H Maarten,
In spite of the CONF being dead I think we should still comment and also make textual suggestions. The CONF will be the basis of the paper, so whatever is done now does not need fixing later.
Here are my comments
Dear Simone, Francesco, Marie-Hélène,
Congratulations for this important result and the breakthrough in
getting R_jpsi right.
Do we understand right that adding the two bins we get a 4.7 sigma
evidence for New Physics? If so this is not a test of LU, this is
evidence of LU. Adding uncertainties up very conservatively we get
4.2 sigma including RK. This is too serious business for a CONF.
Below are the detailed comments.
//---------------------------------------------------------------------------------------------------------------------
We apply the same level of pedantry as for a paper as we consider
to be a 0th draft of the paper.
Please note that questions asked imply something is not clear in
the draft and therefore please consider a clarification in the
paper.
Physics:
L.245: 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).
L.292: It is not clear why you add Lb backgrounds in the J/psi
modes but not in the NR modes. The Lb->pKmumu decay is
observed.
Table 3: The extremely large corrections for L0H need a sentence
of explanation. Also, your systematics are essentially the
uncertainties on these correction factors. It would be nice to
quote them here.
Section 9 is poorly written and hard to understand. You give a
name to systematic uncertainties an explain how you got them but
not not say which effect you try to catch. That's very critical
for this analysis to be understood. For instance the first
sentence of the simulation correction item is totally mysterious.
Relative efficiency to what? Why would unweighted events take into
account simulation statistics?
L.362: "expected" sounds like wishful thinking. Do they or do they
not?
L.364: It is not clear what this is supposed to catch. Why is that
not correlated?
10.2: Why is the \psitwos cross-check not reported?
General:
- Consistently use \mumu, \epem and \ellell. Sometimes you have
charges, sometimes not.
- Sec. 9: it would be useful if the correlated and uncorrelated
uncertainties were grouped together.
Line-by-line:
Abstract: please add some space around the + and - or the
asymmetric errors. Systematics -> systematic. Correlated
between what?
footnote: ^{\rm th}$
L.11: no colon before equations
L.20: suggest "At \epem collider operating ... resonance, the
ratios... " to make clear what's consistent.
L.29: as being -> as it is (or we do not understand the
sentence)
L.34: and K*
L.46: K* meson
Fig.1: the diagrams could be made more pretty.
Eq.2: is that \decay{\Bz}{\Kstarz\mumu} used in indices? The
spaces around the arrows seem small.
L.57: log-enhanced is jargon.
Table 1: State-of-the-art -> Recent. No bold font in title
line.
L.105: "quasi-identical from a production point-of-view": what
does this mean?
L.106: what is the subject of "has"?
L.112: "cell as that of the lepton"
L.115: upstream -> such (or photons emitted upstream of the
magnet)
L.133: events -> signal
L.135: events -> candidates
L.135: also selected if triggered using ... (now this
grammatically means the CALO and TIS efficiencies are 100%)
L.157: following -> second (we were expecting something to
come, explaining "following")
L.162: events -> decays. Statistics -> data samples.
L.171: \bquark hadrons
L.187: add charges in Kpi
L.192: best is undefined LHCb jargon
L.207: "chosen to go as low as": what do you mean? We suggest "For
this reason, the
q^2 interval used to select B->J/psiK* candidates is [6.0,
11.0] GeV2/c4".
L.228: kinematic distributions (?)
L.237: events -> candidates
Fig.2: the one -> that
L.258: paper -> note
Sec.7: use \boldmath
L.293: \Km{}\pip
L.318: You need to explain what K1 and K2 decay to in your
simulation.
L.333: events -> decays
L.334: effect of the trigger requirements
Fig.3: It seems some bins are outside the -5,+5sigma range for the
pull plot.
Sec 9: Do not use itemize to emulate description.
L.366: remove events
L.387: remove "out"
L.398: uncorrelated between what?
L.412: \B factories. But you called them otherwise in L.20. Also
Fig.6.
Fig.5: The labels are much too small. Are the curves without
systematics useful? And if so, are these without any systematic as
indicated or just without correlated systematics?
Fig.6: align labels at end of axis as in other plots.
[1] Kr\"uger
[11] collaboration
[14] Why not 2016?
[39] does "no. 1" convey any information?
Remove p.21
Fig.8: labels and legend too small.
Table 6: in percent
Cheers,
Patrick