Comments Joscha Knolle


Abstract, L11: Is it clear from the abstract that the BDT analysis gives the more precise results than the cut&count analysis, without explicitly mentioning this in the abstract?

L19-20: Are these statements about the top and W boson decays exact?

L192: Is "1.5%" the signal acceptance w.r.t. fully inclusive tttt?

L216: What about other ttH multilepton final states like H->TauTau and H->ZZ?

L253-254: "The correction factor [...] is negligible" -> is it clear that this means 1?

L293-295: I understand this is what you technically do in combine, to assign a prior uncertainty to the ttW and ttZ normalizations, that then gets constrained by the fit. But does this 40% have any physical meaning? Do different choices produce different fit results? If yes, please justify this number. If no, it should be dropped from the paper.

L409-410: What does "are in qualitative agreement with those predicted by Ref. [9]" mean? What is the prediction of [9]? How can the numbers before agree "qualitatively" with something?

Type B

L8-12: First you write that new particles can contribute to the cross section only if their mass is twice the top quark mass, and in the next sentence you write that new particles can also contribute when their mass is smaller. This is unnecessarily complicated. I understand that you want to introduce here the separate interpretation items, so you should highlight what is the conceptual difference between on-shell and off-shell contributions w.r.t. your analysis.

L15-18: Do we really need to cite 12 papers on signatures that we are not sensitive to? I propose to shorten this part.

L34-36: The main improvement over the previous measurement comes from the increased luminosity, no? So that should be the first item in this list.

L64: Please also explain alpha. (also L419)

L149: You should quantify the increased pileup.

L181-183: Please add a statement on the trigger efficiency for three-lepton events.

L186-192: You have three DY vetoes here: a Z veto, a photon conversion veto, and a photon conversion + charge misID veto. Please highlight that also in the description.

L206-207: It might be worth adding a motivation for top tagging like "to identify hadronically decaying top quarks", and even to refer to e.g. TOP-17-019 as an example where this is done.

L222: Since also the systematic uncertainties are assigned jointly to the three groups of other backgrounds, I would propose do just drop "for the purposes of plotting and extracting results in the analysis".

L264-265: The integrated luminosity uncertainty in 2017 is 2.3%, not 2.1%. Also, please cite all three relevant luminosity measurement PASes: LUM-17-001, LUM-17-004, LUM-18-002.

L344-345: Please cite also the updated ttZ measurement, TOP-18-009.

L364: "The results of this paper can be used to constrain SM parameters [...]". First of all, you present constraints, so replace "can be used" by "are used" (also later). Furthermore, I propose to make a very clear distinction between those results that only use the cross section result (which agrees with using "the results of this paper"), and those that perform a separate template fit with a different signal sample (which are not derived from the results as presented in the "Results" section, but constitute additional results).

Type A

Abstract, L5: Spell out "(electron or muon)"

L3: Move the reference [1] to the end of the sentence since we not only cite the digits but also what the number means.

L26: no comma in "tttt process when it is produced" because the subclause is constituent for the statement.

L34: "[29]" -> "Ref. [29]"

L40: You use "cut-and-count" only here. I suggest to change to the same naming employed in the rest of the paper.

L43-47: Is there any reason to split these two sentences?

L48: "data-taking conditions" with hyphen

L206: "top tagging" -> "top quark tagging"

L232: "efficiency epsilon_TL for nonprompt leptons" without any commas

L242-243: "lepton failing the tight requirement" -> "loose-not-tight lepton"

L261: You reference table 2 here, but it then only appears at the end of the next page. Push it up to the start of the section.

L307: "as a function of kinematic distributions" -> "in kinematic distributions"

Table 2: I propose to add an additional column for the correlations, instead of highlighting that by a subscript.

L332: "generally consistent" -> "statistically consistent"

L349: You can shorten this sentence by dropping the "(left)" and "(right)" distinction.

L356: "for the cut-based analysis, and 2.6 and 2.7" with comma for clarity

Fig. 1/Fig. 2: Add a "prefit" to the title line, similar to Fig. 3.

L392: Drop the repeated "require such assumptions"

L396: "For reference" -> "For comparison"

Fig. 7: Add "scalar" and "pseudoscalar" to the plot as in Fig. 8.

Fig. 7 caption L2: "with a scalar (upper left), a pseudoscalar (upper right), or both (lower)" with commas

L445: "standard model prediction" -> "SM prediction"