COMBINATION

General comments:

- Very interesting analysis and very clear and convincing paper, nicely written. - The explains how the final limits are derived needs to be elaborated

Congratulations to Authors and ARC.

Type A comments

o) l 5 comma after Lambda

o) l 6 : s-hat not defined

o) l 6-7 For parton interaction etc – please consider the following rephrasing:

For values of the parton-parton centre-of-mass energy shat much smaller than Lambda, the metacolor force will…

o) l 12 with eta values –> with the eta values

o) l 13 Lower limits –> Conventionally/Traditionally, lower limits…

o) Figure 1, caption: Strictly speaking the figure shows more than just the 'interference'. How about, for instance:

Schematic representation of the total DY and CI cross section.

o) l 17 described as –> written as

o) l 21 CI/DY is confusing: We are adding two terms, so why don't you use “CI+DY” or “|CI+DY|^2”

o) l 22 comma before 'which'

o) l 22/24: I would have put the reference [6] in line 22 after LLIM instead of in line 24, since the former is the first time the model is introduced.

o) l 25 initial-state quarks [hyphen]

o) l 27 please consider the following slight rewording:

In this section, the basic features of the LLIM dimuon mass spectra…are demonstrated by using generator-level PYTHIA [ref to pythia] samples (see Sect. 4) with appropriate kinematic cuts.

o) l 28-29: Not a clear construction of the phrase: at least add “and the pseudorapidity is defined as …”

o) l 29 from the beam axis –> with respect of the positive z direction

o) l 30: “for this example ” : which example ? this measurement ?

o) l 34 illustrate that –> illustrate the fact that

o) l 37 so that –> that

o) l 38: “is observable” → “would be”

o) l 40: Confusing: what is “discussed below”? “previous measurements”?

o) l 43 constructive and destructive cross sections –> constructive- and destructive-interference cross sections [this is jargon I find.]

o) l 49-54: eliminate this paragraph, and insert these lines below line 42

o) l 71 using full –> using the full

o) l 77 cosmic-ray background [hyphen]

o) l 77: Here the pT cut on the muons is >45 GeV while on line 31 was >40 GeV. Anyway it is sufficient to specify the cut once, better in the “dimuon selection criteria” paragraph. Please clarify.

o) l 84 cosmic-ray muons [hyphen]

o) l 84-5 3-dimensional angle –> angle ? [it is a normal 2d angle in the plane that contains the two muons – or do I miss something ?]

o) l 94-5 The predicted number of CI/DY events –> The predicted number of observable CI/DY events ?

o) Table I : it does not seem to be appropriate for a paper. The invented names of specific MC production at a given mass (PYTHIA20, PYTHIA120..), or the number of generated events, or the generated luminosities for the samples are really not important. Consider removing it, and explaining how the MC sample was built in the text, or making the table smaller, just with MC generators used for what signal, and their order.

o) Table 1 (caption): “Description of event samples with detector simulation”: why so contorted? The classical “Monte Carlo samples” is old-fashioned?

o) l 114 pileup is mentioned here for the first time, and should perhaps be written in quotes.

o) l 114: I would have put ”, i.e. multiple p-p…” instead of ”:”

o) l 122-125 I think CMS has a preference for impersonal, passive sentences (we determine a …factor –> a… factor is determined) [possibly also elsewhere] o) l 123-129 : confusing

o) l 132 are given –> are also given

Fig. 5: is it really needed?

o) l 130 'QCD background' is jargon – perhaps we should specify more explicitly which processes we mean ?

o) l 137-9 The non-zero backgrounds are ttbar,…, and Z –> tautau. –> The non-zero backgrounds are ttbar,…, and Z –> tautau production.

o) l139: the statistical → the realtive statistical

o) l 158 greater than – in words

o) l 169 limit-setting procedure [hyphen]

o) l 170 integrated luminosity –> the integrated luminosity

o) l 172-3

The uncertainty in signal and background is dominated by the effect of PDF variations which are evaluated using the PDF4LHC procedure [35]. –> The uncertainty in signal and background is dominated by the effect of the PDF uncertainties, evaluated with the PDF4LHC procedure [35].

o) l 174 + or - –> positive or negative

The PDF variation –> The PDF uncertainty

o) l 185 …the individual…the limit

o) l 186 for an explicit change in each uncertainty, or central value to which it corresponds. –> for a given change in each uncertainty, or central value.

o) l 188 and 190 constructive limit –> constructive-interference limit

o) fig 7 not clear if the result is used only to find the most sensitive value of Mlow. If this is the case, the figures can be removed, because they will confuse the reader.

o) l 196 standard-model sources [hyphen]

o) l 201 improvements on the current published values of 4.5 TeV and 4.9 TeV. – please consider repeating the relevant references at the end of the sentence

Type B comments

o) l 31 Just to be sure: are these the only cuts applied for Fig 2 ?

o) Fig 2 a) & b) are too similar, and it's difficult to appreciate the differences. We suggest to keep only one of the two plots, and maybe in the second one to plot one or two examples of the same curve for destructive and constructive interference. Caption: Values –> Yields; toward pure DY –> to that of pure DY

o) l 102-105:

Q1) This sounds not so clear to me. Reading “Some of the reconstructed events have generated mass below M_low due to the mass resolution” one may think that the there are events that don't go in the denominator of AxM but go in the numerator, hence AxM>1, but from fig.4 it's the contrary. Am I missing something?

Q2) Is there any matching between generated leptons and reconstructed leptons in this procedure?

Q3) In the numerator the events are reconstructed with the full selection while in the denominator only the mass cut is applied? This would justify the “acceptance” part in the name… Or is it only a pure mass acceptance term?

o) Table 2 second column and Fig 4: the values in column 2 can be improved by fitting the points and using the fitted values. For example, the values at 500 and 600 GeV are identical in the column, but they should not be, it's a statistical fluctuation. We suggest to Fit Fig 4 and replace the values of column 2 with the fitted values.

o) Table 2, third column: the k-factors have very small errors, which are due to the very large samples used. However, are we sure that the k-factor does not have additional source of errors? It would be nice to show the theory uncertainty together with the statistical error, or at least to discuss the problem.

o) l 108 'that are simulated in the detector' – I am lost. From the initial part of the sentence I think I understand that the sample of generated events is now different (CI/DY as opposed to DY). But I am confused by this last part – are these events generated only inside the detector acceptance ? if not (ie generated over the full phase space), I would suggest removing this last part of the sentence.

o) l 123-125 A naive question from a non-expert: why is Pythia used and then a k-factor is determined by comparing to MC@NLO ? why not use MC@NLO directly ?

o) l 127-129 Is the same k-factor used for DY and other processes ? not sure about the meaning of this sentence.

o) Figs 2 and 6 the various lines are difficult to distinguish in black and white

o) l 147-150: There should be some statistical procedure that allows to put limits taking into account possible negative fluctuations of the background; especially because in these kind of analysis the signal region is always so poorly populated that this fluctuations are very likely