Dear all,

The Bs->mumu paper got very positive reviews from PRL. For you enjoyment I single out question 5 from Ref A:

5. The uncertainties on the data points in Figs. 1 and 2 should

probably be mentioned in the captions. How are they calculated? The uncertainties are asymmetric in Fig. 1 and Fig. 2 (top), and appear to follow what one would expect from poisson variations of the data. In that case, given that there are a number of empty bins, shouldn't one also include upper error bars for the bins where N=0? Then, in Fig. 2 (bottom), the uncertainties are symmetric, so what do the uncertainties represent here?

Good we invited Ivo. But I think we need a reference for what RooFit actually does.

Cheers,

Patrick

-------- Forwarded Message -------- Subject: referees' report Date: Thu, 23 Mar 2017 11:44:11 +0100 From: Matteo Rama matteo.rama@cern.ch To: lhcb-paper-2017-001-reviewers lhcb-paper-2017-001-reviewers@cern.ch

Dear RC, all, please find below the replies from the PRL referees. The reports are quite positive. A bit surprisingly, the referee B's report (attached) does not contain any specific question or remark on the paper.

Cheers, Matteo

-------- Messaggio Inoltrato -------- Oggetto: Your_manuscript LC16542 Aaij Data: Wed, 22 Mar 2017 22:55:57 -0400 Mittente: prl@aps.org Rispondi-a: prl@aps.org A: matteo.rama@pi.infn.it

Re: LC16542 Measurement of the B 0 s rightarrow mu + mu - branching fraction and effective lifetime and search for B 0 rightarrow mu + mu - decays by R. Aaij, B. Adeva, M. Adinolfi, et al.

Dear Dr. Rama,

The above manuscript has been reviewed by our referees. Acceptance of your paper for publication is likely, but we first ask you to consider carefully the appended comments.

With your resubmittal, please include a summary of changes made and a brief response to all recommendations and criticisms.

Yours sincerely,

Robert Garisto Editor Physical Review Letters Email:prl@aps.org http://journals.aps.org/prl/

IMPORTANT: Editorial "Review Changes" http://journals.aps.org/prl/edannounce/PhysRevLett.111.180001 ---------------------------------------------------------------------- Report of Referee A -- LC16542/Aaij ---------------------------------------------------------------------- For more than 20 years, the elusive decay of the Bs meson into a pair of muons has been searched for in earnest by a number of high energy physics experiments. From the beginning, this was seen as an important channel to search for signs of "new physics", be it supersymmetry or something else. Limits, and ultimately measurements, of this channel put strong constraints on theories beyond the standard model. Along with the failure to find new physics in other searches so far, the measurement of the Bs to mu+mu- branching fraction became even more important.

Just two years ago, the CMS and LHCb collaborations reported the first observation of this decay in a tour de force analysis of their combined data. With more than 250 citations in just two years, this is one of the most cited publications by the experiments. An observation by a single experiment is now reported for the first time with the present measurement. In addition, an interesting excess in the search for B0 to dimuons is not confirmed in the present paper, and the first measurement of the Bs effective lifetime is shown. Both of these results constrain new physics in a relevant way and are important results on their own. There is no doubt that the paper proposed here is one of the currently most important result by the LHCb experiment, and that the result has wide implications on high energy physics in particular, but that it will also attract attention among non-specialists since the result touches upon some of the most fundamental questions of physics.

The paper is well written and at an appropriate level for the subject. There are sections in the middle part which are somewhat difficult to grasp for the non-expert, e.g. regarding the discussion of the various background contributions, but I believe the level of detail is about right, and sufficient references are given for further reading on the technical details. In the end, this is a rather complex measurement of a phenomenon at the 10^-9 level, so one cannot expect all questions to be answered in a short letter. The relevant details however are given, such that the reader can have confidence in the scientific correctness of the results.

I only have a few, relatively minor specific comments:

1. This is just about clarity: About the two normalization channels, the text states "The normalization factors combining the two normalization channels are...". For the reader not familiar with such analyses, shouldn't one specify how this "combination" is statistically performed?

2. Regarding the B0 -> mu+mu- limit, the text explains that "an upper limit... is set using the CLs method". There's a vast variety of "methods" that people call "CLs method" (there's "LEP style CLs" and "LHC style CLs" and so on), which for example differ in the way nuisance parameters are treated etc. I assume the authors are not exactly using what has been described in the given reference from 2002, so I think at least half a sentence of additional explanation is necessary.

3. Just after that, the model dependence of the Bs branching fraction measurement is discussed, leading to a non-negligible additional uncertainty. However, this additional uncertainty seems to be neither included in the final result (one could write it as B +- stat. +- exp.syst. +- model dep.), nor is it mentioned in the abstract or summary. Because it is important for the interpretation of the result in other contexts, why not at least mention it in the abstract, summary, or result?

4. The discussion of the lifetime fit uncertainties should be improved a bit. The text states that "The accuracy of the fit... is estimated using a large number of simulated experiments with properties similar to those found in the data". This is fuzzy and unclear - what are "similar properties"? One has to guess that some unknown properties were varied in some unknown way by some unknown amount. It would help to be more explicit here.

5. The uncertainties on the data points in Figs. 1 and 2 should probably be mentioned in the captions. How are they calculated? The uncertainties are asymmetric in Fig. 1 and Fig. 2 (top), and appear to follow what one would expect from poisson variations of the data. In that case, given that there are a number of empty bins, shouldn't one also include upper error bars for the bins where N=0? Then, in Fig. 2 (bottom), the uncertainties are symmetric, so what do the uncertainties represent here?

---------------------------------------------------------------------- Report of Referee B -- LC16542/Aaij ----------------------------------------------------------------------

See Attachment: lc16542_report_1_b.pdf