Dear authors, ARC, congratulations on this well-written, clear paper. Below we have some questions/comments to your attention.

Best wishes for a speedy publication. The Torino group

Type B comments

---

1. The hardest part of the analysis, to the non-expert reader, seems to be the accurate identification of Z decays coming from merged jets. To convince that the jet substructure identification works, a check on data would be appropriate, if possible. For example, one could measure the ratio Z→jj / Z→mu+mu- with merged jets, taking into account the appropriate efficiencies, and check that the ratio of branching ratios is consistent with expectations. If something like this was done, it is worth mentioning, if not, please consider a similar test.

2. In the discussion of systematic uncertainties, it is not clear how b-tagging, the uncertainty on efficiency of which is up to 60%, can increase the sensitivity by 50%. Can you explain ? Table 2 is very important, but not very clear in understanding the origin and effect of the uncertainties.

Type A comments

---

L4 Invert : “Searches … are of particular interest”

L 77 “ 170 or 180 GeV” → (180) GeV

L 82 “corrected for zero suppression effects” → the ecal corrections are so many and complicated that I would not mention zero suppression, which to my knowledge is irrelevant at pt>150 GeV

L 83-88 I understand this is boilerplate PF descriptive text, but why bother describing electron and muon reconstruction, which are not used in this analysis ?

L111-112 if possible add a sentence to explain briefly why two different algorithms are chosen at 8 and 13 TeV

L132-141 hard to understand

L187 acceptance mentioned here for the first time, it is not defined and you don't say how it is calculated

Throughout the paper subjet → sub-jet

L207 - 208 I would omit saying that this fit function was chosen among others by using a Fisher test. It is a function that describes the background well, period.

L248 source of the systematic uncertainty → remove “the”

L265 “ only the effect .. on signal acceptance” → what does it mean ? that the impact on yield and efficiency was not taken into account ?

Table 2 : In this form, the table is quite confusing. It is not clear if the “magnitude” column is the relative uncertainty on sigma x BR or on the quantity indicated by the last column. In the first case, I would remove the last column. For the label of the second column “Magnitude” is best described as “relative uncertainty” . In the caption: “ the second column, the third column indicates ” → there is an offset in the column count, “Source” is the first column, therefore “Affected quantity” is the fourth, not the third, and “Effect on yield” is the third, not the second.

I concur with the above mentioned congratulations to the authors and ARC. Very well written and well exposed paper.

Type B comments

---

1. in the background modelling section you mention that “ any possible systematic bias from the choice of the functional form is small compared to the statistical uncertainty of the fit” but you never mentioned quantitatively the magnitude (or range of magnitude) of such statistical uncertainty (can only be seen from the plots)

2. the description of the systematic uncertainties (and of the table with the signal selection efficiency systematics) could benefit from a less concise exposition.

Type A comments

---

L 58 nonisolated particles –> non-isolated particles

L 82/87 corrected for zero-suppression effects –> why is this one correction mentioned above the many others that are applied to ECAL and HCAL energy deposits?

L104 candidates in a cone ∆R –> in a cone of radius ∆R

L112 The 8 TeV analysis employs the Cambridge–Aachen (CA) clustering algorithm [25], while the 13 TeV analysis uses the anti-kt algorithm –> what are the main reasons/pros for the two choices? (once such a difference is pointed out in a text, the reader would like to know why)

L118 The requirement on jet “eta” allows to further suppress the background from γ+jets and QCD multijet events –> not 100% clear, further suppress with respect to what?

L158 a resonance somewhat broader –> could you avoid the qualifier “somewhat”?

L183 the background rejection due to b tagging exceeds a factor –> the background rejection in this category exceeds a factor

L186 into the two SRs. –> in this sentence it is probably better to write without acronyms: into the two search regions.

L138 narrow and wide resonance –> narrow and broad resonance (please keep using “broad” when referring to the “wide” resonance)

L241 remove wide, not really needed, it is clear that you are discussing this case

L249 Most of the uncertainties affect the overall signal efficiency, and only the b tagging efficiency uncertainty can result in signal category migration. –> is one worst than the other? is one going to affect more the result?

L270 The background-only fit is consistent with both 8 and 13 TeV data. –> please rewrite it in a less concise format

Table 2: the caption needs to be re-adjusted, the description of columns looks wrong.

L280 The expected and observed limits for spin-0 resonances –> please specify (again) on what are those limits intended to be (on the production cross) - same comment is valid for the Table 3 caption.

L290 This ratio increases –> which ratio? you did not mention it before

L306 in in → remove one in

Figure 5 showing relative contribution of –> showing the relative contribution