Comments from FR and other edits w.r.t. v8 of the paper draft

Typed minutes of the Final Reading:

Edits requested during FR

L 209: --> to suppress the background from low mass hadrons.

Done. Sentence now reads "All pairs of oppositely charged leptons, regardless of flavor, in the ZZ candidate are required to satisfy ${m_{\ell \ell'} > 4 \textrm{GeV}}$ to suppress backgrounds from low-mass hadron decays."

Title: ...couplings in events with... --> ...couplings from events with...


Abstract: ...couplings in events with... --> ...couplings from events with... and L6 remove (j)


L 144-146 Check primary vertex definition

Text was correct.

Add information on size of QCD background

Modified the sentence starting L38; now reads: "Despite a low cross section, a small $\mathrm{Z}\rightarrow\ell\ell$ branching fraction, and a large irreducible QCD background, this channel provides a favorable laboratory to study EWSB because all final-state particles are reconstructed."

A lot of discussion about the message the word "optimized" conveys (in lines 281-286).

Propose modified paragraph: "A total of 36 discriminating variables including observables sensitive to parton emissions between the tagging jets, the production and decay angles of the leptons, Z bosons, and tagging jets as well as quark-gluon-tagging information were considered in the BDT training. Observables that do not improve the area under the signal-versus-background efficiency curve (AUC) were removed from the BDT. The observables sensitive to extra parton emissions provide little marginal AUC increase and are not retained because of the limited modelling accuracy in the simulation. The tunable hyper-parameters of the BDT training algorithm are optimized via a grid-search algorithm. Finally, the BDT performance is checked using a matrix element approach [50-52] that provides a similar separation between the signal and background processes."

L 327: Sentence in the twiki found to be better that the sentence in the paper: "The unitarity bound is determined using the VBFNLO framework [57] as the scattering energy $m_\textrm{ZZ}$ at which the aQGC coupling strength set equal to the observed limit would result in a scattering amplitude that violates unitarity."


L 80-81: Remove sentence about calo jets.


L 257 Remove the last part of the sentence ("only affects the overall yield predictions of the estimates on simulation"). Just state the luminosity uncertainty.


Move "As an illustration from L267" (in fact Figure 2 shows the full distribution) to L 268, when Table 1 is referred (and these cuts are applied). Add also "As an illustration" to caption of Table 1.


L 8 Remove "also"


Reconsider tick frequency in figure


L2 to be changed to "Weak vector boson scattering (VBS) plays a central role in the standard model (SM) and is a key process to probe the non-Abelian gauge structure of the electroweak (EW) interaction."


L4 ...absence of any regularization mechanism,--> ... absence of any other contributions,


L 7-9 Remove commas after [3, 4] and [5}. Suggests --> provides evidence. Remove "also"


L 11 Substitute "diagrams" (when referring to interference) by "amplitudes"


L 13 allowing for --> permitting


L 14 cross sections of --> cross sections for


L 21 Changed to "At the LHC, VBS is initiated by quarks q from the colliding protons; both quarks radiate vector bosons (V = W, Z) which then interact."


L 24 remove parenthesis (j) --> j


L 30 signal and resulting --> signal, resulting L 33 remove (pp)


L 34 cross sections of --> cross sections for


L 35 remains --> remained and


have been --> are L 38 due to --> and


L 44 remove the last part of the sentence ", which is the main interest in the investigation of EWSB."


Figure 1 caption: Substitute "diagrams" by "amplitudes".


L 335 in the four-lepton final state --> in the four-lepton and two-jet final state (with a hyphen??)


L 339-340 was measured to be --> is


L 340 Be consistent everywhere in the paper, use the same notation than in the abstract, i.e. remove fid (subscript), put EW (subscript), include pp->ZZjj->lll’l’jj in the expression. Check also L 312 and be consistent also there. Include the SM prediction, after a comma and in present tense.


Request from Sijin to use the short acknowledgements section as this is a Letter.


Additional edits

Steven Wasserbaech

In Table 1, "ttbar" should be in roman, not italics.


lines 219-221: We say we're "taking" the yield and shape from simulation, but "constraining" them with the data; perhaps we should add to the end of this sentence ", as described in Section~7."

Done, now reads: The yield and shape of the multivariate discriminant of this irreducible background are taken from simulation, but ultimately constrained by the data in the fit that extracts the EW signal, as described in Section 7.

Nevertheless, I can't imagine why it is necessary to say the 120 GeV cut twice. If I'm not mistaken, line 214 should read "Finally, both Z1 and Z2 are required to have mass greater than 60\GeV." In any case, I think "shall" is peculiar and "a mass" is incorrect.

Propose rewording "Finally, both $\text{Z}_1$ and $\text{Z}_2$ shall have a mass between 60 and $120~\text{GeV}$." -> "Finally, the $\text{Z}_1$ and $\text{Z}_2$ candidates must have masses between $60$ and $120~\text{GeV}$."

Sijin Qian

(4) L167, to be consistent with the expression on L209, it can be shortened from "to be half the sum of the transverse momenta of all charged particles in the" --> "to be half the pT sum of all charged particles in the"

  (c) L277-278: (to be consistent with the expression of L209): "and the scalar sum of the tagging jets transverse momenta." -->"and the scalar pT sum of the tagging jets."


  (d) L280: (similar as the item (c) above) "normalized to the scalar sum of the pT of the same objects [49]." -->"normalized to the scalar pT sum of the same objects [49]."


(5) L229-230, L249, and Table 1's header row. To be consistent with the good examples in this paper (e.g. on Figs.2 and 3's legends (the last lines), etc.), the spaces before and after the symbols "+" in all the expressions of "xxx  +  jets" and "Z  +  X", etc. should be removed, e.g. L229-230: (four places) "referred to as Z  +  X, is predominately composed of Z  +  jets events, with minor contributions from ttbar  +  jets and WZ  +  jets processes."-->"referred to as Z+X, is predominately composed of Z+jets events, with minor contributions from ttbar+jets and WZ+jets processes." Other places where also need to be changed by the similar way are L249, and Table 1's header row (the 4th column).


(9) The "year" number should be given for Refs.[12], [17] and [18]. 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.


(11) L517, in [38] "[38] R. Fruhwirth, ..." --> "[38] R. Fru(with two dots on the top of "u")hwirth, ..."


(12) L533, in [44], to be consistent with all other CMS papers, the document name should be changed from "CMS Physics Analysis Summary CMS-DP-2016-020, 2016." --> "CMS Detector Performance Report CMS-DP-2016-020, 2016."

Changed to "CMS Detector Performance Summary CMS-DP-2016-020, 2016."

(13) Ref.[48] is identical with [8], thus should be removed.


(14) L567, in [57], the 3rd word in the article title may should be in the lower case, i.e. "VBFNLO: A Parton level Monte Carlo for processes with electroweak" -->"VBFNLO: A parton level Monte Carlo for processes with electroweak"


Other minor edits

Minor typographic fixes in the bibliography.
L169: A lepton is considered isolated if $R_{iso} < 0.35$. → Leptons with $R_{iso} < 0.35$ are considered isolated.

Pre-FR comments

All comments refer to v8 of the paper.

Provided response to comment, paper text unchanged.

Provided response, paper text changed.

Propose to discuss at FR.

Helen Heath

Type A:

L14 Why "large invariant masses of the boson" rather than "large boson masses"? Is this trying to convey something more than the boson having a large mass.

Changed to "large diboson invariant masses"

L14 "of VBS" -> "for VBS"

Not done.

L21 This line confused me because I thought you'd define VBS as meaning "massive vector bosons" at the start of the section. going back to L2 I can see that the aim is that VBS means vector boson scattering. Is is possible to rearrange L2 so this is more obvious?

Changed to "Weak vector boson scattering (VBS) ...".

L38 "this channel .... EWSB" -> "this channel is a good channel to use to study EWSB"

To be discussed at FR.

L105 "samples of the" -> "samples for the"


L140 "events which" -> "events that"


L155 "selected among" -> "selected from"


L205 "All pairs of oppositely charged leptons in the ZZ candidate are required to have m_ll' 0 > 4 GeV, regardless of flavor". It's not obvious to me why this is included in the Z selection. I wonder if it's worth inserting one sentence.

We propose: "All pairs of oppositely charged leptons in the ZZ candidate, regardless of flavor, are required to have mll > 4 GeV in order to supress the backgrounds from low-mass resonances mediated by the strong interaction."

L248 "a factor"-> "factors of" currently it sounds like the factor is 2.5


L278 "Also exploited in the BDT"-> "The BDT also exploits"


Nancy Marinelli

Type B

Line 162: Maybe a short sentence about particle flow and particle identification would be appropriate here. Particle flow candidates are actually mentioned in line 183 for the jets. So maybe just moving a little the statement would do the job.

We propose to add a PF sentence before L143.

Line 189: [..] combining several channels and methods. Adds no specific information. I would rather stop the sentence at simulated events. Otherwise I would expand a little what channles means.

Stating that several channels and approaches are used in the JEC derivation provides usefull information on the sophistication of the method.

Line 196: While reading here I was a little confused by the way the selection is described and wrote the following question. You say that the event must contain AT LEAST 2 Z, which means that there must be at least 4 leptons. But that implies that you can have more than 4 in the event. But then in line 197 you state “among the 4 leptons etc etc. So what is it that you are doing ? You require exactly 4 leptons or you allow for more than four (as it sounds from the first line) ? By the time I reached line 209 I got the reply to my question. So you look indeed to more than 4 leptons. Maybe this can be stated more clearly around line 196.

It would be strange to state at the beginning of the ZZ selection something that we do not do (veto events with more than 4 leptons). We think think the current formulation is clear.

Line 200: Maybe naive question, still: why electrons must be separated from muons by DR=0.05 ?

Its a technical cut to remove pathological cases where a muon track and an muon (FSR) ECAL cluster give rise to an electron.

Line 205: What is the purpose of this cut ?

We propose: "All pairs of oppositely charged leptons in the ZZ candidate, regardless of flavor, are required to have mll > 4 GeV in order to supress the backgrounds from low-mass resonances mediated by the strong interaction."

Type A

Summary Line 337: I would repeat here the SM value

To be discussed at FR.

Philippe Bloch

L27. you mention VVjj but this is actually only ZZjj

This introduction is meant to be general, while the specific VBS channel presented in this paper is given later in L35.

L144-146. I am confused by this sentence which talks about jets of charged particles only, while lines 183 and following mention that jets are reconstructed from PF (and include therefore neutrals).

L144 is the PubComm recommended text describing the primary vertex finding algorithm. The jets mentioned here are not used in the analysis. The analysis uses regular PF jets, including the neutral component.

L205: may be one could explain what this cut a 4 GeV is fighting against ?

We propose to change the text to: "All pairs of oppositely charged leptons in the ZZ candidate, regardless of flavor, are required to have mll > 4 GeV in order to supress the backgrounds from low-mass resonances mediated by the strong interaction."

Caption Figure 4: The last bin includes all contributions with mZZ > 1400 GeV. It is confusing: do you mean includes ALSO the contributions with Mzz > 1400 GeV ? Would it not be more correct to say : the last bin includes all contributions with Mzz> 1200 GeV.

Changed to "The last bin includes all contributions with Mzz> 1200 GeV."

Dick Loveless

Type B

lines 25-30: Suggest moving the sentence "The hard interaction . . " after the sentence "Figure 1 shows . .". Then write "The hard interaction in VBS involves only the EW interaction, although the qq -> VVjj process can also be produced via the strong interaction, resulting in a large background to VBS searches." It's not clear what "irreducible" or "reducible" mean since you ultimately find a phase space region that allows you to "reduce" the QCD background. It would be good to explain here that the QCD background is very large and a major problem for the measurement.

We prefer the current ordering of the sentences, as it provides a transition VBS→EW production→ Feynman diagrams. We cannot write 'large QCD' backgrounds as it is not generally true (QCD is heavily supressed in ssWW VBS). Is it not the case that the sentence provides an indirect definition of 'irreducible' as the same final state but mediated via the strong interaction? This is the understanding of the reducible/irreducible specifiers used in this text.

line 91: Please add a reference to the new CMS trigger paper: JINST, 12 (2017) 01020.


lines 142-146: This is a good location to discuss particle flow and add ref. [40].

We will a sentence on PF.

line 205: Why is there a requirement that m_(ll') > 4 GeV? Does it matter?

We propose: "All pairs of oppositely charged leptons in the ZZ candidate, regardless of flavor, are required to have mll > 4 GeV in order to supress the backgrounds from low-mass resonances mediated by the strong interaction."

lines 251,252: It is confusing to write of "signal-rich" and background- (signal) rich". What is the "signal-rich region"? What is the background-rich region? Presumably this refers to constrained areas? Can you be specific?

Rephrased to say low/high values of teh multivariate discriminant.

lines 281-286: Can you be more specific about "optimized". Maybe make a list of the "optimized variables".

The choice of variables was optimized by considering some 36 candidate oberservables, addeng them to the BDT training in groups of 3-6 variables and then retaining the groups that brought significant improvements. This is followed by a pruning of the input observables: assuming there are N observables in the BDT, we retrain the BDT N times and identify the observable that add the least amount of separation power. The pruning stopped at the final list of 7 because the losses in separation power start being appreciable when dropping any of the final 7. These are the observables listed in L275. The details of this optimization are in AN2017-002.

line 327: This sentence is confusing. Change "as" to "because". What do you want to say?

The sentence reads: "The unitarity bound is determined using the VBFNLO framework [57] as the scattering energy mZZ at which the observed limit would result in a scattering amplitude that violates unitarity. " The unitarity bound quoted in Tab 2 is indeed determined as the value mZZ at which unitarity is violated. Now this depends on the coupling value in the aQGC model (unitarity is violated for lower scattering energies if teh coupling is large). We set this coupling value equal to the oberved limit we set. We propose: "The unitarity bound is determined using the VBFNLO framework [57] as the scattering energy mZZ at which the aQGC coupling strength set equal to the observed limit would result in a scattering amplitude that violates unitarity."

Type A

Title: "Search" seems too weak for the contents of the paper. You actually make a measurement, so it seems appropriate to say "Measurement of vector boson scattering . . ".


Abstract: Same thing. At the end of the abstract you say "This is the first investigation . . ". Saying "This is the first measurement . . " is a stronger and more precise statement.



line 21: As written the sentence says the protons both radiate vector bosons and ??. Suggest writing "At the LHC, VBS is initiated by quarks q from the colliding protons; both quarks radiate vector bosons (V = W,Z), which then interact.". Instead of a semicolon you could make it two sentences. No parentheses around "q".

Current formulation kept, as it is stated that it is the quarks that initiate VBS.

line 24: Delete parentheses around "j".

Not done.

line 25: Change "and little hadronic activity between the two jets [8,9]." to "with little hadronic activity other than the two jets [8,9[.". I think you mean "other than" instead of "between". What if there is hadronic activity "outside" the two jets? What does "between " mean? In eta? In phi?

Current formulation kept.

line 38: Change "understand" to "measure". The goal of this paper is to measure EWSB. "understanding" requires a broader view with more channels.

Changed to "study EWSB"

line 39: Delete "reducible". See type B comment line 30.

We define reducible, and think its a very usefull distinction.

line 57: No parentheses around "eta".

Parenthesis kept.

line 62: Delete "of pt".


line 103: Add a comma after "24]".


line 113: Delete "a" before "good".


line 117: Change "employing" to "which employs".

Not done.

line 129: Change "with" to "which has".


line 140,141: Change "which" to "that" and delete "described below" or at least put parentheses around "described below".

Done. We cannot drop the "described below" as the ZZ selection is undefined at this point in the text and putting it into parantheses would reduce readability.

line 143: Change "summed physics-objects pt" to "the sum of pt^2 of charged tracks".

Changed to "? with the largest value of sum of physics-object $\pt^2$ is taken to be the ?"

line 144: Write "Jets are constructed using an algorithm [35,36] applied to all charged tracks associated with the vertex as well as the corresponding pt^(miss).".

The proposal changes the meaning of the sentence, we stick with the PubCom text.

line 155: Change "among" to "from".


line 171: Add "data" before "efficiency" and delete "in data".

Not done.

line 174: Change "in" after "scale" to "for".

Similar to above.

line 175: Delete the comma after "sample".


line 183: This is a similar, but more detailed, statement as line 144. It would be nice to combine them.

The paragraphs refer to different concepts: one is the PV finding which uses a simple track jet-clustering and the second one refers to the actual jets used in the analysis.

line 185 - 187: Change "while reducing" to "and to reduce"; parallel construction - to assure ... and to reduce ..., loose identification . . ". Then move the basis of the criteria behind the noun (criteria) it defines. Write "In order to assure a good reconstruction efficiency and to reduce the instrumental background and the contamination from pileup jets, loose identification criteria [41], which are based on the energy fraction carried by charged and neutral hadrons, are imposed on jets.".


lines 189,190: What does "combining several channels and methods" mean? Either write it more clearly or delete it. In fact, this sentence combines nicely with the next sentence: "Jet energy . . simulated events to account for the effects . . . .simulation.".


line 194,195: Reverse the order the phrases so it reads "In order to ensure that the jets are well measured and to reduce the pileup contamination, all jets must have pt > 30 GeV.".


line 220: Again "irreducible". Suggest changing "irreducible" to "large". Delete "the" before "simulation".

Not done.

line 221: Here I suggest replacing "reducible" with "small".

Not adapted.

line 225: Change "reducible" to "additional". How about writing "Additional backgrounds aside from events where heavy-flavour jets produce secondary leptons or where jets are misidentified as leptons.".

We strongly prefer keeping the well-defined disctinction between the two types of backgrounds.

line 227: Add a comma after "background". How small is "very small"? Say "negligible" if this is true. Change "compared to" to "compared with". If you are comparing "like" things, use "with". If you are comparing "unlike" things use "to". i.e. Shall I compare thee to a summer evening.".

Not done.

line 228: Delete "after the selection". If it is negligible compared with the signal that should be sufficient.

We prefer to keep it.

lines 229-233: Can you find a better word than "reducible"? Maybe "non-QCD".

We prefer to keep it, and it is defined in L39-40.

line 250: Add a comma after "[45]".


line 257: Move "only" to after "affects". What are "predictions of the estimates on simulation"? Maybe just delete this phrase.

Removed the second half.

line 260: Again "reducible". Maybe change to "non-QCD". Change "samples in data" to "data samples".

Adopted second proposal.

line 265: There is an extra space before "83%".


line 266: Maybe move "Thus" to the beginning of the sentence and change "further" to "additional".


line 285: Delete "based".


line 288: Delete "A" before "good".


line 298: What does "profiled" mean? Is it important?

We use the standard procedure adopted at the LHC, with details given in the reference.

line 304: Change "being" to "is".

Not adapted.

line 307: Delete "for".

Not adapted.

line 317: End the sentence with "system". Start a new sentence with "Thus the m_ZZ . . . ".


line 319: Is "per-mass" really necessary?

Kept for clarity.

line 321: Again "profiling". Can you define this term?.

The term is used in the ATLAS+CMS paper on the Higgs statistical analysis and we provide this reference.

line 322: Maybe delete "then".


line 337: Can you add the SM prediction here.

To be discussed.

line 446 and 467: Please add years for references 12 and 18. i.e. 2016.

Neither of these preprints has been published yet, i.e., there is no date for these references.

line 465: This paper has been published - JHEP 06 (2017) 106. Done

Michael Henry Schmitt


line 7: I don't think you should call Higgs couplings "gauge couplings".

We propose: "… with couplings to gauge bosons compatible with …"

line 15: It is not obvious that VBS cross sections are model independent. In what sense?

They are model independent in the sense that only the HVV coupling enters. When studying Higgs production one always has the large gluon-gluon-fusion production mode which involves the Hff couplings.

line 42: This statement is not obvious or well known. Perhaps you should give a sentence of explanation or justification.

The spins of the fermions of a spin-1 boson decay allow to (statistically) infer the polarization of the boson. This is what we state here.

line 80-81: "subsequently used ... hadronic jets". This is not true any longer. No one measures jet energies with the calorimeter alone. Please delete this phrase, and also line 82-85: "to be compared..."

This paragraph is from the PubCom detector guidelines. We also think it is usefull to demanstrate the benefit of PF jet reconstruction which directly impacts this analysis of rather low-pT jets.

line 81: "combining information from the entire detector" - this refers to particle flow but you do not use the term. Since you mention particle flow objects on line 87, you should define particle flow more explicitly.

This sentence is from the PubCom guidelines. We prefer to not introduce PF in the detector section but in the section on the event selection. We propose to move L85-87 before L143

line 116: Maybe a dumb question, but why don't you consider WZZ and ttW production? In principle these processes can give you four leptons.

WZZ production with four genuine leptons is included in the signal sample, but supressed with the mjj>100 GeV cut. The contribution from ttW is known to be subleading to ttZ. In general one needs at least one on-shell Z to get some contribution in the ZZ selection.

paragraph 183-188 about jets: There is no statement about a minimum pT for jets. Perhaps you should provide one?

The jet pT theshold is discussed after the JEC, because the cut is done on the corrected momenta.

line 203-204: The mass limits are confusing, and if I understand correctly, unnecessary in light of line 120. I suggest to delete "is required to have a mass greater than 40 GeV " (since later you require M > 60) and also "Both mZ1 and mZ2 are required to be less than 120 GeV " since you state this again in line 210.

The text provides a description of the ZZ selection algorithm. The information on the mass limits is not redundant, as the algorithm starts with loose mass cuts on the ZZ candidates, and once the final ZZ candidate is identified, we impose the on-shell requirement. The ordering of the cuts matters for the ZZ candidate arbitration, i.e., when there are several ZZ candidates per event, notably in the 4e or 4mu final states.

line 249: Are the 10% and the 7% correlated? If so, is this correlation taken into account?

They are uncorrelated, as the production processes are very different between the signal and background.

line 257: "only affects the overall yield predictions of the estimates on simulation" - I don't understand this statement. I would expect that luminosity uncertainties also impact the measurement of the (fiducial) cross section, which is a major product of this analysis.

The text states that there is no shape dependence for the luminosity uncertainty. We do not discuss the impact of the uncertainties on the signal strength or cross section, but the resulting changes to the templates for each process.

line 270: Did you optimize the analysis with respect to cut on mJJ and Delta etaJJ? If so, you should state this in the text.

There are no cuts on mjj or Detajj for the signal extraction in this analysis. There is a very low mjj>100 GeV requirement that supresses hadronic gauge boson decays, but other than that all ZZjj events are used in the BDT analysis. We have a QCD control/validation region as defined in Fig 3 (left), but this selction only serves to cross check the QCD modelling and is not used in the statistical analysis.

line 279: I really dislike the notation RpThard. Do you need this symbol? If not, please remove it.

It is the notation used in the CMS VBF Zjj paper that we refer to. We could introduce a new/different notation but we don't have one that would be worth breaking concistency for.

line 281: How did you optimize the choice of variables?

The choice of variables was optimized by considering some 36 candidate oberservables, addeng them to the BDT training in groups of 3-6 variables and then retaining the groups that brought significant improvements. This is followed by a pruning of the input observables: assuming there are N observables in the BDT, we retrain the BDT N times and identify the observable that add the least amount of separation power. The pruning stopped at the final list of 7 because the losses in separation power start being appreciable when dropping any of the final 7. These are the observables listed in L275. The details of this optimization are in AN2017-002.

Figure 3: On the rihjt I would prefer to see the signal region, ie, mJJ > 400 and |D eta| > 2.4. The signal looks rather puny here and the plot is very similar to the plot on the left as far as background is concerned.

Plese see above comment. There are no VBS cuts in this analysis, all ZZjj events are used in the statistical analysis based on the BDT spectrum.


Title: perhaps should be "for events" rather than "in events"

We search for these events but the search is carried in the data sample of ZZjj events.

Abstract line 2: "is presented"

There are two analyses which are presented (VBS and aQGC).

line 7: "suggests" does not agree with "discovery" so you need to reword: A scalar boson was discovered at the CERN LHC and its couplings are compatible with those predicted for the SM Higgs boson, which suggests that contributions ... "

We don't see a disagreement here and prefer the current formulation.

line 11: Diagrams don't interfere (this would be jargon). Please use "amplitudes"

Not done.

line 23: remove the parenthesis around pT

We prefer the use of parentheses for definitions and abbreviations.

line 110: change "It is" to "and is" -- ie, join the two sentences.

Kept current formulation

line 116: Perhaps put "namely" before ttbar Z.

Kept current formulation

line 142: I suggest to put "CMS" before "inclusive"

Not done

line 159-161: Please reword: ... we require the three-dimensional impact parameter of each lepton track, computed with respect to the chosen primary vertex position, to be less than four times the uncertainty on the impact parameter."


line 161+1: Drop "Signal" since you have no access to whether a lepton is from signal or not.


line 171: Drop "in data" -- wordy and unnecessary.

Not done.

line 182: Please reword: ...exclude from the computation of the lepton isolation.

Reworded to "…excluded from the lepton isolation computation "

line 196: I think "formed from" is better than "formed of"


line 202: You should drop "separately".


line 213: Perhaps put "also" in front of "feature"

Not adopted.

The first paragraph in Section 6 is poorly written and not at the level of the rest of the draft. Here is a suggestion: Several sources of systematic uncertainty are considered, evaluated by varying each relevant parameter and considering the resulting changes to the distribution of the distribution of the multivariate discriminant: changes both in shape and normalization are taken into account. The impact of the variation for each source of uncertainty is summarized below.

We rephrased: "Several sources of systematic uncertainty are considered and evaluated by varying each relevant parameter. The resulting changes to the distribution of the multivariate discriminant, both in shape and yield, are taken into account. The impact of the variation for each source of uncertainty is summarized below."

line 270: Change "signal" to "purity" - adding another cut never increases the signal itself.


line 284: "are" should be "is"

Keep "are" because the sentence lists several (plus observables is plural).

line 285: Please delete "finally".


line 302: Please change "determine" to "obtain" or "calculate".

Not adopted.

line 305: I think "updated" is better than "corrected" - there's nothing wrong with the generator-level lepton momenta.

Not adopted.

line 307: Please delete "then"


Figure 4 caption: It is better to write "mZZ > 1200 GeV " since the bin also includes events with 1200 < mZZ < 1400 GeV.


Page 11: Why is a whole page devoted to this one table?

This is Latex and will be fixed when typesetting the paper in the final layout.

line 334: is -> was


line 338: are -> were


Steven Wasserbaech

Type B

abstract (7th line): It doesn't seem useful to give the value of the fiducial cross section here, as this value cannot be interpreted without knowing the fiducial region.

We think providing the cross section conveys important information (overall size of signal, uncertainties and the split into stat. and syst.). We cannot give a full account of the fiducial region in the abstract. Dropping this information would be a pity.

lines 2-9: I think this paragraph misses the target. It leaves me with doubts and questions rather than excitement about the paper. I hope we can sharpen it up.

lines 2-3: This sentence sounds false. To me the sentence means, "in the SM, scattering of massive vector bosons occurs, but it would not occur without boson self-interactions." It seems pretty obvious that massive vector bosons will scatter, with or without self-interactions! My point is that we seem to be referring to detailed features of the scattering amplitude, but the words "The scattering" don't convey that meaning.

We might be misunderstaning the comment, but isn't it the case that without gauge boson self-interaction there would be no vector boson scattering, no sizable VV → VV process?

lines 4-6: This sentence also sounds false. Unitarity always holds. If I understand correctly, the scattering amplitude for the tree-level diagrams would violate unitarity if there were no Higgs or alternative phenomenon. Perhaps a similar clarification is needed on lines 326-328?

There is certainly a difference between unitarity violation and a theory becoming non-perturbative. However, the aQGC theories are a clear examples for a case where unitarity is violated.

If we are going to mention this unitarity violation, I think we also need to explain its implications. Are we implying that this unitarity violation is impossible? Or disfavored by theorists? Or an interesting feature of the SM? The abstract of [2] says that the weak interactions would become strong in this scenario, not that it is impossible. Is "regularization" (line 4) the correct word for this sentence? Doesn't that word refer to a procedure rather than a phenomenon?

If one considers the electroweak theory of massive gauge bosons without a low-mass Higgs, the amplitudes violate unitarity at the TeV scale. In that sense unitarity does not always hold. Higer-order corrections don't change the picture, they only shift the unitairy limit. There is a difference between violating unitarity and the EW becoming non-perturbative, however the discovery of the Higgs boson disfavors the latter scenarios.

lines 6-9: I am confused by "also" (line 8). What is intended by "also"? Here we are again referring to unitarity as if something must be preserving it, but the unitarity of the tree-level diagrams is not guaranteed. Perhaps it would be clearer to first give the information from lines 10-13, and then to say that the discovery of the scalar boson provides evidence supporting blah blah standard model blah blah.

Unitarity has to be preserved for teh theory to be valid. Now there are scenarios of the EW interaction becoming non-perturbative, byt those are disfavored by the Higgs observation. If the Higgs has HVV exactly equal to the SM value, it will also fully account for unitarity conservation.

line 257: 1. It says "only affects the overall yield predictions of the estimates on simulation". The implication is that somehow this uncertainty doesn't affect our measurements, but of course it does affect some of them. 2. "overall yield predictions of the estimates on simulation" is difficult to digest. So my question is, can we delete all of this "only affects...simulation"?

We removed the second half of the sentence.

Type A

You can and should use macros (\TeV, \fbinv, \stat, \syst, \unit{fb}) in the title and abstract. (They shouldn't be used in PASes, but they're fine in papers.)


Root likes to draw far too many tick marks on plots. For example, fig 2 (left) has bins 0.5 wide, so we really don't need tick marks on the horizontal axis with a spacing of 0.2. Similar comments apply to fig 2 (right) and fig 3 (both parts). (If you want to make fewer tick marks on the vertical axes, that would be fine too!)

abstract, 5th line: "$\ell,\ell' = \textrm{e} ~\textrm{or}~ \mu$" -> "$\ell,\ell' = \textrm{e}$ or $\mu$"


abstract, 10th line: "determined" -> "established"

Keep current

line 2: It's not instantly clear that "VBS" is supposed to stand for "the scattering of massive vector bosons", and maybe we shouldn't make the reader hesitate already on the first line of the paper. We could introduce the abbreviation a little later.


line 17: comma after "bosons"


line 22: We're not supposed to use "Due to" as an adverb. "Because of" would work.


line 33: "on the fiducial" -> "on a fiducial"


line 33: "cross section of" -> "cross section for"

Discuss at FR.

line 34: "have been" -> "were"

Keep current

line 36: comma after "muons"


line 38: hyphen in "final-state"


line 42: "allow for" means "plan ahead for"; we should say "permit" instead


fig 1 caption: I don't think we need to mention "(l,l' = e or mu)" here

We think it adds clarity.

line 62: "of \pt in the range" -> "with"


line 72: ranges of numbers should be typeset with an endash, not a minus sign: "1.3$-$2.0\%" -> "1.3--2.0\%"


line 98: hyphen in "final-state"


line 103: too much space after "Phantom"; change "~ " to "~" or, better yet, "\ " because we don't need to prevent a line break there


line 103: comma after "[24]"


line 103: "and an excellent agreement on" -> "and excellent agreement in"


line 108: comma after "30\GeV"


lines 115-116: I found this sentence to be confusing; proposal: Samples for $\ttbar\cPZ$ and $\PW\PW\cPZ$ production, background processes that yield four prompt, isolated leptons and additional jets in the final state, are simulated with \MG at NLO.


line 128: "per" is peculiar; proposal: "include additional interactions in the same and neighboring bunch crossings, referred to as pileup."


line 129: "the one" -> "that"


line 137: "2~mm" -> "2\mm"


line 139: "22~\GeV" -> "22\GeV"


line 140: "which" -> "that"


line 142: "the one" -> "that"


line 143: Can we please put quotation marks `` '' around "physics-object" on this first occurrence?!

Physics-object is defined in the following sentence, making it clear as is.

lines 159-161: "to be" is clumsy; proposal: "...we require for each lepton track that the ratio of the impact parameter in three dimensions, computed with respect to the chosen primary vertex position, and its uncertainty be less than 4."


line 162: remove the comma after "hadrons"


line 179: "collect" isn't the correct word here

Rephrased to "identify"

line 209: "4" -> "four"


line 210: it is confusing to include the sentences "Finally,..." and "This selection..." in the same paragraph as the stuff about multiple ZZ candidates. Can we please start a new paragraph with "Finally"?

Keep ZZ selection as one paragraph.

line 219: comma after "jets"


lines 219-221: We say we're "taking" the yield and shape from simulation, but "constraining" them with the data; perhaps we should add to the end of this sentence ", as described in Section~7."

Not sure we can use section numbers in papers.

line 223: "kinematical" -> "kinematic"


line 227: comma after "background"

Not done.

line 248: we don't mean "2.5", so we should write "by factors of two and one-half" or "by factors of 2 and 0.5"


line 250: too much space before "[45]"; change " ~" to "~"


lines 253-254: "The uncertainties in the QCD background normalization and the jet energy scale are"


line 264: delete "the"


line 265: too much space before "83\%"; change " ~" to "~"


line 266: "thus" -> "therefore"


lines 268-269: "Table \ref{" -> "Table~\ref{"


lines 276-277:

I suggest that we reorder this stuff: "as well as the Zeppenfeld variables~\cite{Rainwater:1996ud} $\eta^*_{\PZ_{i}}=\eta_{\PZ_{i}} - (\eta_{\textrm{jet 1}} + \eta_\textrm{jet 2})/2$ of the two \PZ bosons"


line 288: "A good agreement" -> "Good agreement"


line 289: comma after "region"


line 293: hyphen in "maximum-likelihood"


line 293: Use of the jargon word "template" in CMS papers was forbidden in the early days, and I believe it is still discouraged by many members of the PubComm steering board. It adds no useful information here. "via a maximum-likelihood fit" tells it all.


line 295: plural + singular mismatch, and also "template" again; proposal: "The shape and normalization of each distribution are allowed to vary in the fit within the respective uncertainties."


line 308: remove the period from the subscript "fid."


line 317: comma after "system"


line 318: comma after "coupling"


line 319: "per-mass bin yields" is somewhat peculiar; how about "observed distribution" (or "observed distributions"?)

We think the current formulation is clear/visual on what is done in the analysis.

line 327: "as" -> "to be"


table 2 caption:

1st sentence: "T0, T1, and T2 and the neutral current operators T8 and T9."

2nd sentence: we can't just connect two sentences with a comma; proposal:

", while the unitarity bounds" or ", and the unitarity bounds"

Second sentence: Done

line 339: "operators, in units of TeV $^{-4}$:"


line 344: period at the end of the sentence

It would not improve clarity/readability.

line 352: "centres" -> "centers"



[1,2] the published titles have no hyphens in "High Energies"


[34] "$\sqrt s$=13 TeV " -> "$\sqrt s =13\TeV$"


CWR comments

Author response

Response to highlighted CWR comment

COMMENT-001 (Vassili Kachanov)

Vector boson scattering – VBS (V = W or Z) and quartic boson couplings are features of the standard model (SM) that remain largely unexplored by the LHC experiments. In the absence of the SM Higgs boson, the amplitudes for these processes would increase as a function of center-of-mass energy and ultimately violate unitarity. The Higgs boson may restore the unitarity, although some scenarios of physics beyond the SM predict enhancements for VBS through modifications to the Higgs sector or the presence of additional resonances.

This paper presents the first experimental investigation of VBS in the four lepton final state. The study is carried out using pp collisions at 13 TeV recorded with the CMS detector at the LHC. The data set corresponds to an integrated luminosity of 35.9 . Therefore, the presented results are very important and is as a significant improvement of the previous results of the CMS and ATLAS experiments.

We have several comments and propositions.

1. It could be useful to present the description of the required final state at the beginning of Section 4. Indeed, the “physical” final state should have four charged isolated leptons and two hadronic jets. However, only at the end of Section 4 it was mention that the selected event could have more then four lepton (the line 196) and should have at least two jets (the line 200). Therefore, we propose to put description of the required final state at the beginning of Section 4: “… the final state should consist of at least two pairs of oppositely charged isolated leptons and at least two hadronic jets …”.

We propose to introduce the following as the first sentence and paragraph of Section 4: “The final state should consist of at least two pairs of oppositely charged isolated leptons and at least two hadronic jets.”

2. It is not clear the situation with jets additional for two selected highest PT jets. Are any additional cuts were applied for these jets ?

No further selections are made, in particular no jet vetos. L200 states “those (events) that feature at least two jets (are considered)”.

3. In the line 126 it was written that “The highest pT electron (…) must have pT > 23 (…) GeV …”. It contradicts to the statement in the line 184-185 : “Among the four leptons, the highest pT lepton must have pT > 20 GeV …”

There is no contradiction: the first numbers are the requirements for the dilepton triggers, which are part of the list of triggers used in this analysis. The second set of numbers corresponds to the offline selection. These cuts are chosen to maximize the overall event efficiecny, which is key in this multilepton analysis. The offline cuts are still very close to the trigger plateau. Finally, the trigger efficiency is finally measured in data and very good agreement with the simulation is found.

4. It is not clear the criteria for the invariant mass of the lepton pair . Indeed, the line 191 has the statement: “… denoted and is required to have a mass greater than 40 GeV ”, while in the line 197 one can see : ”Finally, both and shall have a mass between 60 and 120 GeV

This analysis uses the same 4l selection algorithm as the HZZ analysis, to allow synchronization and to ensure maximal event selection efficiency. The text provides an accurate description of the ZZ selection algorithm used in this analysis.

5. We propose to rearrange the words in the Table 2 description: “Observed and expected lower …” “Expected and observed lower …”

Caption was modified accordingly.

COMMENT-002 (Michal Szleper)

Please recheck the unitarity limits quoted in Table 3, as the ones for fT0, fT1 and fT2 look ridiculously low. According to arXiv:1309.7890, formulae (55) and (57), if dim-8 operators are probed in the MG5 framework, a scale factor of 1/g^4 (numerically 6.36) should be applied to the respective fT0, fT1, fT2 values at the input of the VBFNLO calculator, and a factor 16/g'^4 (numerically 1526.) for fT8 and fT9. Hence, e.g., for fT0=0.44/TeV^4 I do not get unitarity violation until 2.4 TeV. Please make sure you are using the correct scale factors.

Thank you for pointing this out. We wrongly applied the conversion factor of T8,9 to the couplings T0-2. The proper limits are 2.5/2.3/2.4 TeV for T0, T1, and T2 respectively. The paper and AN have been updated accordingly.

COMMENT-003 (Albert De Roeck)

Dear proponents,

Thanks for your paper on vector boson scattering and constraints on anomalous quartic couplings in events with 4 leptons and 2 jets. The paper is basically ok for me.

My comments

- lines 12-16 seems a bit of mix up between unitarity constraints and EWSB discussion.

In the SM, the unitarization of longitudinal weak boson scattering and EWSB are intimately connected. The gauge boson couplings to the Higgs are fully specified in the minimal scalar sector. After EWSB, these couplings and the emergence of the physical Higgs field lead to amplitudes that cancel the divergent amplitudes of the longitudinal modes. The problem of unitarity violation only arises if EWSB is not as prescribed in the SM.

- line 48: These operators are discussed in [6] but to be meaningful to the reader you should consider to explain a bit more on what these operators are, physics-wise (no need to give the formula, which are in [6])

We propose to rephrase: "Finally, the selected $\ell\ell\ell'\ell' jj$ events are used to constrain anomalous quartic gauge couplings (aQGCs) which enhance the diboson production cross section with respect to the SM prediction. The aQGC limits are formulated in the effective field theory framework on the operators T0, T1, and T2 as well as the neutral-current operators T8 and T9~\cite{Eboli:2006wa}."

- line 96: We cross check with the Phantom generator. Add a sentence or two on what is particular for Phantom. Eg. Is it LO or NLO?

We propose to add the information that Phantom is a LO generator.

- lime 100 “merged…?” Merged with what? do you mean here merging of fixed order and parton shower algorithms?

We propose: “The event sample of the QCD-induced production of two Z bosons is simulated at next-to-leading order (NLO) in MG5 AMC with up to two outgoing partons at Born level and merged using the FxFx scheme. The merging scale is set to 30 GeV and leptonic Z decays are simulated using MADSPIN.”->“The event samples of the QCD-induced production of two Z bosons are simulated with zero, one, and two outgoing partons at Born level at next-to-leading order (NLO) using MG5 AMC. The different jet multiplicities are merged using the FxFx scheme where the merging scale is set to 30 GeV and leptonic Z decays are simulated using MADSPIN.”

- line 101 What is the impact of MADSPIN in this analysis? How essential is it? It is clearly good we do it but I was wondering if we know by how far we could have been misled when we would not have used it. I.E I am thinking if we should we also use it in the other analyses e.g. backgrounds for the Higgs. I do not think this is done there right now.

The nominal sample for the irreducible background is a NLO prediction that merges the 0, 1, and 2 jet multiplicities, i.e., the third jet is modelled at LO accuracy. This is a demanding 2->7 process and only made feasible by reducing it to a 2->5 process with on-shell Z bosons. MadSpin now allows to decay these bosons, restoring the Z lineshape and the spin correlations. This procedure of generating on-shell Z bosons is possible because this analysis has the corresponding offline selection. In contrast, the HZZ4l analysis obviously probes the off-shell contributions and a full pp->4l calculation is needed.

- line 122 “that…” -> the one (for the LE)


- line 128-129 Are the HLT requirements really so though (2mm)? This looks more like for off-line cuts. Just checking…

Yes, the dilepton HLT paths with dz requirement have the cut at 2mm, which is however >99% efficient for the signal.

- line 129: triplet triggers: what are the thresholds for this trigger?

The thresholds are 16/12/8 GeV with loose isolation for tri-electrons and 12/10/5 GeV for tri-muons.

- line 136: The primary vertex, mentioned here,is defined later in the paper (line 150). It would be better to put it here already.

L150 does not define the PV, but uses it to define the cut on the impact parameter significance (SIP). Because SIP is part of the object selection, we prefer to keep the structure as is.

- line 161: are these efficiencies determined with tag and probe? Could be mentioned explicitly.

We propose to add the information on the Tag-and-Probe technique.

- line 168: We should give a brief account of the FSR algorithm details here, so that the reader does not need to go to [28] to continue to read/understand the paper, and to connect it later to line 286. Details of course can be left in [28].

We propose: “A photon with pT > 2 GeV and within a cone of ∆R = 0.5 of a lepton is selected if it satisfies quality requirements.”

- line 213: To make sure I understand it: we call all these channels Z+jets, tt and WZ+ jets as “Z+X” (even though tt does not have Z’s) right? At least this is what I would concluded looking at table 1. I guess we refer to table 1 here in the text, but I did not immediately see it in section 5. (it does come only later in section 7).

Yes, ‘Z+X’ is the label for all reducible backgrounds because the dominant (80%) contribution is from Drell-Yan + jets.

- line 254: what is the sickit -learn framework? What is special about it for this application/analysis here? I know it is Python based… Just for curiosity .

The scikit-learn library implements many machine-learning algorithms and, crucially, tools to visualize and optimize the trained models. All studies on the BDT and its optimization was done in this modern library.

- lines 259-261 these lines are correct but I really understood it by going to arXiv:1410.3153 eq (2) and see the formula. Description can be improved.

We suggest add ing the vector/scalar qualifiers: “Also exploited in the BDT is the event balance RpThard, which is defined as the transverse component of the vector sum of the Z bosons and tagging jet momenta, normalized to the scalar sum of the pT of the same objects [39].”

- line 263: We talk about the limited accuracy of the QCD background but in the next paragraph we show that the agreement is not so bad. So what is done at the end to ‘account for the limited accuracy’, and how?

The sentence refers to the the choice of observables that are used in the BDT. We decided to not include any observables relating to hadronic activity beyond the tagging jets, meaning there is no third jet vetoes or other observables relating to extra emissions. These are known to improve the separation power (color decoherence is a phenomenological feature of the pure electroweak production process), but are poorly modelled. We propose to add the following sentence to make this explicit: "...QCD production processes, while taking into account the limited modeling accuracy of the QCD background processes. In particular, no veto on additional jets or other observables relating to extra parton emissions are used.".

- line 264: hyper parameters: this is typically a term of machine learning, butperhaps not too generally known yet. Do we use machine learning in this analysis (e.g. via the scikit framework)? Then I would suggest to spell that out directly and not just hide it under terminology that can be recognised only by experts

We suggest: “This optimization includes the hyper-parameters of the BDT”->”This optimization includes the tunable hyper-parameters of the BDT training algorithm”

- line 302 The Wald Gaussian distribution: is that the inverse Gaussian or what is meant here?

Its the Wald approximation, which is a Gaussian approximation.

God Luck


COMMENT-004 (Joao Seixas)

Institutional Review of SMP-17-006 (LIP)


This is an interesting analysis which lays the methodology for future improvements on the limits of aQGC. A rather innovative selection is used with the use of an MVA approach.

The text is in general good but a thorough revision of the contents and style would be advisable.

The paper would however gain from a bit further explanations about some phenomenological aspects.



T0->9 operators: While is normal to read references for theoretical parameters, the clarity of the publication would really gain from a minimal explanation about the T0, ...T9 parameters. The only thing that the reader knows is that these are effective field operators, but there is no equation, nor explanation as to explain what e.g. their non-zero value mean: Is it conformity with SM or sign of a BSM mechanism/signal ? This is even more important that Table 2 presents the final results in terms of f . Ti / Lambda^4

These are anomalous couplings as stated in the text, i.e. they allow to parametrize deviations from the SM. Any non-zero value corresponds to BSM physics.

Table 2 and related explanations: Given the provided exlanations, the reader understands/guesses the anomalous coupling parameters listed in the table. One doesn't understand however the unitarity bound " which the observed limit would violate unitarity" l308-309. This unitarity bound is defined as a scattering energy: Is it between the hard scatterers ? Furthermore, a minimal as to how this energy relates to f . Ti / Lambda^4 would be welcome. Reading the phenomenological papers would clarify the issue, but having a minimal explanation within the paper would make this latter more self-consistent as to its motivations/goals.

We believe the explanation is correct and succinct. We propose to add $m_{ZZ}$ after scattering energy and to rephrase: “The unitarity bound is determined using the VBFNLO framework [45] as the scattering energy $m_{ZZ}$ at which the observed limit would result in a scattering amplitude that violates unitarity.".

BDT input variables. l256-261: Here is the list of input variables to the BDT. The choice of some of the variables should be more justified. m_jj, |Delta eta_jj| have been justified with plots. The choice of m_ZZ is guessable. For the Zeppenfeld variable, the reader is sent to ref [38]; RpT ^hard is also referenced, fine. But for the last variable (ratio of the pT of the dijet system and scalar sum of the tagging jets' pT on one hand), there is not explanation nor reference; the reader has difficulty guessing the discriminating character of this variable.

On the same issue The problem is that, as the text is written, it is not clear whether these are the only input variables. What is written there is that the ensemble of variables includes these variables. One guesses that it is only these, but it is not at all clear. It is not clear either whether this ensemble includes obvious correlations that have nothing to do with the discrimination process at hand.

We propose to replace the word “include” with “are”, which makes it clear that only the mentioned observables are used. We provide a reference to the VBF Zjj paper for further details on the pT-balance observables, which exploit angular information between the objects.

Systematic uncertainties

230-232: "renormalization and factorization scale variations, etc.": How are these uncertainties estimated? The sentence does not read clearly. Please check/rephrase.

The standard procedure is used. We suggest: “Renormalization and factorization uncertainties are evaluated by varying both scales independently by a factor two and one-half, removing combinations where both variations differ by a factor of four, and amount to 10 (7)% for the QCD background (EW signal).”

l233-234: JER, JES uncertainties: please clarify how they are estimated.

We added a reference to the JEC paper.

l240: "A 40% yield uncertainty". Please explain how this uncertainty is estimated. How did you get to this number?

The text states that it is the combination of the statistical uncertainty in the controll regions and systematic effects due to the composition differences between the signal and the control regions.

Signal and background simulation

General point: Reporting the exact version number of the generators and software would allow others to more easily repeat the authors’ experiments and to determine whether the results might be affected by software bugs which might later be found.

The version of each code was added.

General point: Even with the exact version numbers, the study is difficult to reproduce exactly, due to the lack of details on the sample generation. If the authors do not plan to make the simulated data or generation runcards available, they should report the exact settings either here or in an appendix.

CMS generally does not publish such technical details. All the samples are generated by the official production and have been scrutinized by the GEN group. All technical details on their generation are documented in the databases and the Analysis Note.

l266: Why did the authors decide to use the BDT approach over the MEM? Having performed the MEM approach did the authors try using the MEM weights as additional input features to a new BDT?

We don’t expect any performance gain from chaining the classifiers. The BDT was chosen over the MEM because both provide identical performance but the technical overhead for MEM is considerable.


General point: Information on the BDT training, hyperparameter optimisation, and final hyperparameter values are absent; e.g. how many events were used to for training and testing, what method of HP optimisation was used, did the authors use (nested) cross-validation, what metric was used to quantify “optimisation” and “separation”?

CMS generally does not discuss such technical details of MVAs in a paper. The number of testing and training events was around 100k each and for signal and background. Grid search with ROC AUC was used to optimize the HP. Analysis Note AN2017-002 provides these details on the MVA optimization and the final HP choices.



Title & Abstract

Title & elsewhere: vector boson scattering -> vector-boson scattering

Abstract & elsewhere: anomalous quartic gauge couplings -> anomalous quartic-gauge-couplings

We follow the PubGuidelines recommendations.

Abstract: l, l’ = e, µ -> l, l’ ∈ {e, µ}

We think the current formulation is clear and uses a simple mathematical notation.

Abstract: fiducial cross section -> fiducial cross-section

We follow the PubGuidelines recommendations.

Abstract: standard model prediction-> standard-model prediction

We follow the PubGuidelines recommendations.

l6: There is not an exclusive "two jet" requirement. Please change it to -> at least two jets

The first sentence of the event selection section now states that evenst are required to feature at least four leptons and at least two jets. When discussion the BDT, we make sxplicit that no third jet veto is applied.

l10: The last 2 lines of the abstract are unnecessary, and undefined as it is. It is better to stop at this point. Please change "...are derived." -> "...are determined." The rest can be dropped.

Replaced “derived” with “determined”. The last sentence provides information on the context of this measurement, which is useful to document.


l2: massive vector bosons -> massive vector-bosons

We follow the PubGuidelines recommendations.

l2: VBS relates to Vector Boson Scattering: it does not however match the acronym as it is used in the text. Perhaps you can change it to simply "V" or "VV" (as it is usually referred to), and then also add "where V=W,Z"

This paper considers VBS as tool understand EWSB and thus defines “VBS” as the scattering of massive gauge bosons.

l4: "regulation mechanism, ..." -> "regularization mechanism, ..."


l6: “recent discovery”, it’s been five years, is it really still recent?


l11: -> ... diagrams and those that involve…

It is not clear what “those” refers to.

l13: allowing a test the -> allowing [us/one] to test the / allowing a test of the … (to be performed)

Missing word added.

l15: Higgs -> Higgs boson


l15-16: “measurements at the resonance mass”, do the authors really mean “at”, not “of”?

Rephrased to say measurements of the Higgs boson production and decay rates. (The higgs sector is usually studied at the resonance mass, i.e., by studieing the data at mH= 125 GeV.)

l17: effective field theory, hyphenate correctly

We follow the PubGuidelines recommendations.

l17: "effective field theory approach.": Please add a Reference.

Added a reference to arxiv: 1205.4231 .

l18: “increases the sensitivity” of what?

The VBS tagging jet topology enhances the sensitivity to the aQGC.

l34-35: are reconstructed -> can be reconstructed by CMS

The generic statement is about the ZZ->4l channel in this introduction, independent of the experiment.

l36: leptons candidates -> lepton candidates


l36: misidentification -> the misidentification


l41: the search is in events with 2 or more jets. This should be reflected in the statement where the "llll jj final state" is defined.

Please see previous response to this point.

3 Signal and background simulation

l92: "at LO" but then "...AMC@NLO" is quoted. Please check.

The MadGraph5 _aMC@NLO tool can perform LO and NLO calculations (like Sherpa), the name is unfortunate but correct.

l105: Does the signal sample of ZZjj contain events with more than two jets? This would be particularly relevant when understanding higher order corrections, and additional jet production from ISR/FSR.

Additional parton emissions from ISR/FSR are included from the parton shower.

4 Event selection

l129: “triplet of low pT leptons.” Please state the pT thresholds used in the trigger.

The trilepton triggers only serve to supplement the other triggers and a proper explanation would add at least two sentences that distract from the main argument.

l131: "any trigger"? perhaps: -> any of these triggers

Change adopted.

l150: please define "impact parameter significance"

Rephrased to avoid ambigous terms.

l151 and following: Do you also require a minimum distance between leptons? between leptons and jets? Please specify. -

The minimum distances between leptons are mentioned in L186/7. We will add the jet-lepton distances to the ZZjj selection paragraph.

L156-159: it is hard to follow the discussion of the "factor of one-half", as the calculation is not explicitly shown. Is this explanation really needed? or perhaps you can drop it. As it is, it is confusing. -

We think the explanation and origin of the factor ½ are stated clearly in the text.

l160 & elsewhere: “data”, state whether this refers to simulated data, or collider data.

We believe this and the next sentence make it clear that it is not the simulation.

l166-167: “Muon momenta are calibrated using a Kalman Filter approach using J/ψ and Z decays”, any citation or explanation for what this involves and how it is performed would be useful.

We will add a reference.

l164: add hyphen: -> pseudo-random

The specifier was dropped.

l168: used "to collect"? perhaps you could use "evaluate". Also, please say which algorithm is used.

We extended the explanation.

l202: "the ZZjj selection". Is this the final selection or a pre-selection? It would be better to refer it to as the "ZZjj pre-selection". Also, the names should be consistent throughout the paper.

There is only the ZZjj selection without further restrictions in the signal extraction.

5 Background determination.

General comment: The clarity of this Section would benefit from an expanded discussion.

l205: "MVA template shape" of which distribution? Please specify.

The MVA output is the distribution that is being fit. Rephrased to avoid jargon.

l206: "constrained by the data". This needs to be better described. How is it constrained to the data? Where is this discussed?

The ZZjj selection is very loose and the yield is dominated by the irreducible QCD background. The fit thus “knows” about the overall normalization of this background.

l217: "two control regions": Are they defined as one or two leptons failing ID/Iso requirements, separately? are they complementary CRs? Please clarify, and state clearly in the text.

Yes, these are two independent CR that are orthogonal to each other and the SR as stated in the text. Section was rephrased.

6 Systematic uncertainties

ll237: -> final states, respectively.


7 Search for the electroweak production of ZZ with two jets

245: please rephrase to something like -> The expected signal purity of the ZZjj EW selection is 6% ...

The sentence is clear as is and the ZZjj selection is not dominated by the EW component as written.

Fig. 2: in the caption it would be helpful to the reader to understand immediately which one is the signal in particular because Figures can be taken out and be shown independently. Perhaps you could add to the caption "ZZjj EW (signal)". Also, is this with "at least two jets"? Is there any restriction that is limiting from selecting more than 2 jets? If not, please change the labels accordingly.

We don’t think that defining the signal explicitly is standard procedure. On the jet selection issue, please see previous comments on the issue.

Fig.2 caption: Do the events in the last bins include all events outside the histograms? Please clarify

Yes, the last bins include the overflow and we added this information to the caption.

Fig.3: The above comments, also apply to Fig.3

Fig.3 caption: "full ZZjj selection". This is the first time that this name appears. Please introduce it earlier in the text. This also refers to a previous comment on L202.

There is only one ZZjj selection. We drop the “full”.

Fig.3 caption: please add -> "control region (as defined in the text)"

The CR definition is in the caption.

Table 1: “Data” -> Total observed

We think “Data” is unambiguous.

l253-254: to optimally separate -> to separate optimally

Both formualtion appear to be correct, but we prefer the current one.

l257: are the variables mjj, delta\etajj, mZZ already explained earlier? Please check.

They are in L248.

l262: to achieve the best separation". Please explain in the text "how" this is done.

We think the text makes it clear that a multivariate classifier is used, and the point raised here is the choice of input features.

l263: "taking into account the limited modeling accuracy of the QCD..." This sentence is confusing. How is this taken into account?

We choose to not exploit observables that are known to be poorly modelled.

l264: "hyper-parameters of the BDT". What is that? Please explain.

We propose: ”This optimization includes the tunable parameters of the BDT training algorithm”.

l265: " checked using a matrix-element-based approach..." Please explain in the text.

We think the description suffices to understand that this cross-check was done.

l278: "profiled" is jargon. Please rephrase it.

The ATLAS+CMS statistics note referenced in this sentence used the verb “profiled”.

l289: Please add a reference for the SM cross section prediction.

The cross section is based on the nominal MC which is described earlier.

8 Limits on anomalous quartic gauge couplings

l291: -> events in the final ZZjj selection

We prefer the current formulation, as it makes explicit that all events in the ZZjj selection are used.

l292: Please explain the meaning of these operators.

We think explaining these operators goes beyond the scope of this paper. The effect they have (increased production cross section at large masses) is stated clearly.

Table 2: Please add units to the numbers in the Table. This needs to show in each column. If there is a problem with the horizontal space, please use another line below.

The units are given in the caption, which is part of the “table object” that will be made public. Adding the units (which are unambiguous), clutters the table and reduced readability.

l311: -> for ZZ scattering in the final state with four-lepton produced in association with at least 2 jets.

See previous comment on the issue

l320-324: Please add the units.

Units are mentioned earlier in the same sentence.

COMMENT-005 (Robert Harr)

Dear authors and ARC members,

This is a very nice draft for an interesting analysis. We find that the writing is in good shape and are able to focus on physics related issues. Congratulations and best wishes for a speedy publication.


for Wayne State group.

Type A comments:

on line 13, there's a typo, "allowing a test the electroweak theory”.


lines 138-142: Twice, the tracking is referred to as “silicon tracker” and the third time as “inner tracker”. We think you should be consistent and stick with “silicon tracker” as introduced in Sec. 2.


lines 134-151: These 4 paragraphs are (1) 2 sentences, (2) 1 sentence, (3) 2 sentences, (4) 1 sentence. There could be some merging to avoid such short paragraphs.

Each paragraph describes an independent building block/physics object in the analysis. We prefer to keep them separate.

line 172: The comma is unnecessary.


line 180: “data-to-simulation”?


line 194: The comma is unnecessary.

Will check with LE.

line 214: The comma is unnecessary.


line 221: Should be “events” pluaral.

Sentence rephrased.

lines 233-234: “... and the impact of the jet energy resolution uncertainty is 8%.”


line 245: Omit “the” at the end of the line (before “QCD-induced”).


line 247: Omit “the” before “electroweak”.


line 248: Replace “as well as” with “and”.


line 316: Omit “which is”.


line 318: “... at the 95% confidence ...”


Ref. 5: Capitalize “Collaborations” as in other references.



Type B comments:

line 29: The QCD is refered to as an irreducible background, but the triple gauge vertex and the Higgs vertex in Figure 1 are not quartic couplings, which is the title of this paper. So aren't these background as well?

The paper targets the electroweak production of ZZjj, which proceeds via the diagrams mentioned/shown. Both the triple and quartic vertices are required for gauge invariance. The combination of the Higgs, double triple gauge and quartic gauge interactions are commonly referred to as the vector boson scattering diagrams since the vector bosons scatter via the exchange of a vector or Higgs boson or a direct interaction. Though in general this paper targets EW production which refers to any diagrams that result in the ZZjj final state that involve no QCD interactions. However, the three diagrams shown are the most import contributions to EWK production in the phase space that is chosen, notably the on-shell Z boson requirement.

lines 78-81: What is the point of comparing the resolutions? Isn’t that better left to a paper on jet reconstruction with CMS? And shouldn’t there be a reference on jet reconstruction with CMS in Run II?

The comparison is relevant as it reports the improvements of PF for jet reconstruction. This analysis uses PF and not calo jets. The jet paper referenced is the most recent one (we can only cite papers and other 13 TeV analyses do the same).

lines 84-85: This could be confusing since you switch from L1 decision time to L1 event rate without explaination. Perhaps just add that the L1 trigger selects events in a fixed time interval “and with a maximum rate around 100kHz.”

This is a paragraph suggested in the PubComm guidelines.

line 93: This triboson process is not one shown in Figure 1? Is there a Feynman diagram of it?

In the triboson production contributes to the ZZjj channel if one boson decays to quark jets (V->jj) and this contribution is almost totally removed by the initial m_jj selection. Similarly non resonant production of leptons is remove by the on-shell Z boson selection. Since these processes are negligible and not the physics interst of this analysis, we do not show diagrams for them.

line 102: How is loop-induced gg -> ZZjj related to loop-induced gg -> ZZ? A Feynman diagram of these processes would be nice.

They are the same, except that one has more jets from the matrix elements and partons in the final state. This a subdominant background and there is a restriction on the number of figures.

line 113: What is "underlying event" simulation? We think that this term deserves to be defined.

It is a standard technical term used, e.g., in the Higgs discovery papers.

line 120: At what stage does the pileup get added in at? It is not clear in this presentation.

We think the text is clear and further technical detail is not necessary.

lines 136, 138, 143, 152, 184, 185, 186, 187, 188, and so on: The text casually refers to reconstructed leptons, electrons, muons, hadrons, and so forth, without adding the usual qualifier of “candidates”. This should be corrected throughout.

Adding the specifier "candidate" everywhere would degrade the readabilty of the text and is not common practice. The fact that these are lepton candidtes, espeacially in conjuction with the word "reconstructed" is evident.

equation (1): As written, this expression needs some additional explanation. For instance, if a charged track is identified as a lepton candidate, but not one of the 4 representing the 2 Z decays, is it excluded from the sum? So if the event contains a semi-leptonic hadron decay, the leptons from that decay are not included in the isolation calculation, correct? Is this really what you want this to mean?

The isolation is based on all PF candidates and at this point of the text there is no notion of Z or ZZ candidates.

lines 160-167: We are surprised that there is no mention of B-field calibration in this discussion.

The calibration of the detectors (and the B field map) are not discussed in this letter. The ultimate lepton momentum precision is not crucial to this analysis.

line 173: What are the quality criteria applied? The description provided is rather vague.

We provide a description of the selection, but no numbers of the cut values, as is common (quality cuts on leptons are also not reported for brevity).

lines 176-177: We suggest that the final sentence in the paragraph would be better placed at the start of the paragraph.

We modified the last sentence.

line 192: Can the 4GeV requirement be justified, briefly. We assume it is to suppress decays from B mesons and charmonium, is that correct?

Yes, that is correct.

line 197: We find it odd that the mass requirement is mentioned last. It would seem to be an easy requirement to apply up front. Is there some logic to this?

The ZZ selection algorithm is identical to what is done in HZZ to guarantee maximum event selection efficiency. The mass window cuts are applied last. The order matters of the cuts matters and the text accurately reflects what is done in the analysis.

line 216: Please make it clear what the two samples are. Our guess is that they are (1) only one lepton fails the criteria, and (2) both leptons fail the criteria, but it is not clear.

The sentence states that there are two CR where either one or two leptons fail the selection. We will propose a rewording.

line 240: How is the 40% figure for the yield uncertianty arrived at to account for the limited number of events in the control region? This is not clear from the wording. From a naive point of view, 40% uncertainty would appear to come from a sample of 5 events (1/sqrt(5)). Is this correct?

The uncertainty on the fake ratio has to consider both the numerator and denominator, and the kinematics of the fake ratio.

Table 1: We don’t find this table very useful and wonder if it can be omitted. If it is kept, we recommend that the authors consider using different selections since from Figure 2, it would appear that m_jj>500 or 600 and Delta eta>2.4 and <5.5 or 6.0 might be a better choice.

The current cut values make it such that the VBS and nVBS selections together give all ZZjj events. The gain of tightening the cuts is rather small, as documented in the AN. Also, this changing the cut values after unblinding would bias the interpretation of the data.

line 254: Is it vital that the scikit-learn framework was used? The reference can stay, but perhaps just say that the BDT was trained and optimized.

Previous CMS publications name the respective tool (usually TMVA) and we would prefer to include the tool by name since the BDT is central to this analysis.

Sec. 8: Are allowed variations in aTGC integrated out when determining limits on aQGC?

No, the limits are derived assuming that the aQGCs are the only BSM effect. This is the standard and a very reasonable choice as argued for by the theory community, see e.g.

COMMENT-006 (Bryan Thomas Cardwell)

Ohio State University review of SMP-17-006

Dear authors of SMP-17-006,

This is a well written paper that clearly presents a very important physics result. Although statistics limited due to a large irreducible QCD background, it is the first measurement of its kind at the LHC and is clearly topical. The main aim should be to get this published as soon as possible. We congratulate the authors and our comments are minimal:

l24: “central region” is not clearly defined

We propose to rephrase as “central region”->”between the two jets”.

l31: what about other ZZ final states? are they probed? it seems odd that WW is mentioned but they’re not

To date there are only two LHC channels for VBS of massive gauge bosons: ssWW and ZZjj presented in this analysis. We will clarify the text.

l33: low with respect to WW or low with respect to ZZ->4j? i guess both are true but you should specify what you mean

Both interpretations are valid and intended to be covered by the statement.

l92: the sentence refers is confusing since it refers to a LO simulation using an NLO MC. Please clarify.

That is just the name of the tool which can do both LO and NLO.

l104-105: the sentence is also confusing since it refers to a loop induced process being evaluated at LO, using MCFM which is an NLO generator. Please clarify.

MCFM can do both LO and NLO.

l105: if MG5_AMC is available to check the MCFM for ZZjj, why is MCFM used at all and not just MG5_AMC? Please add explanation.

There are no official MG5_aMC gluon loop induced ZZ samples, only MCFM. The ZZjj sample mentioned includes the two jets at the matrix element, which is a very challenging process (it was first simulated for this analysis) and only a small sample was generated privately. A good agreement with the MCFM+PS sample (as expected from the theory) motivates the use of the MCFM sample.

l185: why not use the “electron (muon)” format as before to be consistent?


l188-198: do you account for the (probably small) Z->tautau->ll contribution or are you not sensitive at that level?

The contribution from taus is negligible at 0.6% of the total yield.

l189: referring to the 4l as a candidate is confusing, suggest for each 2 dilepton candidates instead

It is correct to speak of the 4l candidate, as the algorithm operates on Z pairs.

l230-231: what range of scale/factorization variation was performed to reach these numbers?

The standard procedure is used. We modify as follows: “Renormalization and factorization uncertainties are evaluated by varying both scales independently by a factor two and one-half and amount to 10 (7)% for the QCD background (EW signal).”

l279: perhaps signal strength should be defined?

It is defined in L252

Table 1, row 2 has central values with total bkg = 20, and ZZjj = 4 but total expected = 25. Rounding?

Yes, the background total is 20.48 \pm 4.11 -> 20/pm 4 and the signal is 4.25 \pm 0.53 => 4.3 \pm 0.5 for a total of 24.73 \pm 4.15 which gets rounded to 25 \pm 4. We will ad the missing decimal on the signal central value.

Fig 2-4: it seems strange that no errors on background are shown. are they just too small to be worth including?

We do not show the uncertainty because the final result of this analysis is the signal strength which is obtained from the fit. Not all the information in the fit can be represented visually and contarary to say a differential cross section plots, the distributions shown in this paper are not the end result of the analysis.

Fig 4: what are the vertical lines in the last two bins of m_ZZ, they look like error bars but there is no data in those bins?

They are error bars for bins with zero observed events, as recommended by the Statistics Committee.

COMMENT-007 (Francesco Navarria)

Dear authors & ARC,

This is the Institutional Review on behalf of the Bologna group. We would like to thank and congratulate you for this new paper on anomalous couplings. We do not have substantial comments, but here and there some more details/explanations could be given and the sections containing the results could perhaps be improved. In particular, the procedure for deriving the cross-section value from the signal strength, lines 282 to 287, should have a better explanation of the "fiducial volume" definition with respect to the selection cuts.



Type A English/Style/Formatting (including figures)

L4 regulation -> regularization


L8 Maybe drop "also". The boson might also be responsible for preserving unitarity, but its exchange does it.

We think the 'also' is justified.

L11 The meaning of "only" is not clear to me if I look at Fig. 1.


L13 ... allowing' for' a test' of' the electroweak ...


L17 the effect of which can be parametrized -> and the effects can be parameterized


L18 "increases the sensitivity": it is not clear wrt what -> is very sensitive

The sensitivity to aQGC versus aTGC is increased.

L19 Ref. 6, at least judging from the title, deals rather with WWjj.

The VBS topology enhances the contribution of the quartic coupling w.r.t. the tripple gauge boson couplings. The paper argues this for the case of WWjj, but the argument holds for any VVjj process.

L24, 26, 28 A very minor point: do you need to label V'V' the final state vector bosons? This is used only here, and not even in Fig. 1.


L37-39 "Furthermore, the spin ... of EWSB." This, as far as I can understand, is not discussed at least explictly later in the paper. Perhaps it would be enough to change "allows" into 'could allow' at L38.

We prefer the current formulation. This is the only place where the longitudinal component is mentioned explicitely.

L40 ... in the four' charged' lepton's' final state. (This is also a very minor point, but in principle you could have WZ scattering.)


L43 A multivariate -> After kinematic cuts, a multivariate

We prefer the current formualtion which emphazises the MVA. The kinematic cuts in this analysis are minimal.

L46 I would suggest to suppress "Finally" and to add a 'then' between "are" and "used"

A 'then' would imply a sequence of measurements, but the EW signal strengths and aQGC limits are independent analyses.

L48 In Ref. 6 I couldn't find any mention of the Ti operators. Maybe a few more details about the Tis could be given in the paper.

We now mention the effect of these operators.

L77 summed -> added

This is a paragraph from the PubComm TWiki.

L82-85 Probably "further" at L84 could be suppressed. Otherwise maybe one could rephrase saying that Lv1 decreases the rate to around 100 kHz, and suppressing "from around 100 kHz" at L85.

This is a paragraph from the PubComm TWiki.

L95-98 Something sounds odd. Since you have "yields" presumably relative to the two different programs, why don't you have also "multivariate distribution"'s'?

There is only one multivariate distribution (which is inclusive), but there are more exclusive selections with corresponding yields.

L103 are -> is (x2, the subject is the "interference")


L110 leading order -> LO


L116-117 "All MC samples ... generator." Does it mean that you don't want to use just the shape of the MC simulation, but also the absolute rate?

We use the cross section from the event generator, i.e., no k-factors from higher-order calculations are used.

L127 "dielectron and dimuon": does it mean, since we are triggering on "the same or different" flavours, L126, that e-mu triggers don't have to satisfy the condition?

The mixed-flavor triggers also have to satisfy this criterion, text changed accordingly.

L129 Is there any threshold at all on the leptons in the triplet?

There are, but these triggers contribute very little and adding the full description would distract from the main argument.

L131 "trigger" -> of the trigger conditions above

Sentence removed.

L132 I am not sure to understand what "regardless of the decay channel" means. Does it mean the OR of the three triggers mentioned above?

Phrase removed.

L132 total -> overall


L139-140 agreement -> compatibility (or perhaps better 'difference', which is something numerically defined)


L148 "small": for energetic muons the energy deposits doesn't necessarily need to be small.

The leading lepton pT in this analysis is around 100 GeV, i.e. moderate pT.

L150 "significance": in other papers the significance is defined, perhaps it could be very briefly defined also here.

Changed wording.

L183 "Z()" appears here and at L188, it doesn't seem to be defined.


L200 notably -> namely


L215-224 If I understand correctly the two control regions refer to one or two misidentified leptons, L219, but there is also isolation. Then appears "l_loose" which doesn't seem to be clearly defined. In other words, it is not clear which is its relation to the Z+ll events defined at L216. In that case at least one Z2 lepton fails isolation and identification requirements. Now l_loose passes the identification requirements, but not the isolation ones. Is this correct? Does one have two l_loose control regions? Perhaps the whole paragraph could be rendered more easily readable.

Paragraph rephrased.

L232 "BDT" should be defined here, not at L255.


L253 At this point it almost seems that you are using a different multivariate classifier, compared to the one mentioned earlier. Maybe changing "a" with 'the' could be enough. Probably what was mentioned earlier was just an anticipation, and now we have the details of the procedure.

We prefer 'the' as the concept is explained here for teh first time.

L257 To exploit the correlations, I am sure that there are no cuts on m_jj (apart from > 100 GeV, the ZZjj selection) and |Delta-eta_jj| at this stage, but perhaps it should be restated after the presentation of Table 1, that the only cut is m_jj > 100 GeV.

We prefer not repeating the selection, because the ZZjj selection is essentially the only selection used in this analysis and redfining/recalling it throughout the text might cause the impression that there is more than this one selection.

L264 hyper-parameters -> internal parameters (or just mention briefly what they are)

Sentence changed, see above.

L278 and 301 "profiled" is perhaps jargon.

See previous response.

L284 |eta|^l -> |eta^l|


L311 of Z bosons in -> with two Z bosons decaying into

The current formulation is succinct and uses a standard nomenclature.

Type B Everything else (e.g. strategy, paper structure, emphasis, additions/subtractions, etc)

L185-186 The pT cuts for the highest and second-highest pT leptons cut basically the muon distributions, as electrons have higher thresholds at trigger level (apart from the trigger with a triplet of low-pT leptons) and the cut will not be as sharp as after final reconstruction.

See previois response to this point.

L197 "between 60 and 120 GeV ": not clear why this conditions was not imposed from the start, why do you have an intermediate cut on Z1 at 40 GeV, L190? Maybe the algorithm runs that way for intermediate plots, but I'm not sure that it has to be written that way.

The ordering of the cuts matters and the text reflects this, see also responses to previous comments on this issue

L207 "irreducible": they are irreducible wrt the lepton isolation, but I guess that in general leptons from ttbar or WW will not reconstruct the Z mass. In fact at L209 you mention "background-like kinematics" which presumably could permit some reduction.

They do contribute in the on-shell ZZ selection, if one of the Z candidates is built from the leptons of the W decay. This results in low-mass Z candidates and overall background-like kinematics.

Figure 2 right. It is not easy to visualize the amount of signal in the first and second bin because of the log scale, but enlarging the figure it doesn't seem to be much. Would one really lose some signal by a selection with a 200 or 300 GeV or even higher m_jj cut? The log scale after all is good for visualizing the signal to background ratio. For instance, how does the |Delta-eta_jj| distribution look like after a 500 GeV m_jj cut?

These distributions are for visualization only. The reason to keep the low m_jj part is to allow the irreducible QCD normalization (there is virtually no signal at such low dijet masses).

L284 Given that ECAL covers up to 2.5, while the muon detectors cover up to 2.4, |eta^l| < 2.5 is not really defining a fiducial region with a sharp cut. I guess that a fiducial region should have sharp cuts for ease of comparison. This comment applies as well to the pT cuts.

The fiducial region is close to but not identical to the detector acceptance. These cuts are as stated in the text and apply to both flavors of leptons. The fiducial volumne is identical to the one used in the inclusive and differential ZZ cross section analyses.

Figure 4. What would the distribution after a cut in BDT score look like?

There is no reason to perform any further selections, as the sensitivity to aQGC can only decrease due to huge quadratic growth of the yield with the anomalous coupling and the very low SM expectation (which is almost zero).


L444 ... Nucl. Phys. 'B' Proc. …

Will inquire - other papers (e.g. HIG-13-002) omit the B.

L458 Ref. 25. Please add 'et al.' after "J. Allison". Perhaps a more recent reference, as recommended by the Geant4 collaboration, should also be added: Nucl. Instrum. Meth. A 835 (2016) 186.

Fixed. The official guidelines do not include the 2016 paper.

L478 ... 13' 'TeV ... (Please add a space.)


L481 J. Instrum. -> JINST (as done everywhere else in the paper); please change the year to 2017, suppress "92 p." and add 'doi:10.1088/1748-0221/12/02/P02014,arXiv:1607.03663.'


L484-485 The URL of the paper seems to be:

We don’t understand. There is only a link to the arXiv for this entry.

L494 Probably the reference of the Erratum should be added: Phys. Rev. D 81 (2010) 079905


L506-507 The second doi is the one of the Erratum, so perhaps it should go after the corresponding reference. The reference of the Erratum should read: Eur. Phys. J. C 73 (2013) 2501 (please add spaces and change order)


COMMENT-008 (Greg Landsberg)

Dear Authors and the ARC Members,

Congratulations on a nice search for the VBS in the ZZ channel, which sets stringent limits on aQGC! The analysis appears to be done carefully, and the paper is quite well written, so I only have a few comments.


- General: you need to decide whether you use T0, T1, etc. or T0, T1 etc. to denote anomalous couplings. Right now you use the mixture. I'd suggest the latter as a more standard way of talking about operators. That affects the Abstract, LL10-11, LL47-48, 292-293, 298, 309; Table 2 caption, L2; Table 2 body; Fig. 3 legend; and LL320-324.

We will use the "TX" notation as it is closest to the one used in the theory reference.

- L6: add the CMS long discovery paper to the set of references [3,4].

Will add

- L14: ... invariant masses of boson pairs provides ...


- LL26-27: ... to the EW production of the rmV″V′jj signature, involving ... [Note Roman and also signature, as V' is not a final-state particle!]


- L47: ... described by the charged-current operators T0, T1, and T2, as well as the neutral-current operators T8 and T9 [6].

The operators T0-2 modify all quartic vertices, i.e. including the neutral current ones.

- Figure 1, top left: label the virtual line connecting two Z's with a W. Bottom right pane: label three virtual vertical lines as q.

Will update figure.

- L70: define the muon detector barrel here.


- L75: ... the size [or dimensions] of the towers...


- L93: give the MG5_aMC version here (2.2.3 ?).

Added version (v.2.3.3)

- L96: give the version of Phantom here.

Added (v1.2.8)

- L100: ... Born level and merged with jets from the parton shower description using the FxFx scheme [17].

Sentecne modified.

- L105: given MCFM version here (6.0 ?).

Added (v7.0.1)

- L113: give full Pythia version here, 8.2xy.

Added (8.212)

- LL175-177: it would be logical to move this sentence after the one ending on L182, as you apply thresholds on corrected jets.


- L180: add a reference to the 7 TeV JES paper as well and arrange them in the order of appearance.


- LL209-210: ... feature similar kinematics to that for the dominant background and are estimated using simulation.


- L213: Z+X [should be in Roman!] is really an odd choice for these backgrounds, as they include tt¯ background, which doesn't have Z bosons. Please, use "Other" to denote these backgrounds. Also adjust the figure legends accordingly.

The dominant source of reducible background in this analysis is DY+jets, contributing more than 80% because of its large cross section and the on-shell Z. This analysis uses the same nomenclature as used in earlier multilepton publications, notably the HZZ and ZZ cross section studies. We made the dominance of the DY+jets contribution clearer in the text.

- LL220-223: a bit more detailed explanation would be appropriate here, including the definition of the control sample.

We rephrased.

- L237: give a reference to LUM-16-001 when quoting the 2.5\% number.

Reference was added.

- L272 and Fig. 3: BDT spectrum and BDT score are both jargon; please use "BDT output" instead.


- L278: add a reference to Junk and Read papers as part of standard CLs references.


- Figure 4: move the y-axis label a bit to the left; decide whether you use T or T8 as subscripts in the legend.


- L326: finish the paper with a bang: "... at the LHC and the most stringent limits on the T0, T1, T2, T8, and T9 anomalous quartic gauge couplings to date."




You use italic j to denote jets, while the logical choice would be to use Roman j, which is also the font used in the figures. Suggest switching to Roman everywhere, in particular in multiple occurrences of the ZZjj construct [via search-and-replace], ℓℓℓ′ℓ′jj [LL41,46], etc. See also detailed style comments.


LL3-4: the CMS detector and corresponding;


L8: add blank spaces in front of two opening parentheses in (stat) and (syst);



L2: The massive vector boson scattering (VBS) [the acronym is introduced more naturally this way];

We prefer the current formulation.

L3: electroweak (EW) interaction.


LL5-6: of order 1 TeV [1,2].


L13: allowing for a test of the electroweak;


L14: of differential;


L15: Higgs boson couplings;


L16: the Higgs boson mass.


L19: (aQGCs).


L21: V, W in Roman;


L23: (j) in Roman;


LL24,28: qq→V′V′jj;


L25: the EW interaction.


L33: branching fraction;


L38: allow for the extraction;


LL38-39: contribution to VBS, which is;


L40: four-lepton final state.


L41: The search for the EW production;


L43: add a comma before "which";


The CMS detector:

L67: resistive-pate chambers;

We follow the PubComm guidelines.

L75: the size of the towers;


Signal and background simulation:

L92: The EW production ... is simulated at LO using;


Signal and background simulation:

L92: The EW production; at LO [already introduced on L23];


L96: are cross-checked with those obtained;


L97: agreement on the yields;


L99: add a comma before "with up to";


L101: Z boson decays; between the EW;


L103: add a comma before "and";


L106: {\sc mcfm}; agreement between the two samples is;


L109: add a comma before "are";


L110: at LO using;


L112: couplings probed;


L115: (PDFs); add a comma before "and";


LL118-119: implemented in the;


Event selection:

L131: of 27 and 22 GeV, respectively, help to recover the inefficiency.


L132: events that satisfy;

Sentence dropped.

L138: add a comma before "which";


L151: vertex of less than 4.


L157: half the sum of the transverse momenta of;


L166: Kalman filter approach;


L167: J/ψ meson and Z boson decays. [Note J in Roman.]


L170: (PF) candidates [29,30];


L171: add a comma before "as implemented";


L174: add a comma before the second "and";


LL181-182: and is applied as a multiplicative factor, scaling the;


L192: in the candidate event are required; end the line with a comma;

The selection applies to teh ZZcandidate, fixed.

Background estimation:

L207: the EW signal.


L208: pp→WWZ→jets [Roman];


L212: no comm needed before "or";


L213: Z+jets; WZ+jets [Roman];


LL214-215: suppress this background, which;


L217: add a comma before "except";


L221: of observed events in these;


L222: Z+ℓloose [subscript in Roman];

Notation dropped.

Systematic uncertainties:

L234: uncertainty is 8\%.


L237: The integrated luminosity;


LL237-238: predictions of the estimates based on simulation.


LL240-241: A 40\% yield uncertainty in the reducible background estimate based on control samples in data takes into account;


L242: Z+ℓloose [subscript in Roman];


LL242-243: determine the lepton misidentification rates;


Search for the electroweak production of ZZ with two jets:

L244: Search for EW ZZjj production;


L247: from the EW production.


LL248,250,251: subscripts "jj" in Roman;


Fig. 2 caption, L3: filled histograms~--- the expected signal;

Comment not clear.

Table 1 caption, L2: subscripts "jj" in Roman;


Table 1 body, header row: tt¯Z [Roman]; Z+X -> Other;

Done. Z+X maintained, see prev. argument.

L257: subscripts "jj" and "ZZ" in Roman;


L258: subscripts "Z" in Roman; also ηj1, ηj2;


L259: tagging jet transverse momenta.


L265: using a matrix element based approach;


L267: backgrounds in the search;


L258: subscripts "jj" in Roman;


Fig. 3 caption, L2: subscripts "jj" in Roman; L3: filled histograms~--- the expected;

Done. Comment unclear.

L273: maximum likelihood template fit [compound modifiers made of two nouns do not need to be hyphenated].


L280: add white spaces before the two opening parentheses;


L284: |ηℓ|<2.5.


L288: ℓℓℓ′ℓ′jj in Roman;


Limits on anomalous guartic gauge couplings:

L291: contraint aQGCs in the;


L296: nonzero aQGCs is to;


L297: subscript ZZ in Roman;


L303: aQGC parameters.


L305: subscript ZZ in Roman;


L307: {\sc vbfnlo} framework;


Table 2 caption, L3: are in TeV −4; the unitarity bounds are in TeV.


Fig. 4 caption, LL1,4: subscript ZZ in Roman;



L316: add white spaces in front of two opening parentheses;



L373: start the sentence "Individuals ..." as a new paragraph;



Ref. [20]: add the arXiv reference.


Ref. [35]: add doi and arXiv references; also typeset the journal as JINST {\bf 12} (2016) P02014.


Ref. [37]: is there a doi reference?


Ref. [44]: typeset the erratum as a doi, just like you do in Ref. 40].


That's all. Good luck with the final editorial work and speedy submission!


COMMENT-009 (Emilien Chapon)

Dear authors,

Congratulations for this very interesting and well written paper. I have read it on behalf of the Statistics Committee and I have no comments in that respect, everything seems sound and well described. I only found a couple of typos

l13: allowing a test the electroweak -> allowing a test of the / to test the EWK...


l221: number of observed eventS


l264: I wonder if "hyper-parameter" would deserve a short definition? But I leave it to the authors to decide.

Changed to tunable parameters of the BDT training algorithm.



COMMENT-010 (Livia Soffi)

Dear Authors,

thanks for producing such a nice result and high quality document.

Please find below the comments collected from the Cornell university group.


Livia on behalf of the Cornell Group

Contents - TYPE B

Title & General:

- Would it be nice to add explicitly that this study is performed in the ZZ channel here?

The channel can easily be inferred from the final state given.

- Why is the paper coming as a bunch of low-res images? Impossible to use search.

We cannot reproduce this issue?

- “proton-proton” -> proton--proton with en dash.


- "four-lepton state" or "four-lepton final state"

The Pubcom guidelines prescribe the formulation without a hyphen (idem for similar suggestions).


- Too much jargon is used in this abstract. We suggest to –re-think it significantly and use more precise description. E.g. in events with...

We address specific comments below.

- From our understanding the final state can include also the case where the two Z bosons decay both to either electrons or muons. Therefore there is no need to use the apex when writing pp->ZZjj->lll'l'jj. We would use a more generic "l" for all of them

We use the same notation as the ZZ cross section measurement paper. We will modify "l,l' = e, mu" -> "l,l' = e or mu".

- We won`t use here the "j" to identify the jet since it has been not yet defined and a non expert reader does not know what "j" stands for.

We will define the symbol.

- Please add “in the ZZ channel”

The channel is given in with the cross section symbol.

- It is better not to spell out: "T0, T1, ..." as they are defined in a particular EFT model, this is not a universal knowledge.

There is no ambiguity on the naming/meaning of these operators. We think having the names/list in the abstract provides usefull information to the reader as to the content of the paper. Past CMS results on aQGCs also include the coupling names.

Pdf title:

missing a space "anomalousquartic"


Main Text:

- L21: What about gamma*?

The interest in studying VBS are the weak bosons as they receive the Goldstone boson degree of freedom in EWSB. Processes with photons are not sensitive and past CMS results of gg->WW scattering avoid the term "VBS". The restriction to weak bosons is also common in the theory community, see e.g. a recent review on the topic (1610.08420).

- L24-28: Also in this case we think the apex close to the "V" in the formulas is not needed

We will remove the tick.

- L31: "The fully leptonic..." this is an incorrect statement. Four-lepton decays were probed back in 7-TeV campaign. You must explicitly mention VBS and, better,the four-lepton and two-jet final state.

The Paragraph reviews VBS results, i.e., the presence of tagging jets is implicit. We intend to rephrase.

- L40: You are often mixing the four-lepton and the four-lepton and two-jet final states. Note that there was a dedicated study SMP-16-001, which had an explicit four-lepton final state. The SMP-16-001 study was not focused to VBS though, but this adds quite some confusion.

SMP-16-001 did not investigate VBS nor aQGCs, but aTGCs.

- L43: You either use a multivariate discriminant or you perform a multivariate analysis. You cannot use MVA as A stands for the analysis and not the observable. MVA is an often used jargon for MV discriminants. In some later parts, you have a correct usage.

We rephrased to say "...multivariate discriminant..."

- Fig. 1 clearly misses gamma* diagrams, this would be of a particular importance to the interference and probably (?) more influential than the Higgs exchange.

The diagrams are intended to illustrate the importent processes for the process of longitudinal VBS, which is the process relevant for the study of EWSB. Photons do not receive a longitudinal polarization from the Goldstone boson. The Higgs is crucial to regularize the VBS cross section at the TeV scale.

L55: The symbol "eta" has to be defined before using it


L57-60: It is not clear to us why these detector performances are given for these specific pt ranges. Are they relevant for the analysis?

This paragraph is from the PubComm detector description guideline.

L97: "excellent agreement" does not give any good estimate. Could be possible to quote a range of agreement in terms of % for example.

The agreement is better than 1% once the different Z widths used by the generators are accounted for. The details are in Sec. of AN2017_002. We'd prefer to not go into more detail in the text as this is a rather technical comparison.

L99-102: Check how "MG5_AMC" is written

We follow the PubGuidelines.

L101: If you produce Zs and then let it decay via MadSpin, you force your Zs to be on shell, thus the interference will not be accounted correctly when mll = mZ. How did you get 1%? It may be an underestimation if what you describe (Zs being on shell) is true. Also, gamma* will be at least of this order.

The interference numbers are obtained from dedicated LO samples that do not use MadSpin. Details on the interference study are in on AN2017-17-002 starting L151 ( inv12). Also, the interference mentioned in L101 is between the EW and QCD diagrams.

- L125: What about a trigger table here? It is often difficult to comprehend long descriptions.

We prefer the verbal summary as it summarizes the most relevant trigger paths.

- L126: do we need this half-sentence: "of the same or of different flavors"

Half-sentence dropped because it is redundant.

- L128: corresponding to the leptons-> associated with the leptons?


- L134: "used in the inclusive ZZ analysis" assumes that your reader knows what "the inclusive ZZ analysis" is. What is it? A cross section measurement?

Will rephrase to "used in the inclusive ZZ cross section measurement".

- L135: what is "the luminous region"?

The sentence was replaced by the PubComm recommendation.

- L136: "vertex with the highest sum of the p_t^2 of its associated tracks" that expression has to be fixed, e.g. "vertex which has the highest sum of transverse track momenta, as calculated using the tracks associated to it" or similar.

The sentence was replaced by the PubComm recommendation.

- L135: How the luminous region is defined? Maybe non-CMS reader do not know that

The sentence was replaced by the PubComm recommendation.

- L138: "discriminant which includes" -> a discriminant which is calculated using the following variables: also: move the reference up, e.g. behind "discriminant"?

The reference position at the end reflects the fact that the observables that enter the electron ID are just as important as the discriminant, the cited electron paper covers the entire scope of reconstruction and identification which is shortened in this paper for brevities sake.

- L144-145: We found the "outside-in" and "inside-out" expressions quite colloquial. We would rephrase this sentence.

It is somewhat colloquial, but a vivid description of the procedure and both terms are defined/explained in the sentence.

- L150: We don't think you can assume that people know what "3D impact parameter significance" is. Need to give the expression.

We will rephrase to make it clearer.

- L150: Do we need to spell "3D"?

Will rephrase.

- L151: "less than 4".. are we missing the units?

No, because the significance is the ratio of the observable with its uncertainty, which cancels any units.

- Equation1: Why in this analysis we do not use the standard electron ID?

Eq. 1 refers to the particle flow isolation, not the ID. The only difference w.r.t. to the standard PF isolation is the veto of FSR photons. Multilepton analyses like the one presented here use a FSR recovery algorithm, and the identified FSR photons are removed from the isolation calculation, increasing the event selection efficiency.

- L164: What do you mean for "pseudorandom" smearing?

The specifier was dropped beacuse "smearing" suffices. (Pseudorandom in the sense that its a computer-generated random number.)

- L183: Here for the first time in the paper you introduce the "*" symbol. We would explain what do you mean for that.

We dropped the star because the "Z candidate" refers to a reconstruted entity.

- L185: Why this pt threshold is lower than the one in the trigger in L126?

Please see response to earlier comment on this point.

- L187: Why the DR separation between electron and muon is larger than the one between l1 and l2?

The electron-muon DR separation requirement suppresses rare cases where a muon FSR photon is reconstructed as an electron, notable after undergoing conversion in the tracker. The lepton-lepton DR separation is a technical cut to remove rare cases where the same lepton gets reconstructed twice.

- L190: It looks like that you inherit this description from 4l paper as you have mZ > 40 GeV while later (ln 197) you say 60 > mZi > 120 GeV. Please, clean this up.

This analysis uses the same 4l selection algorithm as the HZZ and ZZ analyses, for maximal eventselection efficiency.

- L196: write out 4-> four

We will rephrase to use the word "ZZ candidate" to be concistent.

- Figure3: Did you consider to remove all the cuts from the CR plot on the left and simply write "CR". Then explain the cuts in the caption?

This notation was requested during the analysis revirew. We prefer having these rather simple cuts in the figure, as it might be used out of context. With the cuts in the figure, it can be understood by itself without much further explanation.

- L200: "notably": does that mean 2 jets are required in addition to the "ZZ selection"? We would say that.

Yes, the ZZjj selection extends the ZZ selection by the two tagging jets.

- L202: By this point there was no mention of gamma* and it is clear that you are not even trying to tell it apart. Technically, it is better to call the selection as VVjj, where V = Z, Z*, gamma*. Can you tell what is the gamma* fraction for Z2?

The name of the selection "ZZjj" reflects the topology that it selects. No statement about purity is made. Finally, the contributions from the photon continuum under the Z peak is negligible.

- L224: you have to say what that reference analysis is

We think giving the references is sufficient.

- L227: what is the "classifier"?? What is "the statistical analysis"? You have to describe these.

Will replace classivier with "discriminant" that is previously defined. Will add "..., are used as systematic uncertainties in the statistical analysis.".

- L227: "MVA output distribution". In some parts of the text you have correctly used MV discriminant, but here we have a jargon again. Is there a reason for a cut&count coupling study? You can definitely add more sensitivity in the future if you consider smarter (and more) observables.

Will make nomenclature coherent. What is meant by "cut&count coupling study"? There are no cuts besides the very loose ZZjj selection in this analysis. We systematically studied the multivariate discriminant to ensure its optimal performance, as documented in the AN.

- L236: 6/4/2% etc is a confusing notation. Looks like a division.

Its is a compact notation and the sentence is unambiguous.

- L233: etc: you mention "the energy scale uncertainty", "the luminosity uncertainty" as if everyone knew that these exist and what they are supposed to be. You have to cite some reference and/or explain.

We added the references which provide the details on how these uncertainties are evaluated.

- L240: this sentence is 4 lines long. Please break it up into shorter sentences- it's very hard to follow.

We will rephrase.

- Table 1: where do the uncertainties come from?

We will add a sentence explainign that they are the quadratic sum of the statistical and systematic uncertainties.

- Section5: We would write in this section more precisely the individual fractions of each background component.

Table 1 provides this numeric information and FIG1 and 2 show the distributions. This analysis is based on a shape measurement with very loose selections and providing numbers in Section 5 without further context could mislead the reader.

- Section7: The title of this section is too much similar to the title of the paper. Could we simply change it to "Analysis Results" or something similar?

This paper includes two sets of results (EW production and aQGC limits). We will modify to "Search for the EW ZZjj production".

- L257 and following: can this be turned into a bullet list?

We prefer the prose form and will suggest a more concise formulation.

- L267: what does "to the search" mean here?


- L306: We would specify what "the other couplings" are

We rephrase: ".. by setting all other anomalous couplings to zero."

Text - TYPE A

- Abstract: investigation -> search

We prefer investigation.

- L8: "this boson"->"the Higgs boson"

The point of these VBS measurements is to check that the Higgs is indeed the Higgs from the minimal scalar sector. Saying that the boson is the Higgs thus makes this measurement redundant.

- L12: remove "thus"


- L13: "test the"->"test of the"


- L13: "allowing a" -> "thus allows to"


- L16: "resonance mass"->"Higgs boson mass"

Sentence rephrased.

- L18: similar to ln 13.


- L30: "and CMS" -> "and the CMS"

The current formulaiton appears to be the more frequently used one.

- L37: "a precise"


- L42: "using" -> "with"

Stay with current formulation.

- L63: Do we need a space in "nonshowering"?

This is a paragraph from the PubCom detector TWiki.

- L115: remove "as the default"


- L120: "for data" -> "for the data"


- L156: Remove "jet area"

Quotation marks removed.

- L147: "on the track" -> "to the track"

Consult with LE.

- L206: "are taken"-> is taken

This sentence should stress that both are from the simulation, thus stick with 'are'.

- L215: We would skip the repeated word "reducible".


- L216: "from the data"


- L218 subscript Z_2 appears out of the blue.


- L267: "to the search" -> "in the search"


- L275: "from the data"


Review comments

Bill Gary, 16 May on v3 of the paper

Title: add space between "anomalous" and "quartic"

Abstract: o add serial comma in 10th line: "...T8, and..."

o last sentence: "...first investigation OF..."

Text: 3 "non-Abelian"

6 "...the CERN LHC..." [mention CERN on first mention of "LHC"]

18 "...allowing tests for..."

29, Don't use jargon "reducible" and "irreducible" background without defining what they mean. Please add explanations.
We added a definition.

36 "channel" [singular]

41 Here you should now state merely "LHC", not "CERN LHC"

41 "data set" [two words; see style manual]

45 "EW production" [no hyphen]

Fig. 1 caption: o "...diagrams FOR the electroweak-..."

o I assume you want a hyphen following "electroweak-" in line 1 so that the caption reads (effectively) "electroweak-induced".

o I would write " the bottom left DIAGRAM." [not "panel"]
We propose "The scattering of massive gauge bosons as depicted in the top row is unitarized by the interference with diagrams that feature the Higgs boson (bottom left).". This avoids having a sentence with three "diagrams". 82 " select events of interest..." [it is too subjective to state "most interesting"]

105 Are you using US or British conventions ? "center" in line 5 is US, while "modelling" with two "l"s is British. The US spelling is "modeling"
Changed to AE throughout the document. 110 lower case "matrix element"
Fixed. Did you mean "reweighting" and not "reweighing" ? This sounds like jargon "matrix element reweighting functionality". What is that ?
Changed to " … employing matrix element reweighting to obtain …". 110 no hyphen following an adverb ending in "ly": "...finely spaced..."
Fixed. 126 "subleading" with no hyphen but it is better not to introduce jargon that is virtually not used: "...and the next-to-highest \pt lepton pTl>12 (8)\GeV."
Fixed. 129 "\pt thresholds" [no hyphen]
Fixed. 130 "passes" -> "satisfies"
Fixed. 130 "pass" -> "satisfy"
Fixed. 152 the cone is defined by "=0.3", not by "<0.3"
Fixed. 152 You alreadydefine \phi on line 72, so you should not be defining it again here.
Fixed. 157 "...factor OF one-half..."
Fixed. 161 don't use unexplained jargon "tag-and-probe". Merely state "measured in data in bins of..."
Fixed. 180 don't introduce an acronym used only once. Eliminate "(JES)" here and write out "jet energy scale" on line 234.
Fixed. 199 " performed on a subset..."
Fixed. 200 add comma following "selection"
Fixed. 200 "The two highest \pt jets are referred..."
Fixed. 205 "panel" -> "diagram"
Fixed. 219 add comma before "respectively"
Fixed. 223 reads better as "...and IS described..."
Fixed. 226 "...classifier. The resulting..."
Fixed. 228 "...envelope ... IS ..." [not "are"]
Fixed. 233, 235 "uncertainty IN" [not "on"]
Fixed. 237 "...IN the trigger efficiency is evaluated BY..."
Fixed. 238 "efficiencies" [plural]
Fixed. 251 "...production, i.e., the ratio..."
Fixed. 278 "...measured TO BE..."
Fixed. 281 How about "The measured signal strength is used to determine..." You haven't used the signal strength for any purpose to this point (merely determined what it is), so "further [or "furthermore"] used" does not seem appropriate.
Fixed. 286 remove "just"
Fixed. 286 "reconstruction-level"
Fixed. 290 eliminate distracting "furthermore"
Fixed. 293 "involve" [no "s"]
Fixed. 301 "Gaussian" [upper case "G"]
Fixed. 310 "A search was performed for..."
Fixed. Ref. [44]: take "C" out of boldface
Fixed. ------++ Bill Gary, 12 May on v0 of the paper Eliminate "(EW)" from abstract
Fixed. 12 "...allowing a test of the electroweak..." [use English grammar, not French grammar :)]
Fixed. 29 "Collaborations" [upper case C]
Fixed. 30 You need to define what you mean by "Run 1". Since you never use that term again it would likely be better to rewrite the sentence without using the term.
Fixed. 169 "particle-flow" takes a hyphen
Fixed. 169 "anti-\kt" [use the CMS macro "\kt", which will give a CMS-style upper case Roman "T"
Fixed. Sometime (line 103, 215, etc.), "jj" in "ZZjj" is in italics, and other Times (104, 243, etc.) not. They should be consistent.
Fixed. 314 write out all acronyms again, for the summary "standard model"
Fixed. [7] take "D" out of boldface; similar for [8,37,40]
Fixed. Figs. 2-4. The axis labels and numbers should be bigger. Remove "Preliminary". In Fig. 2 (right), Fig. 3, and probably Fig. 4, "CMS" should go inside the plot frame. The "CMS" and "39.5 fb-1 (13 TeV)" should be bigger.
CMS is in frame, removed 'Preliminary', increased font sizes. ---++ Darien Wood, 12 May on v0 of the paper Dear SMP-17-006 authors, Here are my comments on the paper draft 0. I think it is in good shape and I do not object to assigning a language editor. Below are some specific comments, mostly on wording. Best regards, Darien Abstract: No need to define "EW", since it is now used again in the abstract.
Fixed. The definition was there because the label "EW" is used in the phrase "..the electroweak production is measured to be $\sigma_{EW}=$...". We removed it because it should be clear what the "EW" subscript stands for. line 7, suggest "this boson" -> "contributions from the exchange of this boson"
Changed. line 16: no need to define the "EFT" acronym since it is not used again
Fixed. line 24: "only involves the" -> "involves only the"
Fixed. line 32: "laboratory" -> "favorable laboratory" (I think any channel can be said to provide a laboratory)
Suggestion adopted. line 36: "allow to extract" -> "allow the extraction of"
Suggestion adopted. Figure 1: Do you plan to add particle labels to the Feynmann diagrams? It would be good to distinguish the quarks from the leptons, to show which V's must be Z's, and to label the Higgs.
Fixed. line 101: "less than 1% of the total yield" Don't we really want to know how it compares to the EW yield, rather than the total yield? Later the paper says that the EW purity of the sample is around 6%, so the interference contributions could be around 1/6 of this, if I understand correctly.
We think the comparison to the total yield is more relevant, as the interference is concentrated at low BDT scores as shown in Fig. 3 of the AN. We propose to modify the sentence starting L99 to include this information on the kinematics: "The interference between the electroweak and QCD diagrams is evaluated using dedicated samples generated with MG5_AMC at LO and is found to contribute less than 1% to the total yield with background-like kinematics and is therefore neglected. " line 114-115: either go with plural ("generator" -> "generators") or singular ("Each MC sample is weighted ...")
We think the current formulation is correct, as it emphasizes the fact that all MC sample are normalized to the cross section returned by the particular genertor used to produce it. line 211 and legends in Figures 2, 3, 4: "Z+X" is an odd name for a category that included ttbar, since ttbar events have no Z's. What about "ll+X"?
The "Z+X" label defined in L211 and is also the label used in HZZ and the ZZ inclusive analysis. While it is true that $t\bar{t}$ is a contribution to the reducible background, it is only a minor one, contributing less than $10\%$ as illustrated in Fig. 27 and Fig. 28 of AN-2017-002. line 236: "efficiency are" -> "efficiency is"
Suggestion adopted. line 239: "limited statistics" -> "limited numbers of events"
Suggestion adopted. line 281: It sounds strange to compare a "volume" to a "level". What about "The fiducial volume definition is almost identical to the selections imposed at the reconstruction level"?
Fixed. line 296-297: "allowing to interpolate" -> "allowing interpolation"
Suggestion adopted. Figure 4: Does the last bin include overflows? If so, this should be said in the caption.
Its does, we added "The last bin includes all contributions with $m_{ZZ}&amp;gt;1.4~TeV$." line 309: "We presented results for a search for ..." -> "A search was made for..." (The summary generally should be in passive voice, and should say what was done and not what was presented.)
Fixed, thank you for pointing this out. ---++ Isabel Josa Mutuberria, 11 May on v0 of the paper Dear Philipp, all, I went through the paper draft, it reads very smoothly and the final product is really nice. Let me send you a couple of minor comments (see below). Let´s wait until tomorrow to see if other members from the ARC have any additional comments. Best, Isabel
Thank you L 19-23 paragraph: L 19 -20 up to ... which then interact. OK. This is a general statement about VBS processes. But L 22-23 (...VBS is characterized by the presence of two forward jets (j) in addition to the gauge bosons (qq --> VVjj) and little hadronic activity in the central region [7, 8].) We can also have the VBS process qq -->Vjj , with only one Vector Boson in the final state. The sentence only holds for the specific final state of this paper. Also, as it is written, it may be understood that the two Vector Bosons that interact (L 20) are the same to be observed in the final state (L 23). Can you rephrase slightly this sentences ?
We added 'outgoing' and a prime to the final state gauge bosons to clarify that the incoming bosons don’t need to be the incoming ones. We think the statement of having two bosons in the final state is true, because the sentence explicitly refers to scattering, i.e. a 2→2 process, excluding the weak boson fusion process. L 36-37 Furthermore, the spin correlations of the reconstructed fermions allow to extract the longitudinal contribution which is the main interest in the investigation of EWSB. Sorry for my ignorance, but, can we (do we) extract the longitudinal contribution in this paper ? or is it a general (true) statement about VBS in 4l final states ?
We do not explicitly extract the longitudinal contribution yet. This sentence is intended to show the motivation of studying VBS in this channel and what will be done with more data. L 232-233 The uncertainty on the QCD background normalization and the JES uncertainty are the dominant systematic uncertainties in the search. I think this is a new sentence wrt PAS. I checked the AN, Fig. 50 (b) post-fit full distributions. I can see that the largest pulls come from the PDF and renorm. EWK signal, then normalization of ggZZ, jet energy scale. Pulls from qqZZ (renorm and PDF) are much smaller. What do you mean by ``The uncertainty on the QCD background normalization’ ‘ ? To which of the items in the pulls does it refer to ?
We think the pre-fit pulls should be compared, as they do not depend on the randomness of the data. From Fig. 50(a) one can see that the QCD ggZZ normalisation is the leading systematic, followed by the JES uncertainty. L 245 Fig. 2 shows the dijet invariant mass mjj as well as the absolute dijet pseudorapidity separation … Change the orderingin which the two plots are presented mjj <--> delta etajj to follow the same order of the plots themselves. /Users/pigard/CMS/zzjets_analysis_note/utils/trunk/tmp/SMP-17-006_temp.pdf L 276 nuissance. Only one s?
Fixed. References.- 6 and 11 seem to be the same one.
Fixed. 36 seems to be somehow corrupted. LaTex stuff: Please, check the names of the particles (electron in particular), that should be in roman (or use \Pe) in many places.
Fixed. Also the notation for ZZjj is not uniform, sometimes jj is in math mode, others in roman font.
Fixed. -- PhilippPigard - 2017-05-12

Edit | Attach | Watch | Print version | History: r23 < r22 < r21 < r20 < r19 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r23 - 2017-07-25 - unknown
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Sandbox All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2023 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback