Comments by Adinda de Wit

Type A:
Abstract: Do we normally define acronyms in the abstract? It looks odd to me (I recognise 'SM' is used often, so from that point of view I understand why it is done)
L9: "Heavy scalars and pseudoscalar bosons" -> "Heavy scalar and pseudoscalar bosons" (ie scalar bosons and pseudoscalar bosons?)
L34: presented in [29] -> presented in Ref. [29]
L 207: "expected significance" -> "expected sensitivity"
L331: "overlaid" -> "stacked" (I think)


Type B:
General for figures: the % written in the legend is not easy to read. In the bins empty of data, should there not be an error bar corresponding to ~1.8 events?

General for interpretations: these are all so quickly described it is very hard to follow them for a non-expert. If it is possible some elaboration on the models considered (beyond a single paragraph per interpretation) would be useful. 

L7: "Moreover, due to ... particles and operators" -> it seems weird to state that the "production" can be significantly enhanced by BSM effects due to the SM production cross section being small. I sort of get what you're trying to say ...


L61-65: I don't understand the discussion here. Is it necessary to state that the cross sections reproduce those that were used in a previous paper? What if the cross sections didn't reproduce those in a different paper but gave a better understanding of the theory? It seems redundant. In addition, the sentence "These cross sections  - the two Higgs doublets" contains 4 commas and is therefore hard to read. Finally, tan beta is introduced but not the mixing angle alpha

L 152: Why can't the isolation requirement be described? It is not easy for the reader to look up reference 56. 
L 157-163: This discussion is not easy to follow. Would some equations help here?

L219: What happens with ttW in the BDT analysis. It is not constrained via a control region, so is it constrained in the signal regions or is it fixed to the SM expectation (within uncertainties)?
L 254: "derived from a control sample enriched in Z->e+e- events with one electron or positron having a misidentified charge" -> How do you derive such a control sample? 

L 290-293 "For each of the Rare, Xgamma and ttVV categories are taken from the largest theoretical cross section uncertainty in any constituent physics process". Why was this done? It would have been possible to treat the constituent processes separately in the statistical analysis and draw them together in the plots. Does the current choice of treatment lead to very large cross section uncertainties being applied to constituent processes which actually should be subject to a much smaller uncertainty?

L293-295: How is the 40% uncertainty chosen, or is it random? (Often when we say 'constrained from data' we don't apply an external constraint to the normalisation)

L 295: "For ttH, we assign a 25% normalisation uncertainty to reflect the signal strength measured by CMS" - This begs the question of how the ttH process is normalised. Is it normalised to the SM cross section or the measurement by CMS? (SM cross section would be more correct as the tttt analysis and the ttH analysis use the same data and would therefore be correlated via certain uncertainties). 

Table 2: How is the impact on the cross section defined? Is it by removing the uncertainty and verifying how the uncertainty on the cross section changes, or by moving the uncertainty to its +1%/-1% values and checking how the cross section changes?


L336: "This fit provides a measurement of the significance of the observation relative to the background-only hypothesis": it seems odd to start by talking about the significance before something has even been measured. Sentence might read better if re-ordered.

L354: Quoting two observed cross sections is a bit strange. Which one is "THE" measurement? (At least they are compatible in this case). Two different expected numbers are fine (comparing the sensitivity of the analysis), but for the observed considering the same dataset, giving only one number would be clearer. I understand that the BDT analysis is used for the interpretations, it might be better just to quote that number.