed through a Lagrangian, we used the two as if they were indicative of the same thing. And this is confusing. Since the sentence starts as "The SM is an effective low-energy theory" replacing Lagrangian -> theory should be OK.

Line 8 (on V1): "A few examples", this is wrong since you give two examples. Instead write: `` Examples include Randall-Sundrum (RS) gravitons GRS [12, 13] coupling to ZZ and WW, and technimesons [17, 18] coupling to WZ. '' (on 18-SEP-12 draft): This now starts in line 9. However the problem is still there. In the new context it should say ``Other examples includeâ?¦''

Done

In the remainder of this paragraph you effectively say these ideas can be tested at the LHC if lambda is order 1 TeV and that limits have ready been set at higher scales. So the reader will naturally conclude that there is nothing new to contribute, but you anticipate this by adding a final paragraph, ``it should be noted however�?�''. Let me suggest that this is not a very strong argument as presented and that this part of the first paragraph should be carefully rewritten.

First of all, we should not have said "O(1 TeV)", we should have said "O(TeV)", because CMS (in general) and this analysis (in particular) are sensitive in the beyond-1-TeV region. This has now been corrected.

The "around 1 TeV or above" scale is relevant to mention here for two reasons: it is theoretically motivated (as explained in this paragraph) and the analysis has high sensitivity. We do have to note that in other models (Technicolor, etc) which favor lower-mass new particles, this analysis does not have any sensitivity. We have added a comment to make this more clear. We are obviously open to suggestions as to how this could be improved.

Line 16 (on V1): Z �?' ll (l = μ, e) or νν �", is not a boson. Z is a boson. This needs to be rewritten (on 18-SEP-12 draft): This problem still exists but now in in line 18 where you say, on boson being a Z\rightarrow l+l-�?� Z\rightarrow is l+l- is not a boson, it is the decay of a boson.

Done

Line 17, 18 (on V1). When you are using the term boosted, I believe you mean something much more specific than how the term is used in relativity. Obviously a charm quark with 100 GeV is highly boosted in the detector. What you are attempting to describe is the fact that above 800 GeV jets from quark pairs merge into a single massive jet. You do attempt to say this with the use of the I.E. clause but I think you need to spell this out much more clearly.

(on 18-SEP-12 draft:) This problem is still there. Since this is an unconventional use of the word boosted it needs to be more directly addressed. The line numbers of this problem are shifted.

This has been rephrased now.

Line 23 (on V1). Why is the word ``leg" used? Just remove it. (on 18-SEP-12 draft:) Above problem still exists. Now starting in line 26. Why not say ``each reconstructed lepton interferes with the isolation definition of the other lepton�?�.''

Done

In line 37, Gkk should have commas on either side of it.

Done

Line 55 (on V1): I just don't know what the sentence starting at the beginning of line 55 means. The trigger does not ``become'' anything. (on 18-SEP-12 draft:) This problem still exists now starting in line 68. I can ultimately figure out what you mean but make it easy for the ready, there must be a more direct way of saying this

Done

Line 70 (V1) or 87 (18-SEP-12 version): Do we contemplate something other than ``collision'' events? Why is this word here?

Done

Line 79 (V1) or 97-98 (18-SEP-12 version): Muons are fundamental particles and can not be built. Presumably you mean muon tracks or objects and should say so.

Done

Line 90-91 (V1) or 105 (18-SEP-12 version): You did not find anything in the simulation. You found something in the ``simulated data''.

Done

Line 106: Is it obvious what "generator level jet" means or would this benefit from further exposition.

"generator-level" jet: a jet which is created from generator-level "stable" particles (kaons, pions, muons, etc.)

This is standard jet terminology and we believe no further explanation is needed in the text.

Line 109: I don't know what "in-situ" means in this context. Why not just say, Jet energy corrections are derived from the energy balanceS

Done

Line 125: the sentence is confusingly written. What does it mean to say ``is reconstructed as a jet with pt>250 GeV''. This implies that the reconstruction algorithm has forced the pt to take a value greater than 250 GeV. Of course what you mean to say (I think) is that you apply a requirement, I.e. Make a cut.

Done

Line 130: This is unnecessarily complicated, you mean to say ``by maximizing the quantity Ns/sqrt(Ns+NB) (where S''

Done

Line 145: To be consistent with what you write below you should refer to the signal region rather than the signal sample.

Done

Line 147: Do you need to say 2.8 is in radians.

We do not think this is necessary. This is explicitly mentioned in the definition of the phi angle in Section 2. Furthermore, this is a quote from Wikipedia:

As the ratio of two lengths, the radian is a "pure number" that needs no unit symbol, and in mathematical writing the symbol "rad" is almost always omitted. In the absence of any symbol radians are assumed, and when degrees are meant the symbol ° is used.

Line 158: Simulation is not an animate object and can not hold opinions or make judgements. For example it is wrong to say ``According to simulation''. Instead you must say something like ``By analyzing simulated data we determine thatS''

Done

Line 178: Its an ``analytic function'', not an ``analytical function''

Done

Line 189: The sentence that starts ``The fit describes wellS'' does not have a natural English syntax. Instead write: ``The tail of the Mvz distribution, which is the region of interest for the new resonance search, is well described by the fit function/''

Done

Line 194: Again ``According to simulation''. Make the same correction as line 158.

Done

Line 195: no s on production.

Done

Line 238: This sentence is very difficult to read. It should be rewritten.

Done

Line 251: This is a strange concept: ``its statistical fluctuation is around 15%''. You presumably mean to say that the statistical fluctuations on rho vary within 15%.

15% is the statistical variation of rho. The sentence has been rephrased.

Line 257: This paragraph reads line all the systematic uncertainties are enumerated here. Here is a suggested rephrasing: There are several systematic uncertainties in the expected signal yields which are common across channels. These are the error on the luminosity measurement, the JES effects on jetsS

Done

Line 261: This is a difficult to read and confusing sentence. Try writing something like: To determine the effect of the JES uncertainty, the Pt of every jet is scaled up and done by the total uncertainty S and then the full analysis selection is reapplied.

Done

Line 266: Don't say ``no larger than'' say less than

Done

Line 288: write: ``In this case, due to threshold effects, the systematic effect is found to be around 7%''

Done

Paragraph at line 321: This seems very defensive and seems to undermine the argument. You want to present things in a positive and affirmative way. For example, point out that this analysis is sensitive to different variations of RS models than gamma gamma and ll. Since that is the case there is no reason to directly compare the limit to those.

Thank you for this suggestion. We have rewritten this paragraph and we think we are now making a stronger case.

The exposition of those model variations starting in line 324 might better be placed in the introductory material. Doesn't this provide a motivation for the work performed that if put in the introduction provide the reader with an incentive to read one?

We would like to keep a distinction between the general motivation for this search (ie. the fact that TeV resonances decaying to VV is a general feature across many models, as discussed in the Introduction) from the interpretation (which is only partial, as it considers only a few benchmarks). This sentence here refers to a specific case of RS model, which is a VV-coupling-enhanced BSM model. We think that discussing the specific case in the Conclusions section is relevant for communicating the implications of the results in the particular model. However, we think that the motivation for the search is already strongly supported by the theory papers to which we refer (ref. 13-19).

In figure 4, at the top, the sqrt(7) TeV is on the left side of the plot but in figure 5 it is on the right. You should be consistent.

Done

Comments for CWR

English/Style comments left for Language Editor to decide/act upon

- last line in the abstract: not sure we should really stress that we are the first here. We know it irritates our competitors and we had some gentlemen's agreement on that, but the pubcom can tell you if this was just for titles of papers or also for the abstract content

- end of page 1 and page 3: footnotes were used to be forbidden in our papers. If this is still true (plse check with your pubcom chair) then we should indeed remove them.

Lines 3 and 4: "dynamic" is an adjective, not a noun (or very rarely). Suggest change to "dynamicS" (plural).

Lines 33-47 The description of the CMS detector is untypical short, which may be justified. However, it may be worthwhile to give some indication of which aspects of CMS detector performance are important for this analysis.

Line 48 "Collision" is an odd description; how about "Experimental"?

Table 1 The text and table caption refer to "samples" which would usually be understood to mean the collection of simulated events generated for comparison with the data. However, Table 1 mainly lists the information on the signal properties (the signal mass, the cross-sections and k-factors). The only item which indicates these are related to samples is the list of masses, which are presumably the ones generated. This should all be stated more clearly. Also, the k-factor is not defined, the column "Mass" presumably refers to the graviton or W' mass and should be labeled as such for the two sections.

Line 98 The vector arrow over E^miss_T should be lengthened or centred on the E

Line 110 Change to "one matching pair" (singular)

Line 274 "a single event-counting experiment" is jargon to some extent; how about "a simple count of the events". A similar change should be done for the captions of figs 4 and 5.

(21) The title of article seems a little too long and has not mentioned the pp collision at 7 TeV yet. I'm not sure whether it would be somehow better to retain the 1st half of the current title plus the beam information, i.e. "Search for exotic resonances decaying into V + Z using final states with a jet and a lepton pair or missing transverse energy" --> "Search for exotic resonances decaying into V + Z in proton-proton collision at sqrt(s) = 7 TeV"

(22) In Abstract (a) The 1st line, "final state" --> "final states"

(24) L3, "is associated to" --> "is associated with"

(26) L15, it may should be changed from "In this note we present ..." --> "In this paper we present ..."

(32) L33, to be consistent with the capitalization of other Section titles, it should be changed from "2 CMS Detector" --> "2 CMS detector"

(36) L63, as NLO has just been introduced on the last line (L62), I'm not sure whether this line can be shortened from "next-to-next-to-leading order (NNLO) corrections," --> "next-to-NLO (NNLO) corrections,"

(38) L67,

(a) similar as the item (32) above for the capitalization, "4 Physics Object reconstruction and Event Selection" --> "4 Physics object reconstruction and event selection"

(b) To be consistent with other Section titles, the distance above the title from the end of last Section should be increased.

(48) L139, similar as the items (32) and (38) above, "5 Background Estimation" --> "5 Background estimation"

(58) Table 5, (a) in the 2nd header row, the 2nd column, to be consistent with the expression in the 1st column, the capitalization should be changed from "Mass Window" --> "Mass window" (a) in the 2nd header row, the 2nd column, to be consistent with the expression in the 1st column, the capitalization should be changed from "Mass Window" --> "Mass window"

(b) The location of Table 5 may should be moved to Page 9 since its first citation in the text is on L196 in Page 7, so that the 1st citation would not be apart from the Table itself by the full 3 pages.

(59) L270, similar as the item (38b) above, the distance above the title from the end of last Section should be increased.

(61) Table 6

(a) Similar as the item (58a) above, "Mass Window" --> "Mass window"

(b) Similar as the item (58b) above, the location of Table 6 may should be moved to Page 10 since its first citation in the text is on L197 in Page 7, so that the 1st citation would not be apart from the Table itself by the full 4 pages.

(a) In the header row, the right-most 2 columns, similar as the items (58a) and (61a) above, the capitalization should be changed from "Obs. Limit (pb) | Exp. Limit (pb)" --> "Obs. limit (pb) | Exp. limit (pb)"

(62) L289-290: the results here is not close enough to say "comparable".

(27) L22-L24: Perhaps, sounds better to say "For the same reason, when reconstruct the Z->ll vector boson events, each leptonic leg interfering with the isolation definition to the other legs has to be excluded from the isolation calculation in order to avoid introducing inefficiencies."

(28) L25, to avoid the duplication of words as at the beginning of L15, it may be changed from "In summary, in this paper we consider three final states:" --> "In this study we consider three final states:'

(30) L31, please give the definition for the "reduced Planck mass".

(31) After L32, this paper has 16 pages without counting the Acknowledgments and References Sections; for the long articles, at the end of Introduction Section, normally it has a paragraph to briefly introduce each of other Sections in the paper, e.g. "In this paper, Section 2 is ...; Section 3 ...; ... are described in Section 4; ... a Summary is in Section 8". But this paper seems missing this paragraph yet.

(41) L80, the "system" should be plural, i.e. "inner tracker and a track from the outer muon system." --> "inner tracker and a track from the outer muon systems."

(42) L101, because the text before is for physics object reconstruction, the title of Sub-section should be expended from "4.1 Dilepton channels" --> "4.1 Event selection in dilepton channels"

(44) L120, similar as the item (42) above, "4.2 EmissT channel" --> "4.2 Event selection in EmissT channel"

(52) L188, I'm not sure whether a word of "and" is better to be added or not, i.e. "the values of NB, NC, ND to the ones from the SM" --> "the values of NB, NC, and ND to the ones from the SM" or "the values of NB, NC and ND to the ones from the SM"

(64) Table 7, similar as the items (58b) and (61b) above, the location of Table 7 may should be moved to Page 11 since its first citation in the text is on L197 in Page 7, so that the 1st citation would not be apart from the Table itself by the full 5 pages.

(65) L307, the "state" should be plural, i.e. "for new exotic particles decaying to the VZ final state," --> "for new exotic particles decaying to the VZ final states,"

68(c) All 4 Tables should be moved to the positions before Fig.4 since the 1st citation of Fig.4 in text is on L287, which is later than the 1st citation of Table 11 on L286.

Comments from Don Summers (CADI link) %Completo_5_%

English/Style/Formatting/References

Ref [3] Salam: Might want to add page 367

Done

Ref [29] arXiv:1011.3540 now published as Comput.Phys.Commun. 182 (2011) 2388

Fixed. (Now Ref[31])

Ref [44] Reference should be JHEP 1101 (2011) 080

Fixed. (Now Ref[48])

Ref [46] hep-ex/9902006 now published as Nucl.Instrum.Meth. A434 (1999) 435

Fixed. (Now Ref[50])

Comments from Albert de Roeck (CADI link) %Completo_5_%

Analysis %Completo_5_%

- Is there any overlap/connection or even cross check with the monojet analysis? In particular for the MET case you basically have just a monojet (up to QCD radiation) and MET, ie a similar topology.

Indeed the two analyses are quite similar. However, there are two very important differences. First, our analysis cuts (that aim to reconstruct two partially overlaping jets appearing as a single fat jet) reduce the background by a very large amount and limit the phasespace to events with one very-large-mass jet. Second, the mono-jet analysis is not stricly speaking a real mono-jet analysis, since there is no veto on the second jet (there is a veto on the third jet). This was done in order to keep a good sensitivity to DM production. The two analyses are therefore dealing with different phasespace regions. Unfortunately, since they were performend at different times, no direct cross check was performed and several analysis details ended up being very different (e.g. the background prediction). However, we agree that this is a good suggestion, and we will keep it in mind for a future update.

- is the V->qq always taken to be a monojet in this analysis? Ie there must be some phase space where you actually see the two jets separated, but maybe it is not an important part for the phase space you are looking at (high pT)?

In this analysis, we only consider the boosted topology where the jets are merged. There is another CMS EXO analysis (EXO-11-102) which considers the non-merged jets case, ie. ZZ->qqll. The plan for 2012 is to combine all the searches in the double vector bosons channel. For the 7 TeV/2011 round, the combination will be done only with the boosted V + Z -> ll, νν analysis.

- How (with which recombination scheme) is the jet mass calculated in this analysis? Maybe useful to give as information

We use the E-scheme, i.e., the four-vector of the jet is the sum of the four-vectors of all its constituents. This is explicitly written in line 89 (now line 101). The mass of the jet is then calculated as the standard sqrt(E^2 - p^2).

- Is there any discussion on the overlap of the leptons? Can you loose events because the lepton signals overlap (in particular electrons?) Is that what is meant at the end of 6.2.1? Otherwise I would worry if we may be loosing events at the trigger or the selection due to merged lepton signals, and how well we control that.

No, the end of Sec. 6.2.1 is discussing the (hadronic) V mass cut efficiency.

For high mass resonances (i.e. G_KK mass greater than 1.3 TeV) the dilepton channel does, indeed, suffer efficiency losses due to the lepton isolation. This has a different effect for electrons and muons.

For the muon channel we use a SingleMu trigger to avoid the loss of events related to the isolation definition at the trigger level. In the offline selection, the muons are reconstructed using the cocktail algorithm and the high pT muon ID recommendation without the isolation: the muons have no isolation criteria applied in the analysis. This does not increase the background contamination because the muons are constrained to form a Z; studies were made to compare non isolated muons and muons using relative isolation and no difference was observed in the mass distributions. In general, with the current muon algorithms some loss of signal efficiency in the high mass region is expected, however it is much less than for the electron case.

For the electron channel, we use DoubleElectron triggers. An additional complication is that in the offline selection we use GSF electrons with very loose ID criteria (WP95). The isolation criteria are applied at both the trigger level and the offline selection. In general, the electron channel observes larger inefficiencies for resonance masses greater than 1.3 TeV compared to the muon channel.

In the 2011 analysis this issue was not considered very important becomes that the problem becomes relevant in a mass region in which we are not very sensitive. For the 2012 analyses, there have been developed (in the context of other boosted searches) two different double lepton triggers (electrons and muons) with special isolation criteria to address this issue. Additionally, the POGs are now focusing on improving the reconstruction of high-momentum and non-isolated leptons. This issue will be better studied and monitored in the 8 TeV update of the analysis.

- Table 1: The k factors here are considerably larger than what we have used before in eg for G-> gamma-gamma in the 2010 analysis, which was based on a dedicated analytical calculation. Is FEWZ actually the right tool to calculate these factor for this process? Has FEWZ been updated to include those particular processes?

The k-factors for the RS used here are the same used by the RS->gamma gamma analysis in 2011 CMS analysis. They were not produced with FEWZ. It is the W' k-factors that were produced with FEWZ, because the assumption in SSM is that the W' couplings are the same as the (SM) W couplings. The W' k-factors used here are common with other CMS W' searches (e.g. W'->lnu and W'->WZ->2lnu).

The text was updated to avoid any confusion.

Changed

"Mass-dependent k factors for the next-to-leading order (NLO) and next-to-next-to-leading order (NNLO) corrections, calculated with FEWZ [29], are applied."

to

"Mass-dependent k factors are applied. For the GRS analysis, next-to-leading order (NLO) corrections are calculated using the two cutoff phase space slicing method [29, 30]. For the W' analysis, the next-to-next-to-leading order (NNLO) corrections are calculated with FEWZ [31]."

- Table 2: why are so many cross sections only at LO for V+jets? we have NLO cross sections for eg W+5 jets now?

NLO k factors don't exist for high-pT madgraph samples. We don't expect the k-factors of the inclusive sample to hold necessarily in the tail of the pT distribution. Having said that, these values are given as a reference but not used directly in the analysis. We use a data-driven method for the background determination.

- line 127: what is the effect of an isolated track cut effectively on the data? Do we reduce the number of events significantly? It is one of these cuts where one like to make sure that it is under control, systematically.

This cut is quite standard and it is one of the selection ingredients that are common with the mono jet analysis. It is employed to remove leptons (particularly taus) from V+1*jet, more than anything else. Given the high tracking efficiency and the good MC description of isolation variables at CMS (NB: we use standard tracking isolation here) we did not feel that a dedicated study or customized solution was required for our analysis. Note that this does not affect our background prediction, as it is determined with data-driven techniques. For the signal, we checked with MC that the efficiency of this cut after the other selection cuts, is larger than 90%.

- lines 182-184: is the MT cut here really as given in this interval or is this a typo? The signal part is > 900 GeV so I expected the sideband rather to be < 900 GeV

Sorry, it was a typo. Corrected.

- line 189: this correction factor is rather large, not close to one. Is that easy to explain?

The simplest way to explain this is that for the background the jet mass and MT are correlated, and the value of the rho parameter ("the correction factor") reflects that. In some more detail, the mass of a QCD jet is a function of its energy (for details of the calculation, see Ellis:1991qj). The "extra" jet in the W+jets and Z+jets background is from QCD processes (extra parton emission), so the above reasoning applies here. The jet energy is entering the calculation of the jet-MET transverse mass. Therefore, the two parameters are expected to be correlated.

- Figure 3b does not look very nice. There seems to be some general deficit in the data at high mass… Is that no worry? Of course you wont' have a fake signal, but perhaps something does not work so well in the closure?

In the closure test we only considered the comparison between predicted bacgkround from data and MC in an integrated fashion. In this comparison we found that the scaled Monte Carlo is reproducing the data reasonably well. In the particular high-mass region the expected background from Monte Carlo is 6 events and we observe 3 events in the data. We think that the discrepancy is not significant enough to suggest a problem in the closure test.

- line 230: It is said that the statistical fluctuation are around 15%? Why are these so large?

This number is produced by using the statistical uncertainties in the event counts which are used for the rho calculation. For example, the uncertainty that corresponds to sideband B is sqrt(125)/125 ~ 9%. By summing up in quadrature all the uncertainties we end up with 15%.

- line 260: Where is this 9% coming from? boosted samples studied?

Yes, this selection cut efficiency was studied as a function of the boost with real data. The details of this study can be found in EXO-11-081 Analysis Note. In short, we performed a study with ttbar samples in semileptonic decays. We compared the V-mass cut efficiency in MC and data for a W from the hadronic t decay with one PF-AK7 jet using the jet mass and other criteria from the VZ analysis.

- line 290 and further: I said this already before:

For the RS KK we still limit ourselves to the region k/M_PL < 0.1 which is fine for quoting a result, but please note that the views on this limit -- traditionally used since the year 2000-- is changing:, see eg the talk of J. Gunion (page 6) given at the Higgs meeting earlier this year: https://indico.cern.ch/getFile.py/access?contribId=36&sessionId=7&resId=0&materialId=slides&confId=175490 and the reference K. Agashe, H. Davoudiasl, G. Perez and A. Soni, Phys. Rev. D 76, 036006 (2007) [hep-ph/0701186].

The values for k/MPL can go up to 1 and even above…Maybe we could test this in this paper?

For the 2D limit plot in RS models (Figure 5), we go all the way up to k/M_Pl = 0.3. Beyond this value the resonance width starts becoming noticeable, ie. it is no longer negligible compared to the experimental resolution. In order to probe this phase space region, we would have to produce and analyze signal samples with the corresponding (increased) width/couplings and check how the efficiency chages in this k/M_PI regime. It is too late to do this for the 2011 round, but it is definitely worth incorporating in the 2012 update.

English/Style/Formatting/References %Completo_5_%

- last line in the abstract: not sure we should really stress that we are the first here. We know it irritates our competitors and we had some gentlemen's agreement on that, but the pubcom can tell you if this was just for titles of papers or also for the abstract content

Leaving for Language Editor to decide.

-line 2: this sounds a bit funny in view of the discovery of a Higgs-like particle and gives the impression that there is a disconnect in our experiment. Such models are now strongly disfavored. I suggest to find a better opening sentence. It is still of interest to look for these decays of course, eg in EDs, which in general also need a Higgs.

The main reason to search for VV resonances at high masses has been historically connected to EWSB models. Nowadays the interest for these resonances is more related to ref [11-18] as an extension of [1-10] more than anything else. We think that this is an interesting perspective to include in the Introduction. Moreover, it is not quite true that a light Higgs excludes these models. As for SUSY, it just calls for a (relatively) light partner. For instance, the case of a composite Higgs is discussed in this paper, which considers the 125 GeV mass for the Higgs and its consequences. Our main point is that these models are not (yet) dead.

- end of page 1 and page 3: footnotes were used to be forbidden in our papers. If this is still true (plse check with your pubcom chair) then we should indeed remove them.

Leaving for Language Editor to decide.

- line 76: "multiple interactions" is a term used most specifically for interactions within one and the same pp collision, not for pile-up. Hence I suggest to phrase it differently here, not to create confusion… (call it maybe multiple collisions?) This also comes back at line 147.

Fixed.

- line 77: I assume you use also a track condition for the electron? does not seem to get mentioned here.

Yes, it has now been added.

- line 84: suggest in THE isolation cone

This phrase has now been removed.

- line 86: suggest ..definition ALSO excludes…

This phrase has now been removed.

- line 91: true momentum: is this the momentum of the patron or the jet? Best to specify…

Added a clarification in the text.

- line 162: this formula falls out of air here without any explanation. Suggest to give information the general features for this choice.

Explanation added.

- line 201/210: "data driven" is slang and not used in our papers. More like "methods based on data" or alike…

This has been rephrased.

- As we compare to the boosted top decays of heavy resonates, shouldn't we also give the limit? why only discussed in the conclusion?

It is difficult to mention a single number as a comparison since several models were considered (as usual). We have added the reference to the boosted top decays CMS analysis.

Comments from Imperial College (CADI link) %Completo_5_%

Analysis %Completo_5_%

Abstract The lower limit for the RS model is given here (and later) as 700 GeV. However, in the analysis description, the lowest value used for any of the tables and plots is 750 GeV. To give a limit at 700 GeV, the analysis description should be extended to lower masses.

More generally, what sets the lower limits here? Both plots in fig 4 indicate no turnover of the expected limit around 700 GeV so it implies the analysis could be extended lower.

It should read 750 GeV for Randall-Sundrum. This has now been fixed.

The MET channel doesn't have sensitivity for masses below 1 TeV, and the lowest RS mass considered for the dilepton channel was 750 GeV. The reason for this is that the the mono/fat-jet V->qq (V=W,Z) reconstruction technique is appropriate/efficient for boosted VZ topologies which corresponds to W'/RS resonances heavier than 700-750 GeV. Below this value the fat-jet reconstruction efficiency decreases dramatically and the analysis loses its sensitivity.

Lines 17-24 Any non-stationary W and Z will give a boosted pair of fermions; the 800 GeV values seems irrelevant at first reading. This sentence needs to be clarified to explain what difference the large boost of 800 GeV actually makes, i.e. that the jets begin to merge, which is the missing critical piece of information. It would help to clarify which frame the small opening angle is in.

Also the fact that the search is in the boosted regime where the dijet/dileptons are overlapping is not well explained and the "advantage" of this fact to discriminate against the backgrounds is not well motived.

We modified the sentence

"For resonances heavier than ∼ 800 GeV, the two-fermion system produced in the decay of each V is boosted, i.e. the fermions are emitted within a small opening angle. "

to read

"For heavy resonances the two-fermion system produced in the decay of each V is boosted, i.e. the fermions are emitted within a small opening angle in the laboratory frame. The hadronization of the V → qq ̄ quarks would then produce a single jet with mass close the V mass, very different from a typical QCD jet. Monte Carlo simulations suggest that ~70% of the decays would produce a merged-jet topology for resonances heavier than ~ 800 GeV."

Lines 23-24 Does the isolation requirement problem refer only to the analysis or also the L1 trigger selection?

This is referring to the tight(er) isolation requirement applied by the standard muon selection offline. The isolation at the HLT is looser than that and we cannot in any case avoid it. We think that this is too much of a detail for the introduction, since we have not discussed the differences between trigger and offline selections yet.

Lines 29-32 The description of the two models probed: SSM and gravitons decaying to ZZ, is too minimalistic. No details are given on the models, parameter assumptions, etc. Given that it is the main result of the paper, it would be good to do a better job here. In addition, the SSM is not mentioned explicitly in the first paragraph so it is not clear how it related to the discussion there. Make sure this model is explicitly named and referenced there. However, if this is meant to define it for the reader, it would probably be better to say something like "the Sequential Standard Model in which W'... and a model in which GRS"?

Also, does it actually needs the references to Weinberg et al [1-3], and the Higgs mechanism [4-9]. For one thing the historical references do not give a complete summary of the Standard Model as it is today (e.g. QCD), even if these were the pioneering publications, and in addition what does the details of the original SSB mechanisms have to do with this paper?

"Results are presented in terms of two benchmark scenarios: the Sequential Standard Model (SSM) W′ → WZ and GRS → ZZ."

We changed the above sentence as follows:

"Results are presented in terms of two benchmark scenarios: i) the Sequential Standard Model (SSM) in which a new gauge boson W′ with the same couplings as the SM W boson decays to a WZ pair. i)) a RS graviton G decaying to ZZ. In both the scenarios we search for resonances heavier than 700 GeV, where the considered boosted topology becomes relevant. For the RS graviton we consider values of the coupling parameter k/M_Pl up to 0.3.

Concerning the second part of the comment: the main reason to search for VV resonances at high masses has been historically connected to EWSB models and, despite the limited benchmark models that we used, the result is more general. We then decided to give a broader view of the possible implications of this results. Nowadays the interest for these resonances is more related to ref [11-18] as an extension of [1-10] than anything else. We think that this is an interesting perspective to include in the Introduction.

Line 52 Give the thresholds for the single muon and double electron triggers.

The thresholds changed with time. We prefer not to give a long list of thresholds, which would just look out of place. So we added (after line 52) "The trigger thresholds changed with time, as a consequence of the increasing peak luminosity and the changes in running conditions. The tightest thresholds used in the trigger (e.g. 40 GeV for the single-muon trigger and 17 GeV for the dielectron trigger) are looser than the corresponding analysis requirements."

Line 56 In the VZ->nunu (Etmiss) channel there is no information on the trigger but it just says it is 99% efficient for events with pt> 160 GeV and E_t_miss > 300 GeV. It would be good to know which masses can be probed which such cuts and what is the signal selection efficiency that can be achieved.

We have added the trigger information for the MET analysis. We have also added "which allows to probe resonances heavier than 1000 GeV with an efficiency above 20%."

Lines 60-65 Is this GEANT tune special compared to other analyses? For the non-expert reader it does not convey much and the most important thing is probably whether the choice of events influences the analysis in any way. This comment applies to the other details cited in this paragraph which may be important to the reader intimately acquainted with this or similar analyses but does not tell the average reader a lot.

This is the GEANT version used by CMS. We refer it as in any other paper. We think that quoting software version is quite standard in CMS papers, at least for reproducibility.

Table 1 What are the uncertainties on the cross-sections and how are these derived? Some indication of how the uncertainties are fed into the limit should also be given.

These are pythia cross sections. The cross section uncertainty does not enter the limit setting calculation (Brazilian flag plot) but it would correspond to an error band on the theory lines in Fig 4. Any theorists should plug his/her line with error on the green-yellow band, which is the actual result of the paper. For the same reason, it is quite conventional (as discussed within the LPCC workshops with theorists, for instance) not to include the theory error on the limits, since other theorists might not agree with the estimate we wanted to use. This is, at least, the convention in EXO searches, where these plots come without an uncertainty band associated to the theory xsec line (which is simply given in the plot as a reference). In other domains (e.g. SUSY) this is different (there a +/- 1 sigma error limit is also quoted). We decided to present our results in a way consistent to other established (and published) EXO searches.

Line 95 Is it really true that MC JES corrections are used? Which corrections are actually applied in this study?

This is indeed not correct. We changed it to "Jet energy corrections are derived from the in situ measurements with the energy balance of dijet and photon+jet events. "

Line 98 Which Etmiss is used? Are there any corrections applied?

We use particle-flow MET. This is implicitly said, since the E_Tmiss is defined as the sum of the transverse momenta of the reconstructed particles. To make it more clear, we changed particles -> PF particles (as in line 88)

Line 123 In the Etmiss channel a single track veto of pt> 10 GeV is applied. i) How is the selection efficiency and uncertainty of this selection calculated and used in the limit setting? ii) This cut is likely to be very PU sensitive but nothing is discussed about the effect or how it is determined.

As for any other cut, the signal efficiency is taken from MC. For the data, we used a data control sample to which the same cut is applied. We checked the stability of the full analysis selection (signal efficiency and bkg prediction) vs the number of vertices and we see no big dependence. In any case, any effect is common to our background region and signal region on data. In addition the MC correction rho is computed with a PU reweighed MC. This will be more crucial for 8 TeV data, where more accurate procedures (e.g. PF isolation) will be used.

Table 2: why are LO cross sections used to normalize the main Z/WW/WZ/ZZ backgrounds?

NLO k factors don't exist for high-pT madgraph samples. We don't expect the k-factors of the inclusive sample to hold necessarily in the tail of the pT distribution. On the other hand, we use a data driven bkg prediction. These values are given as a reference but not used for the result derivation.

Line 162 What is the motivation for choosing the background function given? It looks quite non-intuitive so some explanation is probably needed. If it is purely heuristic, then state this.

The main reason to use this function is the experience from other analyses with dijet searches. This function is empirical but pdf-related, since the pdfs is what we are probing in the end. On the other hand, there is no profound reason to prefer this particular function over others. We do consider other functions for systematic effects, as was done for the dijet, hadronic VV and black-hole searches. In order to acknowledge the historical origin of this choice, we changed the text to

"Following the example of other resonance searches[], we fit the mVZ distribution to an analytical function and the result is used as the basis for the modeling of the expected SM background distribution: "

adding a reference to the dijet 7TeV paper.

Figure 2 The fits in the low mass region 400-800 GeV do not match well to the data or the expected background. Something needs to be said about this.

We have added a sentence

"The fit describes well the tail of the m_VZ distribution, which is the region of interest for the new resonance search. A discrepancy is observed at low m_VZ values. Any modeling imperfections, quantified as the difference between the best-fit function and the data distribution in the m_J sideband, are taken into account by assigning a systematic uncertainty."

GENERAL comment More confidence in the background method would be achieved if the curves would predict the low mass WZ/ZZ continuum which CMS has measured. Has this been tried?

The background-determination method that we are employing was originally (and successfully) used in the H->jjll analysis. We believe these questions were successfully addressed in that context.

Lines 178-185 Why are the MT sideband regions (C and D) so far off the signal regions in terms of the MT cut, i.e. 20-70 GeV compared with >900 GeV? This is a long lever-arm to assume proportionality in the signal region; it would seem more robust to use e.g. 400-900 GeV (or is this a typo?).

It was a typo. Corrected now. The sideband for the m_T is 700-900 GeV.

Line 214 It is stated that the dependence on PU is negligible but no details are given. This should be better motivated.

Details are in the AN. We think that the paper is already long enough as it is. If this was a 8 TeV paper this would be more relevant. But since this is the 2011 low-PU run we think we can avoid going in the details. This is in any case discussed in the systematic section

Figure 3 The vertical axis needs to be reduced and the legends improved.

Fixed.

Section 5.2 In the ETmiss channel mj and MT(jet,MET) are used in a ABCD method but there is no discussion on the systematic uncertainty of the correlation rho of the two variables. Has this been estimated?

This is discussed in Sec. 6.1.2.

Lines 267-269 What is the systematic for the 750 GeV point? This is not mentioned in the text here or in table 7 but a result for this mass is given in table 10. This needs to be made consistent.

This was an oversight and has now been fixed.

Lines 288-291 How are the limits betw

Edit | Attach | Watch | Print version | History: r43 < r42 < r41 < r40 < r39 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r43 - 2012-11-04 - ChristosLeonidopoulos
 
    • 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 &© 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback