http://cms.cern.ch/cds/TOP-13-010
Ernesto
Type B
As it is currently written, it is not clear if this measurement really provides a benchmark in the search of ttH production. If we understand correctly, for ttH observation the ttbb cross-section is needed, not the ratio. The measurement and the prediction for ttbb are still affected by large uncertainty so a first step is the measurement of the ttbb/ttjj ratio (and its comparison with theory). We think that this should be stated clearly as otherwise the reader could think that a measurement of ttbb with an adequate level of accuracy for the ttH search has been achieved while this is not clearly the case (at least to our understanding).
One should state clearly in the introduction the menu of measurements which will be described
- measurement of the ttbb cross section with pT_jet > 20 GeV and comparison with MadGraph prediction (which is relevant by itself as it is needed for ttH searches)
- measurement of the ttbb/ttjj for pT_jet>20 GeV and comparison with MadGraph (which is more systematic free measurement)
- measurement of ttbb/ttjj for pT_jet>40 Gev which is what is compared to available NLO QCD predictions
Type A
Abstract Last line “… are compared…”: no statement on the level of agreement between measurement and prediction is given. Why?
L9: ref 7 is a huge volume… Quoting the cross-section and its uncertainty would make the scope of the paper more clear.
L21-22: while it is true that many uncertainties cancel in the ratio, it is not clear how a measurement of the ratio would help in determining the value of ttbb cross-section which is the quantity needed for ttH search.
L31-36: add the accuracy (LO, NLO, ..) of MadGraph and Powheg event generator.
L59: we are enthusiatic about the label “ttLF” for indicating events with two jets from light parton in the acceptance (not sure that “gluon” can be considered really a flavour) Moreover “L” and “F” seems really to indicate different flavours (suggestion: just call them “ttLL”). Addendum (Marco): enthusiastic → not enthusiastic (?)
L67: “Bremsstrahlung” why capitalized? Addendum (Marco): the original german word, since is noun is capitalized (don't know what is the rule in english for loans)
L70: “…using Z boson…” probably you mean Drell-Yan events in a mass range around 91 GeV (so just specify the mass range)
L88: “hard b-tag selection” usually the b-tagging selction are specified by “tight/medium/loose” working point. For non-CMS users “hard” is meaningless as you specify the efficiency, contamination performance. For CMS users who want to reproduce your results, “hard” is useless. Suggestion: remove “hard” and substitute it with the WP label.
L93: is “gluon” technically a “flavour”?
L96: “GeV/c” → “GeV/c^2”
L104-113: it seems there are few “shortcuts” in the description of how the measurement is extracted. If we have understood the procedure correctly 1. only events with exactly four jets are retained 2. the four jets in the kinematic acceptance are sorted in decreasing significance of the b-tagging discriminator 3. ttbb and ttjj signals are extracted by a template fit of the CSV distributions of the two jets with lowest probability of being a b-jet according to the CSV value 4. (It seems there is still an ambiguity in the procedure.) The two distributions can be fitted at the same time with ttbb and ttjj templates having their proportion (the ttbb/ttjj ratio) as parameter of the fit OR the ttbb and ttjj are evaluated separately and you define ttbb/ttjj ratio from the ratio of the individual cross-sections.
L125: “poorly known contribution of Z+bb” do you mean in the MadGraph MC sample?
L144+8: “After correcting… discriminator.” This probably clarifies our previous comment on L104-113 on how the ratio of the cross-sections is extracted. Perhaps one should write it before and here just discuss the systematics on the ratio.
L145: specify at which QCD order the MadGraph and Powheg predictions have been computed.
L174: add a statement on the level of agreement found with the predictions.
Table 1 Please specify if the ttjj signal is the sum of ttbj+ttcc+ttLF. Caption: “The given uncertainty… is statistical” → “…only statistical” Also: the statistical uncertainty is reported only on the total while it should be written for each contribution.
Figure 1: label on the y-axis “Jets” should be “Jet” (singular)
—
Stefano C.
Type B
Congratulations to the author for this important measurement. The paper is well written and easy-reading. Some general comments follow.
L1-23 (introduction): The paradigm: h(126) → couplings → top coupling → ttH → ttjj as a candle for ttH observation makes sense to me, but it can be explain more clearly by reviewing both the order of the sentences and the connection between them. For example: - consider swapping the order of the 2nd (L 3-4) and 3rd (L 4-6) sentences - the 1st sentence (L 1-3) seems completely disconnected from the rest - a strong indication of the Higgs coupling to fermions and top in particular comes from the rate of production which is driven by gluon-gluon fusion (fermiophobic Higgs were excluded almost immediately): please consider including a sentence also on this - I would add a couple of Feynaman diagrams for your signal
L 122-144 (systematics): many source of systematic uncertainty are poorly explained where not simply mentioned as numbers (more details in Type A). I would expand a bit this section in order to give more details
Type A
Abstract: I am not sure you want to explicitely tell the reader, already in the abstract, the method how you extract the signal from your selected events. Usually you put there: brief description of the measurement and motivations, dataset used and summary of the main results
L 14: I would remove “however”
L 20: is “dilepton triggers” without any introduction a bit jargon?
L 102: is ttH expected contribution calculated inclusively with respect to the Higgs decay? I would specify
L 107: Is there a particular reason for doing so? Statistics in the MC? I would specify the reason explicitely. Has it been checked the effect of fitting them separately?
L 110: on which basis the 1st additional jet is chosen over the 2nd? pT? Please specify
L 114-116: why ttbb efficiency is greater than ttjj? One would naively expect the opposite, since b-tagging is applied on the former…
L 129-130: how do you assess the uncertainty on the c-jet scale factor?
L 133-134: how is pile-up affecting the result and how is this uncertainty estimated? Please specify
L 137-139: both these sentences sound obscure to me
Marco M.
General comment:
Overall I find the article difficult to read, mainly for the presence of many sequences of short sentences, sometimes poorly logically linked one from another
From the title one would expect the main result of the paper is the absolute cross-section for the process pp→tt+bb. The abstract instead starts from the ratio. Since the relevant quantity for the search of pp→ttH is the absolute cross-section we would recommend to invert the order in which the results are presented. The abstract lacks some quantitative figure of merit of data/theory agreement.
L 52: hard (→ high transverse momentum?)
L 57: in → among
L 83: (nominal) mass of the Z boson
L 92: one should explain why the choice of a unity scale factor for charmed jets.
L 93-94: repeat the same concept as in L 122-128.
Table 1: it is not clear how undecayed top outside the acceptance can contribute to the raw event count within the acceptance (tt others class)
Giacomo
Type A
l.49 Is [14] the only theoretical work exploring the subject? You were referring to [8-14] early on. I would avoid quoting just one work out of many. Moreover it's probably unnecessary to quote anything at this point
l.58-l63 Are the categories mutually exclusive? If not, do overlapping eventw exist? If yes, where do you put (for example) a “ttccLF” event?
l.62-l.63 dominantly→ dominant. I don't understand these 2 lines. From where do the merging jets come from, and why should this merging make this contribution dominant? In general, I think this paragraph can use some rewording.
l.91 The scale factors → Data/MC scale factors
l.100-l.101 From After to decay: this sentence does not make much sense as it is. I suppose you mean that 97% of the selected events contains 2 b quarks, but you will need to rephrase this sentence.
Table 1 It would be nice to put a line with the expected ttH contribution in the different channels, as you quote 12 ttH event in total
l.104-l.113 This section should be meoved after line 116.
l.164 Table 2 contains all the cross sections. You may want to move this sentence at the bottom of the paragraph and say that all the values are reported in table 2 or something similar
l.166 You already quoted the predicted value. You can remove it from here.
Type B
Abstract: It's not clear what the main measurement is about. The title states that the paper is about ttbb xsection, but the abstract suggest it is about the ratio of ttbb and ttjj xsection, with the 2 xsections measurement as a side results. I would suggest the abstract needs rewording.
I think it's missing some (even brief) discussion about what uncertainties are introduced when correcting for acceptance, and which ones get modified when modifyng the jet pt threshold. A few lines explaining how systematics are affected can be helpful
General remark: From the paper, it seems that reference [14] is the only relevant theoretical piece on the argument. Nor in the motivations nor in the results any other paper is cited. Is this the case? Moreover, this single paper (that is not even yet published) is referenced to at least 6 times in the body of the text. I would suggest to include more theoretical references (if they exist) and to remove some of the redundant references to [14]. As a side note, in the reference one of the authors is missing (Bevilacqua)
Marco C.
As a general remark the paper, although quite complete and of descent quality, looks more like a list of different items and does not keep the reader focused on one single subject at the time. Indeed the paper presents different subjects with results both on absolute ttbb, ttjj cross-section determination and their ratio but it mixes them together rather than differentiate what is relevant for future ttH and what is relevant now for MC comparison. (in the abstract, in the body of the paper, in the conclusions) More specific the perspective of a measurement of ttH production is linked only qualitatively to the actual measurement of ttbb cross-section, and no quantitative arguments are given to link the actual precision reached at CMS for ttbb with a sensitivity in the potential observation of ttH. Even the language with many short sentences more than a wider reasoning,I found difficult to follow. I would suggest to show few plots to illustrate the analysis cut, to demonstrate that the analysis is solid and to facilitate the understanding of the reading Globally I found the paper a good draft but it still needs a wider review.