Dear authors,
We had our group meeting at Nikhef about this nice paper
and we collected our comments and suggestions which we add in
the attachment.
Tjeerd Ketel

General:
The "introduction" section is rather long. It can be shortened,
as it contains substantial information of the detector, trigger
and simulations, which should be moved to the "data sample
and event selection" section.

The requirements in trigger and further data selection
should be put together without mentioning those that will
be superseded by a later selection. The selections on pT 
for instance should be found at only one place and not spread 
over different places.

Nowhere is mentioned that B+->K0pi+ also stands for its
charge conjugated (even CP conjugated) counter part.
Lines 112-115 do hint to it, but not so prominent.

Title:
We present only a "branching fraction ratio" and not 
"branching fraction measurements".
If "ratio" were to be added, also the order of KSpi+
and KSK+ should be reversed.

Abstract:
"A search for ... is also performed using a
subsample ... resulting in ..." could be replaced by
"A subsample ... is searched for ... and results in ...".

Line 2:
Replace "an interesting laboratory" by "a laboratory"
or "a tool".

Line 10:
Remove "More precise measurements ... of these decays
... pi K puzzle".
It seems that "these decays" refer to other decays
than of those in our paper; maybe I am wrong. 
A "discussion" section is not included, but the "results
and summary" section does not come back on the K pi puzzle.
It does not become clear how our B+ decays will shed light 
on the puzzle.

Line 11-15:
This is not a new paragraph as it continues the discussion
on the combination of different measurements started in the
previous paragraph.

Line 15-16:
The decays in Ref.[10] are B0 decays and not B+ decays.
Is this then relevant for our paper?

Line 17-19:
Why "naive averages"? Cannot we trust them?
In the analysis paper we quote the number of selected events:
3229 K0pi+ from Belle. This number is larger than our 1804
fitted events. Also the uncertainty of ACP(K0pi+) of Belle
is smaller than ours. Should we not discuss why we are
so inefficient on these decays, while for charged 2-hadron
decays we have at least a factor of 10 more events than Belle?
 
Line 48-51:
Move "At least ... without the considered track"
to line 64-65 where the selection is refined.
We can use the definition if Chi^2_IP in those lines.

Line 53:
Please, define "fully simulated" which is used in several 
following places as "full simulations".
Or remove "fully" and "full".

Line 60-62:
Move "This analysis is based on ... 8 TeV" to line 23
and make sure that the subsample of 1 fb^-1 is related
to the 7 TeV part.

Line 70, 76 and 79:
Remove "signal".
The requirements apply to what is later called both signal and background.

Line 75:
Replace "Tracks that are consistent with leaving hits in the 
muon detectors are also dropped from consideration" by
"Tracks that seem to pass through the calorimeter by the presence 
of hits near their extrapolations in the muon detector
may correspond to leptonic decays and are, therefore, not selected."

Line 77-78:
Replace "Tracks lying in a fiducial volume with large detection
asymmetries are vetoed in the analysis" by
"Only tracks inside a fiducial eta-phy region with small detection
asymmetries for B+ and B- are accepted".
The fiducial volume is usually the one you trust most.
A volume or region has to be defined by its coordinates. 
A clarification of the detection asymmetry, which otherwise has
to wait to line 155, does help here.
It would be good to exchange this line with the first line of the 
next paragraph, as the pT requirements continue.
Better would be to move both lines to line 64-65, where the pT 
selection is discussed.

Line 79:
The "scalar pT sum" can be avoided by requiring that "in addition
to the individual pT requirements the sum of the absolute values
of pT should be larger than 4 GeV/c".

Line 81:
Replace "greater" by "larger".

Line 85:
Replace "Pairs ... are reduced by ..." by
"Pairs ... are shown in simulations to be suppressed by ...".
Only the number of pairs can be reduced.
It is instructive to indicate how obtained this knowledge. 

Line 90:
Add to "background"
"by using variables mentioned above and a few which are introduced
below".

Line 95:
The specification "in a cone defined by a circle of radius 1 rad
in the pseudorapidity-azimuthal angle plane" should start with
"for tracks ...".
Still it does not make sense. Please, rephrase this sentence. 
Pseudorapidity should have nothing to do with it. 
I can imagine a cone in space with an opening
angle theta of 1 rad, but not with a circle of radius 1 rad.
Is the pT in the equation for ApT also a scalar sum?
In that case Sum pT could also be used in line 79.

Line 137:
The long subscripts for "in ratio of efficiencies 
r_{B+->KSK+ as KSpi+}" can be replaced by 
"in ratio of B+->KSpi+(K+) efficiencies r_{KmisID}" 
followed by the ratio with "pimisID".

Line 248:
Replace "This is the first search and upper limit on 
a charmless and bottomless Bc+ meson decay" by
"This is the first upper limit on
a Bc+ meson decay into 2 light quarks."
A search is not very serious if not even an upper limit
can be reported.
"Charmless and bottomless" is clumsy and difficult to 
combine. The second best could be "heavy-quark-less".

Line 263,264:
Replace "searches for other charmless decay modes, such as
B+c->Kbar*K+ and B+c->phiK+, may reach the level of
sensitivity needed to observe such decays" by
"other two-body B+c decay modes to light quarks as in
Kbar*K+ and to phiK+ for example may be detected".

Fig.2:
Replace "(Right) Confidence region bands (solid line)
obtained using the Feldman and Cousins [36] approach 
and including systematic uncertainties. 
The dotted and dashed lines are the central value
and the statistical confidence region bands, respectively.
The observed number of events and the corresponding upper limit
are discribed by the gray lines" by
"(Right) The number of events and the corresponding value of rB+c.
The central value (dotted line) and the upper and lower statistical
confidence region bounds are obtained using the Feldman and Cousins [36] 
approach (dashed lines).
The solid lines includes systematic uncertainties.
The gray outline of the box show the obtained upper limit of rB+c 
for the observed number of 2 events."

Comments of Patrick below will be included in the same way as the other
comments and sugegstions above.


Hi Tjeerd,

Here are my textual comments on the Ksh paper. There's one more 
fundamental thing: They should not assess a systematic error for the BDT 
selection. That's already accounted for in the statistical error unless 
they believe your BDT efficiency measurement is biased. In that case 
there are better methods to obtain a systematic error. A BDT is nothing 
but a cut and they do not add systematics for all their other cuts.

L.2, 164: B meson
L.11: I would find Feynman diagrams helpful.
L.52: b-hadron
L.62: say charge conjugation is implied.
L.81: Why must the B be displaced from "the PV" by 1mm and its vertex be 
"well separated" from all PVs? Try to rewrite that in a sensible order, 
starting with the Ksh vertex and then make statements about the B.
L.105 N_S and N_B should be in math mode (at least the N, maybe not the 
B and S).
L.132: MeV/c^2
L.137: These r_{B->...} are ugly and not re-used anywhere.
L.141: This is surprising. We would have expected an excess of K+ over 
K- in background.
L.157: Formally -> At first order
Do you really need to spell out Eq 4 and 5 separately? And why do you 
merge Det+Prod in one asymmetry? It's 2 different things.
L.168: You should not assess a systematic error for your BDT selection. 
That's already accounted for in the statistical error unless you believe 
your BDT efficiency measurement is biased. In that case there are better 
methods to obtain a systematic error.
L.177: Are they in reasonable agreement before or after they are reweighted?
Table 1: Make sure all numbers are in math mode
L.212: in B->J/psiK decays.
L.215-216: That sounds like only B->J/psiX decays can be used at LHCb 
(you could have normalised to B->D(KK)pi to take one example). Remove 
that sentence and say what you do rather than what you do not do. We 
suggest you rewrite the paragraph saying you normalise to KsK to J/psiK 
and then correct for the kaon asymmetry for Kspi.
L.232: why only 1fb^-1?
L.234-238: That's a long sentence in the wrong order. Fit describe the 
fit and then the FC method.
L.248: B_c^+ meson
L.257: remove "corresponding to data".
Fig.2: increase space above caption
[37] Add arxiv http://arxiv.org/abs/arXiv:0907.2256

Cheers,

Patrick