Dear all,

The r_K* paper is out. We already gave the proponents a very hard time asking a lot of questions, but they are very good at replying in a satisfactory manner. Please read the paper draft at https://cds.cern.ch/record/2258163 and send comments to Maarten.

See also his slides at https://indico.nikhef.nl/getFile.py/access?contribId=2&resId=0&materialId=slides&confId=577. The numbers have slightly changed (toward the SM). Also note that in the meantime a bug was found in paper https://arxiv.org/abs/1606.04731 and now the two measurements of BF(K*mumu) agree (and the published avlue moves _away_ from the SM).

Note there is quite some disagreement between some people (including me) and the proponents about what numbers to report. They measure the BFs of K*ee in 2 bins but do not report them. The numbers are yet very interesting and we should not let theorists determining them wrong. Also, the ratio r_Psi(2S) has been determined including the systematic uncertainty. This value is interesting as the constraints on Psi(2S)->ee vs mumu are very weak (11%). The main argument of the proponents is religious: had they found this ratio not to be 1, they would have used it differently. I think that would have been the case for R_K* too. They would not have published a value of 10.

Cheers,

Patrick

//===================================================

Here are my comments:

Dear Simone, Francesco, Marie-Hélène,

Congratulations for the very well written paper. We have already given you a hard time with many questions in previous rounds and are generally very happy with your answers. There are a few items remaining.

Please note that questions asked imply something is not clear in the draft and therefore please consider a clarification in the paper. We suggest extensions of the text at some places as this paper is likely to be tread by people not familiar with LHCb.

Physics:
 - L.91: In *one* of the downstream tracking stations? Is that enough?
 - L.264: (already in CONF)  This procedure can cause biases if this fraction of 1-2% is not identical for all your four samples. See https://arxiv.org/abs/1703.01128. A systematic uncertainty must be assigned (and will be small). You should at least say these fractions are the same for all samples.
 - L.264: if most MC are K<->pi swaps, that indicates your PID requirements are too lose. You should say that to avoid giving the impression we have a poor K-pi separation.
 - L.330: why do you need to take the fraction of number of bremsstrahlung photons from simulation? The data tells you that. Or do we miss something?
 - Fig.4: in the bottom plot, you have a white region below the total PDF around 5500 MeV which corresponds to about 50 candidates per bin (zooming in). That's more than the about 10-20 signal candidates in that region. Do you miss a component? Or, if that's signal, then why is the white band narrower at 5400?
 - Why don't you report the measured BFs in q^2 bins? That's very interesting.
 - It is a pity you do not report r_\psitwos which contains a lot of physics.
 - We request you do a combination of the significances of the two bins to avoid theorists doing it themselves and wrong. A combination with R_K would also be useful, at least to be given in talks (and thus in additional material in CDS).

General:
 - We prefer the PDG convention B->J/psiK*, but understand why you spell it as K*J/psi for symmetry.

Line-by-line:
L.19: do we usually write \B factories or \B-factories? We prefer the former (as in \B decay).
L.26: there are many more papers one could cite for LQ models. See for instance https://arxiv.org/abs/1703.09226 and references therein. Obviously you can't cite them all.
L.29: The comment on (pseudo)scalars could be expanded and moved to the paragraphs above where you discuss the physics.
L.45: Bremsstrahlung radiation is a pleonasm. Maybe emission?
L.48: We see what you mean, but a factor cannot be reproducible. Either "unknown constant factor" or "unknown, in a reproducible way".
L.52: either "between 0.045 and 1.1" or "in the range [0.045,1.1]"
L.56: found how? You measure it or you compute it based on the PDG. This is resolved only much later in the text. Maybe "known to be small" or just "is small".
Table 1: [20-22] is not vertically centre-aligned
L.114: "optimal" is a bit vague. We wonder what external people will understand.
L.131-132: propose: "to both electrons", then remove "track" (you don't add to the track anyway).
L.132-136: This needs a rewrite. First it's unclear if you mean signal K*J/psi or simulated. Second, how can you find that it's q^2-independent while looking at K*J/psi?
L.141: the thresholds are also the results of a physics optimisation, not only due to occupancy. The present sentence sounds like we have the same rate for all L0 triggers.
L.143: propose: a fraction that only...
L.165: remove "one" (the second one ;) )
L.176,184,192: never use \Bz as a noun. Here add "candidate".
L.180: "HCAL. However,"
L.195: {\rm th}
L.197: which binning scheme? You have not yet mentioned bins.
L.204: space between 800 and (500)
Fig.2: The default rainbow colour scheme does not work in BW. Choose one of the many printer-friendly ones. On the x axis the K appears bold (could be the printer). Always add charges to particles.
L.223: Section ??
L.227: "allows to" is disfavoured in proper English.
L.237: the \B and \Kstarz candidates and the di-lepton pair.
L.237: We assume you mean the FD wrt the PV for the \Bz. But flight distance wrt what for the K* and the di-lepton?
L.256: events -> candidates
L.257: We stumbled over "required on the 2D plane" as we first thought you meant a physical plane. Propose: "A requirement applied in the plane formed by the corrected..."
L.260: This is not a significance. -> "to maximise the precision on B(B->K*ee).
L.261: Number of expected signal events -> expected signal yields (avoid using "events"). Same for background.
L.269: why the long spaces in the decay descriptors? Probably need some \mbox{} around them.
L.293: \Kp{}\Km
L.294: candidates -> signal
L.296: Add that you fit separately for ee and mumu. Also it is not clear at this level that you mean the two q^2 bins, and not the three intervals including the resonant mode. Maybe say here that it's a simultaneous fit with J/psiK*.
L.330: cite Wouter here.
L.306: well, not all, as you describe below.
L.320: event -> decay
L.322: this sentence comes as a surprise here and does not convey any information. Could you be more specific on how you use it to determine its leakage?
L.330: events -> candidates.
L.332: "trigger PDF" is a bit short. You mean PDF for each trigger category.
Fig.4: font on x axis is too small. Add charges
L.338: \Kstarz -> kaon (not all excited kaons are K*, as the name of the K(1270) suggests)
L.341: events -> decays
L.342: "simulated candidates" or just "simulation"
Fig.5: why is the bottom plot on in log as the corresponding Fig.4 plot?
L.348: we presume "decay kinematics", but more information is required here.
L.362: double ratio
L.362 (already in CONF): "expected" sounds like wishful thinking. Do they or do they not?
L.365 (already in CONF): dislike abusing itemize to emulate description.
L.382: considered. However
L.389: events -> decays
L.394: Double -> double
L.400: events -> candidates
L.416 and 428 and Table 4: use \boldmath
L.419: 1 -> unity
L.420: are therefore considered to be a sign
L.429 and following: LL is an acronym and thus should be in roman. Or consider calling it $\ln L$.
Table 5: [0.517,0.891] and so on to make clear it's a range.
Section 10: Remove the itemize environment. Paragraphs would work as well (you can merge items 3 and 4)
L.458: systematics -> systematic uncertainties
L.462 and 467: remove statistical
L.475-486: would prefer if you just referred to App B and say the plots are there. With the present way, you force the reader to turn pages a lot.
L.491: first run is undefined here. Say 2011 and 2012 as in the introduction.
Table 6: events -> decays. Now that you have all numbers for simulation and data it would make more sense to give the numbers for data and those for simulation in parentheses.
Fig.8 and following: Background-subtracted. simulation.
[1] Model-independent (the inspire record is wrong. We informed them.)
[12] why are these and only these refs concatenated?. Remove spurious no. 7.
[13] remove no. 11
[15] remove the note [,166(2003)]
[16] remove no.8
[40] NeuroBayes