• Be consistent with the use of past vs present tense.

• It won't significantly change the result, but why use powheg+pythia for ttbar together with MC@NLO for single top? It would make more sense to either use MC@NLO for both processes (since there is interference between the two) or else to use AcerMC for single-top, as it is considered to describe the %$ P_T$% spectrum for additional b-jets better than MC@NLO.For the future this should be considered. This enhances the utility for this study as a background measurement for top and higgs.

<-- p { margin-bottom: 0.08in; } -->

is it unreliable or amount of MC not feasible.

What is Z wrt? is it primary vertex?

//1 QCD event pushes up the estimate? - no it's a log plot.

549 The sentence should perhaps be "The predicted ttbar cross section
of 165+11-16 is used instead of the less well determined but
consistent value of 171+-...

which one did you use.

//Fig 7: The last point in the first jet pt for miuons is ok for electrons
but disagrees in muons because of a spuriously weighted event in the
highest bin!

//Up to this point the cuts for MET were released. What cuts are used
for this figure compare to the ones used in making the templates?

//The pt is not uncorrected -- it is corrected for detector effects
but not to parton level. This is a confusion throughout the paper
and one cannot be sure when looking at a plot what the real pt
definition is. In fact the utility for background computations is
limited by the fact that any physics measurement in top or Higgs
will use a reconstructed parton correction. Here the step is
to do a comparison with MC but distributions for the corrected
parton are useful. It is also confusing what JES correction is applied.

///They had
some odd features statistically that misled us to interpret
some of the extremes of distributions.

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

• Be consistent with the use of past vs present.

<-- p { margin-bottom: 0.08in; } -->

L26 – “Thus, ”

L36 – Suggest moving the order a bit, for example: “It is an extension of an earlier ATLAS measurement on both the electron and muon... based on 1.3 pb-1.”

L49-L51 – Line about CDF and D0 results seems a bit out of place here. Maybe move it to L28 after describing the importance of the measurement?

<-- p { margin-bottom: 0.08in; } -->

L67 – composed --> made

L72 – “, respectively”

L75 – trigger --> triggering

L81 – upwards?

<-- p { margin-bottom: 0.08in; } -->

L150 – Suggest removing “as required in this analysis”

<-- p { margin-bottom: 0.08in; } -->

L190 – at --> for a pT...

<-- p { margin-bottom: 0.08in; } -->

A. Jet Selection

<-- p { margin-bottom: 0.08in; } -->

<-- p { margin-bottom: 0.08in; } -->

L271 – accounting --> counting?

L273 – “... for all jets in the W->enu data and Monte Carlo event samples.”

274 Can we see the JVF for the stacked backgrounds? It would be useful
to see where the QCD template goes.

<-- p { margin-bottom: 0.08in; } -->

Figure 2 – Caption: “... low statistics and poor signal to background ratio.”

<-- p { margin-bottom: 0.08in; } -->

L383 – match --> matched.

L412 – samples of events --> event samples.

L426, L429 – I don't think you need to put this statement twice, you've already referred to Table III in lines L405-L406.

L479 – for consistency, multi-jet --> multijet.

L481-L483 – Suggest re-phrasing.

490 The background events with a muon and an energetic jet dont survive?
To what degree and what happens when we have 2fb?

<-- p { margin-bottom: 0.08in; } -->

L499 – multi-jet --> multijet.

L526 – Again, don't think this line is needed since you've introduced Table IV in L507-508.

<-- p { margin-bottom: 0.08in; } -->

Figures 12, 13, 14, 15, 16 – I think first line of caption needs re-phrasing. Maybe: “Uncorrected pseudorapidity difference between...” or something like that...

<-- p { margin-bottom: 0.08in; } -->

L853 – distribution --> distributions.

<-- p { margin-bottom: 0.08in; } -->

Figure 19 – In the caption you mention “and the ratio of theoretical predictions to data” but there is no such thing in the figure, or is it?

We believe that the corrections on the jet energy mean that the measured pt corresponds well to the true pT. So I'd leave that as is.

On appendix A. I don't really see why it is included at all. If it causes confusion I would suggest to drop it rather than raise it in status to the level of a parallel analysis. If they wish to leave the appendix in, maybe they should say a bit more in lines 885-887 about why it's interesting to see the pt20 results as well as well as the pt30 ones.

reliability of simulation of an electron as a jet is very good. You just need to get the calorimeter total energy response right (more or less).

For jets faking electrons the reliability is much worse because the simualtion has to get fine details of the ahdronic shower correct (e.g. every variable that we cut on in the electron finder)

372-381 Confusing as to what is actually changed. Perhaps the cuts should
just be given in a table.
Fig 4: This one has a supressed zero but Fig 2 does not.
The qcd distributions are not smooth and should either be
smoothed (interpolate with sys on interpolation) or discarded
with some upper limit for sys.

They dont correct back to parton but "unfold". How does that impact us?

in tt+jets we will use a similar unfolding approach to this analysis (we use pt >25 jets)

228 Beam spot is 60mm in Z and resolution in Z is 0.1mm

238 When we do the R stuff, will we have to recalibrate it all?

247 Note use of y not eta

312 How well matched to the MET and Mt cuts are we?

313 and on : We should think about Soft lepton tags

371 Is the definition of fake e different here and in top or what we do?
What they did is not clear but they do relax E/p and overall do a isolation
relaxation.

371 Note that they get shapes from the data and then fit met.

480 For fake mu they use the impact parameter.

499-504 They make a pythia dijet sample to look for a muon. Has any
work been done in atlas to try to identify b,c,light in jets?

505 Note the difference in the QCD for mu and e. We should make sure
we keep them separate as they clearly had stuff leaking into distributions
and that was really an artefact of the stats for the estimates.
Hence e and mu are different in pots in fig 6/7

561 By the time we use 1fb or 2fb the consitency of the kinematic distributions
may be a new issue.

500-600 How does the QCD estimate here compare to the top algorithm for QCD
estimates.

Fig 7 and onwards are untagged - this is useful but we will need to see flavor
tagged as well.

584 Final state QED differs for e and mu so again we need to do separately

599 What do they do really ? how does it compare to what we need?

637 Uncertainties for where we start on ttH are HUGE: 30% compared to 2%
for 0 jet. Also for ttbar? But we know ttbar.

652 Interesting to see the quark/gluon difference in calorimeter response.
D0 claims that CDF's 2j bump in w+j is due to this being done incorrectly

670 Note that the full analysis is repeated for the systematics. We need
to make a to-do list for the systematics from this section and see if
it agrees with what we have for our current todo list -- if we have one.

• It won't significantly change the result, but why use powheg+pythia for ttbar together with MC@NLO for single top? It would make more sense to either use MC@NLO for both processes (since there is interference between the two) or else to use AcerMC for single-top, as it is considered to describe the %$ P_T$% spectrum for additional b-jets better than MC@NLO.

Fig 16 caption left -> top, right->bottom.

I Introduction

33 strike 'the full 2010 data sample' -- lumi is the interesting bit now.

II The ATLAS Detector

III Data and Online Event Selection

139 The use of online and on-line is not consistent. It is also imprecise
to say online - it should be defined in terms of the algorithms
and corrections in the introduction. Similarly for offline.
Saying selected "online" is just saying triggered?

IV SIMULATED EVENT SAMPLES

212 Is the min bias here matched to what was measured (by Glasgow?!)

V Off-line Event Analysis

Fig 2: data points that are zero should not be shown.

B. Missing Tranverse Energy and %$ M_T $%

C. Wenu

366 How bad is the reliability of the simulation of a lepton as a jet.
Or do we just expect this?

372-381 Confusing as to what is actually changed. Perhaps the cuts should
just be given in a table.
Fig 4: This one has a supressed zero but Fig 2 does not.
The qcd distributions are not smooth and should either be
smoothed (interpolate with sys on interpolation) or discarded
with some upper limit for sys.

D. Wmunu

E. Detector level comparisons

544 The discrepancies are within systematic are they?

549 The sentence should perhaps be "The predicted ttbar cross section
of 165+11-16 is used instead of the less well determined but
consistent value of 171+-...

Fig 7: The last point in the first jet pt for miuons is ok for electrons
but disagrees in muons because of a spuriously weighted event in the
highest bin!

Up to this point the cuts for MET were released. What cuts are used
for this figure compare to the ones used in making the templates?

The pt is not uncorrected -- it is corrected for detector effects
but not to parton level. This is a confusion throughout the paper
and one cannot be sure when looking at a plot what the real pt
definition is. In fact the utility for background computations is
limited by the fact that any physics measurement in top or Higgs
will use a reconstructed parton correction. Here the step is
to do a comparison with MC but distributions for the corrected
parton are useful. It is also confusing what JES correction is applied.

F. Unfolding

611 Why would electrons need such a huge correction?

F. Overall Systematics

Fig 17: is this a linear or quadratic sum of the other systematics? It should
say "linear" sum or "quadratic" sum

Fig 18: The "no pythia" should be on this one and appears on others.
Also the tick marks don't appear all thetime on all points.
This is true of many figures.

Fig 19: Was the ratio figure left out?

Fig 20: Odd statistical issue in Blackhat-Sherpa for the first bin. Either
need to rerun or remove the point but of course it is odd not to
have all points matching data...

VII Cross section results

869 Would be good to reflect these conclusions in the abstract.

Fig 26: In spite of what the abstract says, sherpa is very poor here

Appendix A

Should show us the QCD templates again as well. They had
some odd features statistically that misled us to interpret
some of the extremes of distributions.

COMMENTS Relevant to Glasgow but not for authors

134 What jet pt do we use in ttH, WH and ttbar? Is the 20 and 30 relevant?
They dont correct back to parton but "unfold". How does that impact us?

228 Beam spot is 60mm in Z and resolution in Z is 0.1mm

238 When we do the R stuff, will we have to recalibrate it all?

247 Note use of y not eta

312 How well matched to the MET and Mt cuts are we?

313 and on : We should think about Soft lepton tags

371 Is the definition of fake e different here and in top or what we do?
What they did is not clear but they do relax E/p and overall do a isolation
relaxation.

371 Note that they get shapes from the data and then fit met.

480 For fake mu they use the impact parameter.

499-504 They make a pythia dijet sample to look for a muon. Has any
work been done in atlas to try to identify b,c,light in jets?

505 Note the difference in the QCD for mu and e. We should make sure
we keep them separate as they clearly had stuff leaking into distributions
and that was really an artefact of the stats for the estimates.
Hence e and mu are different in pots in fig 6/7

561 By the time we use 1fb or 2fb the consitency of the kinematic distributions
may be a new issue.

500-600 How does the QCD estimate here compare to the top algorithm for QCD
estimates.

Fig 7 and onwards are untagged - this is useful but we will need to see flavor
tagged as well.

584 Final state QED differs for e and mu so again we need to do separately

599 What do they do really ? how does it compare to what we need?

637 Uncertainties for where we start on ttH are HUGE: 30% compared to 2%
for 0 jet. Also for ttbar? But we know ttbar.

652 Interesting to see the quark/gluon difference in calorimeter response.
D0 claims that CDF's 2j bump in w+j is due to this being done incorrectly

670 Note that the full analysis is repeated for the systematics. We need
to make a to-do list for the systematics from this section and see if
it agrees with what we have for our current todo list -- if we have one.

General Comments on the paper layout and Style

Comments on the analysis/methodology

• It won't significantly change the result, but why use powheg+pythia for ttbar together with MC@NLO for single top? It would make more sense to either use MC@NLO for both processes (since there is interference between the two) or else to use AcerMC for single-top, as it is consdiered to describe the %$ P_T$% spectrum for additional b-jets better than MC@NLO.

Specific comments on the text

Abstract

line 5: "up to 5 jets" seems to imply that the measurement is in bins of n excusive jets, however in fact the measurement if in bins of n inclusive jets. Obviously "up to 5 or more jets" doesn't work, but there must be a better way to write this.

Data and Online Event Selection

line 101: "On the average" -> "On average"

line 110 and 116: here you separately use "on-line" and "offline" , in section titles you use "online" and "off-line"; be consistent!

Off-line Event Analysis

Jet Selection

line 271: use \mbox to prevent wrap around of the inequality for JVF

Missing Tranverse Energy and %$ M_T $%

line 300: if the reader looks at reference [39] they will find extensive discussion of calcualtion and performance for "missing transverse momentum", not "missing transverse energy"

line 301: use \mbox to prevent wrap-around of the %$ M_T $% formula

References

reference [18] is not the appropriate one for single top production with MC@NLO - you should use, in addition:

S. Frixione, E. Laenen, P. Motylinski, Single-top production in MC@NLO. J. High Energy Phys. 03, 092 (2006)

and

S. Frixione, E. Laenen, P. Motylinski, B.R. Webber, C.D. White, Single-top hadroproduction in association with a W boson. J. High Energy Phys. 07, 029 (2008)

The draft document may be found at this URL: http://cdsweb.cern.ch/record/1390459

The draft document may be found at this URL: [[http://cdsweb.cern.ch/record/1390459][<http://cdsweb.cern.ch/record/1390459>]]
It is version no. 4 entitled: 'Study of W-boson production with jets in pp collisions at sqrt(s)s = 7 TeV with the ATLAS detector'

Additional ATLAS internal supportive material can be found at:

http://cdsweb.cern.ch/record/1361687

Please read and comment. All comments must be received by lunchtime, Friday 21st October, and will be briefly discussed in the glasgow ATLAS meeting.

thanks

Sarah

-- SarahAllwood - 2011-10-19