Dear Marcel and others,
Have a look at the comments I collected this afternoon about the Delta ms draft.
Tjeerd ====================================================== Comments on Delta ms
Dear authors, We discussed your high quality draft at Nikhef and come with a few comments and questions about the content. Also suggestions for the text are given.
Many comments also apply to the Delta md paper. Please, be so kind to apply similar corrections to both drafts.
Comments on the content: ======================== line 12-19: Are the cuts really equal to those of the road map? If they are not, add an overview of the cuts in an appendix.
line 20: Why don't we include the Same-Side flavour tag as we did in Refs [3] and [4]?
Fig. 2: The combinatorial background candidates show a similar \eta_c distribution. Is this good or bad? Please, discuss this. Naively, we expect a peak at 0.5 for random combinations.
Eq. (6) belongs in the introduction.
Comments in the text: =====================
Abstract: Change to 'The mixing frequency \Delta m_s is measured with 35 pb^-1 data taken during the LHCb physics run in 2010.' '... with a proper time resolution of 36-44 fs. By using ..." 'The preliminary mixing frequency ... is measured.'
The blinded analysis is too prominent in the abstract. It is just one of the tools among many. Typing error frequency. The number 35 pb^-1 should be known between 34.5 and 35.5. Why don't we stick to one number for present analysis, 37 pb^-1 ? 36-44 fs does not look like an average. The presented value is the preliminary one for these data.
line 1: Replace by 'The measurement reported by CDF in 2006 [1] is'
line 17: '... its momentum and transverse momentum and on those values of its decay particles'.
The pion and kaon are not stable particles. line 20: '... have been optimized ...'
line 21: 'This calibration have been adjusted ...'
Eq. (2) and text and Eq.(8): Replace 'P_i^B' by 'P_i^bkg'
P^B for background when the B-meson is the signal is confusing.
line 29: '..., q is the decision of the flavour tagging algorithm, and eta_c is its calibrated mis-tag probability.'
It was not clear that q is the decision outcome of the algorithm.
line 39: 'Without proper time uncertainty, without event selection bias and without flavour tagging, the proper time probability distribution of the B_s signal is described by,'
Important to exclude flavour tagging here in the text. What is 'perfect'? Why '... or ...'?
line 43 (sub 10): 'The width of this time distribution is defined here as the width of a pull distribution which directly corresponds ...'
We need a definition of the pull distribution mentioned in the caption of Fig. 1(left).
line 43 (sub 15): 'into account by convolution with a single Gaussian G, the proper ...'
It is nice to mention the convolution operation in the text.
Figure 1: The figures are too small. Remove '~' before 35 pb^-1. Put 'LHCb preliminary', 'sqrt s = 7 TeV', and '35 pb^-1' in one box.
line 46: '... candidates as shown in Fig. 1(center,right).'
Eq.(6): 'P_t(t|\sigma_t,\epsilon) = ...'
Without this extra \epsilon, Eq. (5) and Eq. (6) disagree. We have to change the name when we change the definition.
Eq. (7): '\epsilon(t)= \frac{N_rec^MC(t_rec)}{N_gen^MC(t_rec)}\approx \frac{N_rec^MC(t_rec)}{\frac{1}{\tau}........'
'reconstructed MC events' is unusual in an equation and in this case the meaning is not clear. Approximate \approx is more appropriate than proportional to \prop.
lines 48-51: Rewrite these sentences. For example: 'The efficiency function is determined from Monte-Carlo input. If we chose this \epsilon(t) as the proper time acceptance, the systematic uncertainty in the asymmetry of mixed and unmixed events used for the determination of the mixing frequency will be small as \epsilon(t) cancels in the ratio of the difference and the sum.'
The efficiency function changed into the proper time acceptance without a word. 'potential not completely correct' is a strange description. 'generous uncertainty' does not make sense. Asymmetries are calculated from ratios of differences and sums. Such a statement helps to understand line 48.
line 68, 69: Explain what you mean by 'Generator level mass templates'. Avoid to refer forward to line 75 with 'offset observed in the reconstructed B0s mass in data'
The 'templates' are probably mentioned in Figs 3 and 4 An example, how this is done for one resonance background, may help to understand what is meant by this jargon.
Figs 3-5: The text and numbers are too small.
line 89: 'The signal proper time probability distribution function is changed by the addition of the tagging information as'
Eq. (6): 'P_t(t|\sigma_t,\epsilon,q,\eta_c) = ...' The equation is incorrect for q=\pm 1.
Same comment as for Eq.(6) The product of cosh and cos should not occur. The definition of q=0 is strange. q is either -1 or +1. Maybe a solution can be found without q=0 and by using the tagging efficiency, \epsilon_sig.
line 118: Add 'Or that we have a statistical fluctuation.'
line 124: Correct 'rage' to 'range'. Also in \Delta m_d draft. Change 'generously' in 'conservatively'.
line 127 Change the word 'need'.
It gives the wrong message. If you needed it the analysis should have been done with 2G.
line 131, 132: It is not clear why 'extreme scenario's and very hard cuts' are used.
Table 3 and line 142: Why are \eta_c and \sigma_t(. ?) taken together in one contribution? As asked before; why would we expect same \eta_c p.d.f. for signal and background?
line 155: Add 'preliminary' as proposed for abstract.