Review Twiki for SMP-17-004

The twiki is used to prepare the final reading for SMP-17-004: "Observation of electroweak production of same-sign W boson pairs in the two jet and two same-sign lepton final state in proton-proton collisions at 13 TeV"

Color code

  • BLUE answered
  • PINK to be discussed
  • RED did not agree

Table of content

Helen Heath ( Date of comment 05 Sep 2017

Thank you for your comments, Helen.

Hello I only have a few minor comments on this paper. Best wishes Helen

L1 Is the SM "exceptionally" accurate? Exceptionally compared to what?

we removed "exceptionally"

L23 "reported observed" should be one word or the other not both.

we removed "reported"

L55 "amount" -> "number"


L66 "muons system"-> "muon system"


L71 "dominating" -> "dominant"


L82 "They are"-> "These techniques are"


L88 remove comma after "third"


L112 insert "the" before "finite"

added "the"

Claude Charlot ( Date of comment 05 Sep 2017

Thank you for your comments, Claude.

Dear Colleagues,

Congratulations for this very nice result and very clear and well written letter. I have rather few comments that I append below.

Cheers, Claude

Type B:

- Fig.1: as written the caption as written is a bit confusing about the QCD-induced production (it looks like the single, triple and quartic gauge coupling diagrams also refers to QCD-induced production, which obviously should not). I suggest to remove the symmetry and say something like '.. of the EW-induced W boson pair production (left, middle-left, miffle-right) and of the QCD-induced background (right).'.


- L26: I would repeat the channel used here -> '.. a study of VBS in the same-sign WW final state in proton-proton ..'

adopted, wrote "same-sign $\PW$ boson pair final state".

- L39: it is not clear wether LO refers here to QCD or EW or both, I suggest to write it explicitly (since after you discuss the effect of complete NLO corrections).

changed the order in the sentence to make this clear.

- L99-100: the fact that the charge misidentification for muons is negligible was already mentionned at L90-91, I would just remove this part of the sentence, keeping the numbers for the electron case -> '.. and about 0.35 in the endcaps region.' It would also be relevant to mention that you use the tight electron charge criteria, since it is not standard, perhaps after the sentence ending at L66.

Removed the muon statement. Did not include a statement on the charge id for electrons. Seems too much detail for this letter.

- L142-143: add whether leptons are dressed or not for the fiducial definition.

we are using bare leptons and added this to the text "In this definition, the leptons are defined at particle level post final state radiation and ..."

- L165-166 and Table 2: CMS has now much tighter limits on fT0, fT1 and fT2 from the ZZjj VBS recently published analysis. These results should be reported instead of the Run 1 results.

Included the ZZjj analysis and added reference

- L183-185: it is not clear why the H+ analysis is mentionned here and the result of the H++ analysis compared to it. The model is the same but what do we learn here comparing results of two different searches?

we learn that there is multiple signatures sensitive to these models. The direct comparison was a request from the CWR.

Type A:

- front page: provide meaningful PDFKeywords

pdfkeywords={CMS, physics, vector boson scattering}}

- L16: 'An excess of events with respect to SM expectation..'


- L27: 'at the CERN LHC' (see pub guidelines)


- L33: you do not need to put '\eta' between parentheses (see pub guidelines)


- L71: '.. are nonprompt leptons' -> '.. arise from nomprompt leptons'


- L72: I would drop 'candidates' here -> 'Two same-sign leptons, electrons or muons, ..'. (otherwise candidates should be used everywhere, eg at L74, 80, .., that becomes rather heavy)


- L74: 'from other charged or neutral particles'

"and" is correct here

- L82: 'to veto top quark events.'


- L119: 'simultaneously' would better be moved after 'varying' (as it is it looks it is simultaneously up and down..)


- L124: 'An uncertainty from the parton distribution functions in the signal efficiency ..'


- L131: '.. with a fit to the (m_jj, m_ll) two-dimensional distributions.'


- L139: 'The ratio of measured event yield ..' (singular)


- Table 1 caption: 'after the selection.'


- L173-174: the specification of the model used would be better given before, at least before the number of parameters of the model is used and perhaps even before the sentence starting at L169.


- L197-215: there is a formatting problem with the acknowledgement.

I do not know how to fixed this.

Darien Wood ( Date of comment 05 Sep 2017

Congratulations on this important result. It should be submitted promptly. The paper is well written and I have just a few comments.

Best regards,


Type B:

Figure 1: It is not incorrect, but the caption does not parse easily to map to the four diagrams. This could be fixed by reordering the figures (QCD first, then EW 1, EW3, EW4), with a caption like: "Representative Feynman diagrams for (left to right) QCD-induced W boson pair production and EW-induced W boson pair production for single, triple, and quartic gauge couplings."

we modified the caption

Lines 96-98: "The background contribution from charge misidentification is estimated by applying a data-to-simulation efficiency correction to charge-misidentified electrons in bins of eta estimated using Drell–Yan events." This sentence can be confusing. Do you mean to say that the charge misidentification rate is measured from DY events? If this is the case, then maybe that information is better in the following sentence. E.g. "...electrons in bins of eta. "The charge misidentification rate, estimated using Drell-Yan events, is between ... for electrons, while negligible for muons."


Lines 121-124: You take the uncertainty from interference as a constant fraction (4.5%) of the signal. To estimate even the relative size of the interference effect, don't you have to assume a signal cross section? Naively, I would expect the interference to scale as the square root of the signal cross section, and not in direct proportion.

The fraction depends on mll and mjj. There is a dependency on the couplings we are estimating, but it's a second order effect. After all, we are seeing a very small fraction. But maybe more importantly, we are talking about the effect after the selection, which makes the dependence unknown a priori.

Lines 147-151: For the theoretical prediction of the fiducial cross section, why do you give the LO and not the NLO, when you seem to say you have computed the NLO corrections? Can I just reduce the quoted LO prediction by 13% to get the NLO prediction?

This is true to first order. There are second order effects from the acceptance. The MC is at leading order and we compare the cross section to that value. The reader can compare the results with their calculation. It's somehow a consistent approach. The NLO calculation is a response to our PAS and we added this last minute. We can discuss whether we want to change the sentence.

Table 2: Shouldn't the first column be f_xx/\Lambda^4? This is the form of the coefficients of the dimension-8 terms given in [6], and I thought the f's were supposed to be dimensionless.

correct, fixed.

Type A:

Line 6: "enhancements for VBS" -> "enhancements in VBS" (?)


Line 59: "single and double lepton triggers" -> "single- and double-lepton triggers"


Philippe Bloch ( Date of comment 04 Sep 2017

Thank you for your comments, Philippe.

I have very few comments for this short and well written letter. Comments are listed below

Best regards


Fig.1: I think one could mention which diagram is which (EW, QCD, quartic) in the caption, after labelling the diagrams. Should one not add H and gammas with the Z exchange, to be consistent with the text L46?

We adjusted the caption.

L57: probably obvious, but the symbol “j” for a jet is never introduced. May be put (j) line 10?

I don't think that is necessary.

L106: I think a sentence is needed at the beginning of this paragraph, to explain that all the following percentages refer to the systematic uncertainties on the predicted yields.

A general statement would be incorrect. We should briefly discuss if such a statement is needed.

L131-132. one should explain better what the two parameters fit brings, compared to a simple counting experiment (i.e. the excess). May be a remark that the mjj distributions are different for the signal and the WZ background?

we explain the the purpose in line 133 and 134. We had a number of iteration with the ARC on this point and would appreciate a proposal.

Finally a question: same sign W events have been used in the past to identify double partons interactions. There is no mention of this background. It is probably negligible due to the jj selection, but I believe a sentence on this would be useful.

the background from double partons interactions has been evaluate and is negligable in this measurement. That's not a surprise because of the VBS requirement. Do you really want to add a sentence? The caption for the mll and mjj distribution lists the background.

Type A

L17. these particles . Why the plural (it refers to “a new resonance”)

Singular or plural is correct. Changed the sentence to "Doubly charged Higgs bosons are "

L39. missing blank after QCD-

I did not find this.

Sarah Eno ( Date of comment 31 Aug 2017

Thank you for your comments, Sarah.

Congratulations on a very interesting result, and a well-written paper. My comments for the FR are given below.


• Would the left-most figure look better of the neutrino and lepton lines were of more similar length? Maybe you could radiate the W a bit sooner? For the second plot, wouldn’t it look nicer if the W’s were radiated soon as well? In general, these diagrams could do with some adjustment of spacing.

I like the standarized way of plotting these diagrams.

• Caption, figure 1. Shouldn’t this be “production of EW- and QCD-induced same-sign W boson pairs”

adjusted the caption.

• According to cms pub comm style, the y axis label for figure 2 should be “Events / bin” ( )

added /bin

• Figure 3. Cms pub com prefers you not to use the “x” sign in the y axis label. “. Do not use either the \times or the \cdot symbols in formulas where multiplication is the default operation, e.g. write “\sigma \mathcal{B}" for the product of cross section and branching fraction, and not “\sigma \times \mathcal{B}" or “\sigma \cdot \mathcal{B}".

adjusted the figure

• Figure 2. Do you want to include curves with the aQGCs on the m_ll plot?

we went through all possible permutations of this question in the review.

Tables • Table 1, 6th row, last number. Do you really want 3 digits on the uncertainty? Ditto for first row?

I would leave this for esthetic reasons. No strong opinion.

• Table 1, second column. The sum of the listed backgrounds is larger than the total background


More important comments

• Abstract. To me, the first sentence says that it is the first observation at 13 TeV, not the first at any energy. Maybe ,’s after “pairs” and “TeV” would help?

Let's discuss in the meeting.

• Paragraph lines 37-56. Do you want to mention the normalization of the samples here? Also, which are actually used for background yields and which are just used for closure tests, etc for the data-based backgrounds?

You are right. A description of the normalization was removed in the review process. I think we should add this between at 105 or in 127. We will make a proposal by the time of the reading. Our proposal is to add to v6 l50 "The simulated samples of background processes are normalized to the best theoretical prediction." and l104 "All other background processes are estimated from simulation applying corrections to account for small differences between data and simulation, as described below."

• Line 67. So, you don’t use type-I met?

We do use type-I MET.

• Line 98. Should be “a data-to simulation efficiency correction,in bins of \eta, which was estimated using Drell-Yan events containing charge-misidentified electrons? (or are you applying it to charge-misidentified electrons in the control sample? If so, how do you determine which one was charged-mididentified? The Z electrons are required to be opposite-sign? The W can be either charge… not sure what you are doing here” Also, is this “control region” used to get the charge-mis-identification background? Or the background where one of the Z leptons is missing? Or both?

Please have a look at page 19.

• Line 110. Is it obvious why the top veto for DY applies to VBS same-sign W’s?

we are estimating how well we model the veto efficiency for light quark jets.

• Line 126. Are the simulated background yields used? Or are the data-based ones discussed in the paragraphs lines 94-105? Do you normalize the simulated yields to the background predicted by data-based methods?

this is the same question as asked before. We will make a proposal on how to fix this before the FR.

• Line 140 is this number both for the signal and for the background from WZ? Or are there different resulting normalizations?

added "signal" to the sentence.

• Lines 141-144. Are the leptons dressed?

we are using bare leptons and adjusted the text.

• Line 152. Does this apply just to the NLO corrections? Or is for the LO estimate to the signal efficiency?

This is only for LO.

• Lines 157-166. Was the MC used here also madgraph?


• Conclusions do you want to give a range of values for the improvements on the aQCD’s? looks like a factor of 4-5?

Yes, I think this would make for a strong ending of the paper. Let's discuss the wording in the meeting.

Grammar comments

• Line 8. I find “pairs events” a bit awkward. Do you mean “to identify events containing two W bosons of the same sign”?

removed "events"

• Line 43 “differs from the signal on a kinematic basis” -> “and has different distributions for its kinematic variables than the signal” or “can be differentiated from the signal on a kinematic basis”

modified following Dick's proposal.

• Line 58 a W boson decay or from a W boson with


• Line 85. Because of the , after “technique”, this sentence to me looks like the presence of the muon is required. Maybe add “possible” before “presence”?


• Line 124. Wouldn’t this be better as “An uncertainty on the sign efficiency times acceptance due to uncertainties in the parton distribution function…”?

changed this to "An uncertainty from the parton ..."

Dick Loveless ( Date of comment 28 Aug 2017

Thank you for your comments, Dick.


This is a very nice result and a well-written paper. It was a pleasure to read it.

Type A

line 1: Delete "exceptionally", which seems like overkill.


line 5: Change "through" to "using".


line 42: Add a comma after "analysis", and change "and" to "whereas".


line 43: Instead of saying "differs from the signal on a kinematic basis" you could say something like "can be kinematically separated from the signal.".

adopted this proposal

lines 53,54: "Simultaneous" is not really correct. How about "Proton-proton interactions occurring in the same beam crossing bin as the event of interest (pileup) are included in the simulation samples.".

Adopted this proposal.

lines 55,56: Then you can write "The pileup has a mean of approximately 27 and corresponds to the conditions observed in the 13 TeV data collected in 2016.".

Adopted the change but wrote "The simulated pileup" because to make it clear what corresponds to what.

line 60: You could reference the CMS trigger paper here: JINST 12 (2017) P01020.

added the reference.

line 66: Change "muons" to "muon".


line 71: This sentence on backgrounds is out of place in the paragraph defining the signal. It would be a good topic sentence for the next paragraph on backgrounds. lines 82,93: Move line 71 here and move the first sentence on b quark backgrounds below nonprompt and WZ backgrounds.

moved the sentence to line 127 to summarize the results in Table 1. We can rediscuss.

line 88: Add a hyphen to "hadronically-decaying".


line 91: Change "nonnegligible" to "not negligible".


line 94: Delete "fully". What is a partially identified lepton?


line 97: Delete the hyphen in "charge misidentified". In the preceding and following lines this phrase does not have a hyphen.


lines 99,100: Delete "while negligible for muons". This is already said in line 90. If you want to say it again, change "while" to "whereas it is".


line 120: Does "up to" refer to a range? Probably not. Better to change "up to" to "a factor of".

this is indeed a range.

lines 124,125: I think what you are saying is: An uncertainty in the parton distribution function contributes 5% to the signal times acceptance.".


line 137: Change "excess in data" to "data excess".


Table 1 caption line 4: Add a comma after "listed".


lines 162,163: Write " . . .nine coefficients, shown in Table 2, are obtained . . . ".


line 272: The "pp" should be in Roman font.


lines 303,304: A date "2017" should be given.


Type B

It would be nice to use the acceptance from the generated events to report a total cross section with an associated error. If this is not possible, it would still be better to report the measured fiducial cross section and errors according to lines 147,148.

following Franciscos proposal for now. We will discuss during the reading.


line 43: How small is the QCD background?

after event selection, the QCD background is about 7% of the signal, less than 4% of the total background.

-- MarkusKlute - 2017-09-04

Edit | Attach | Watch | Print version | History: r7 < r6 < r5 < r4 < r3 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r7 - 2017-09-07 - MarkusKlute
    • 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 &© 2008-2020 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