Review for TOP-17-003

Colour code

• green : comment is addressed
• blue : work in progress
• purple : answering question, but no action taken
• orange : Assigned person to answer the question

http://cms.cern.ch/iCMS/analysisadmin/cadilines?line=TOP-17-003

Comments on PAS v6

General comments

Rebeca (27/07/2017)

• In both the title and the abstract, you write "of a Top Quark with the Higgs boson", and that sounds weird, either use "the" top quark or "a" Higgs boson. There are other occurrences in the text and I suggest to also fix that.
• Since you train signal Vs all backgrounds, are the samples for training of the BDT orthogonal to the ones used in the analysis?
• In the figures, are the FCNC contributions multiplied by any factor?
• Check the recommendation for pu systematic, I think it is 4.6%
• How the combination is almost in the 2sigma level from the observed above the expected, when the first two categories are both below, and the b3j4 is spot-on. Is the combination driven basically by the b3j3 region?
• Why the uncertainty on the signal strength is so asymmetric?

Textual comments

Rebeca (27/07/2017)

• Abstract - Single top production for the signal process is considered for the first time to experimentally probe top-Higgs FCNC couplings -> Single top production is considered for the first time as signal process to experimentally probe top-Higgs FCNC couplings
• Abstract -The final state -> A final state
• Abstract -Add the year of the data-taking here from the predicted background
  
  
• l.2 if it is recently discovered, it has to be new
• l.3 standard model
• l.3 remove "particle"
• l.3-6 This sentence is probably not necessary.
• l.14 I think you need to rephrase here, after spin space it is not easy to read.
• l.16 is probed in this analysis
• l.16 an associated production -> the associated production
• l.21 the search at ATLAS -> the ATLAS search
• l.28 pb should stay close to 36.98
• l.30 The anomalous top-Higgs FCNC couplings are expected to be small not sure if we need this info here unless it is linked to the rest of the paragraph.
• l.32 A consideration -> The consideration
• l.50 Generation of signal -> The generation of signal
• l.52 in generation -> in the generation
• l.55 =0 should not be separated in another line
• l.57 The SM top quark pair production (also decay right?) because your TT process is SM production, but FCNH decay
• l.65 a single top production -> single top production
• l.70 the W boson -> W boson
• l.71 withW, Z, generated with MADGRAPH, as -> with W and Z, generated with MADGRAPH; as
• l.74 simulations -> simulation
• l.76 pythia version should go together with pythia
• l.78 t-channel
• l.83 this also should go together with the number and not in the next line
• l.126 CSVv2 b -> I am afraid this is jargon, you should define in a few words what does the CVS algorithm do (appears also in Fig 2, l164)
• l.143 you never defined "MC" (and went through great pains to just have "simulation" until now, which is of course better) (appears also in Fig 2, l.168)
• Figure 2: cosmetics needed, the hatched area is very harsh; the b3j3 label is a bit disruptive (can you move it somewhere else?); bMVA TopHLepbb is too much jargon; and the y-axis range should be enlarged in general.
• Figue 3: on predicted -> on the predicted
• l.159 Systematic uncertainty estimation
• l.160 Systematic uncertainties considered in the analysis are the ones that affect the normalization .... distributions. -> Sources of systematic uncertainty that affect the normalization of the predicted background as well as the shape of the distributions, are considered in the analysis.
• l.196 remove "No significant deviation from the predicted background is observed."
• Table 1, 2 I would suggest to add the uncertainties with numbers and not percentages, and try to fit the numbers all in the same line. Also add toral background just below Data, or Data after total background.
• Figure 5: make lines thicker, legends much larger, and Hut and Hct should go somewhere else, they are flying there in the middle and looks a bit weird.
  corrected (Predrag, 03/08/2017)
• Reference [40] There is a paper already for this one: https://arxiv.org/abs/1610.04191

• Line 3: Standard Model -> standard model
• Line 21: "... and H->tautau" -> "... nd H-> tau tau decays"
• Line 97: only deposits in the calorimeter towers?
• Line 125 "selection criteria includes" -> "selection criteria include"
• Fig. 3 and 4: maybe worth to repeat that signal is normalized to backgrounds expectation.
• Fig. 4: why b4j4 only for KHct?
• Line 184: is it really 1% of the total uncertainty? So, from table 1 and 2, about 0.1-0.2% on the background yields?

ARC comments on PAS v3

Physics/Journal comments

Javier (15/06/2017)

• Figure 2, top left: doubt, why does ST-tuH have many more positive leptons?
  Positively charged leptons come from top decays, whereas negatively charged leptons come from anti-top quarks. Top quarks are anomalously produced by an up-quark from one proton interacting with a gluon from the other proton (analogously anti-top quarks come from anti-up quarks). Since much more up quarks are present in protons than anti-up quarks, more positvely charged leptons are present in the ST-tuH signal. (Kevin, 15/06/2017)

• Which journal?
  We would prefer to submit a Letter. We are open to your suggestions. PRL might be an option. (Kirill, 16/06/2017)
• L192-195: is there any constraint on the length of the paper?
  As we would like to go for a Letter, we kept the manuscript short. (Kirill, 16/06/2017)
• L124: no description of the reconstruction of the neutrinos?
  We have briefly added this information in L 124. Since the method of reconstructing neutrino pz is a well known thing, and as we would like to aim for a short-length paper, we avoided detailing the exact steps of this reconstruction procedure. (Kirill, 16/06/2017)
• L131: maybe you could mention the most powerful variables used in the BDT?
  The details are mentioned just a little bit later in L 138. L131 was to introduce the general methodology. (Kirill, 16/06/2017)
• L203: is this 20% valid only for t->uH or also for t->cH?
  It is only valid for t->uH. The single top contribution to the t->cH limits is much smaller. We have corrected this sentence. (Kevin, Kirill, 15/06/2017)

Textual comments

Javier (15/06/2017)

• L17: I'm afraid this nomenclature for TT (with capital letters) is not maintained for the rest of the paper, since all plots show tt (small letters)
  We have harmonized the naming used in plots and text (Kirill, 16/06/2017)
• L69-70: Which generator was used for these processes? DY processes are not mentioned either
  Information added (Kirill, 16/06/2017)
• L71-73: is there background which does not use Pythia8? From table 2 on AN I don't see any
  Sentence modified. The initial version of it attempted to emphasize the distinction between POWHEG and other generators for the first emission treatment (Kirill, 16/06/2017)
• L35,L173: Reference for 2016 Lumi [11] is incomplete
  Fixed (Kirill, 16/06/2017)
• L75: Reference for tune CUETP8M2T4 is wrong and incomplete
  Fixed (Kirill, 16/06/2017)
• L105: pT for muons is 30GeV, isn't it?
  Fixed (Kirill, 16/06/2017)
• L108: is the same cone for e and mu? The AN declares 0.3 and 0.4 resp.
  Fixed (Kirill, 16/06/2017)
• L109-110: is the same RelIso cut used for e and mu? The AN quotes different values for electrons
  Fixed (Kirill, 16/06/2017)
• L116: it should be noticed that the 2 b-jets are among the 3 identified jets, shouldn't it? Well, I see it is mentioned in next section...
  Adjusted (Kirill, 16/06/2017)
• L148: Systematics, in v1 of the PAS there was a nice table summarizing the effect of systematics which was very appreciated, can we have it back again?
  During the pre-approval we have received the comments that these tables are possibly too busy to be added to the paper. That's why we decided to be more explicit in the text rather than having the detailed table with all the uncertainties. Moreover, it looks only feasible for us to have such table for pre-fit uncertainties and not the post-fit ones as there is no straightforward way (to our knowledge) how to decouple the uncertainties into different components after the fit. (Kirill, 16/06/2017)
• L191: senstivity -> sensitivity.
  Fixed (Kirill, 16/06/2017)

• L8: sounds better without “to be”
  To be or not to be ? Agreed. (Kirill, 16/06/2017)
• L31: … couplings are expected to be small and therefore THEIR contributions …
  Fixed (Kirill, 16/06/2017)
• L86: a description of the PV reconstruction is missing. Waiting for the updated recommendation?
  Added from https://twiki.cern.ch/twiki/bin/viewauth/CMS/Internal/PubDetector (Kirill, 16/06/2017)
• L94: maybe would help introducing the anti-Kt algorithm at this point?
  Added a paragraph from https://twiki.cern.ch/twiki/bin/viewauth/CMS/Internal/PubDetector (Kirill, 16/06/2017)
• L105: is selected -> is selected offline ?
  Modified (Kirill, 16/06/2017)
• L109: description of Irel corrections?
  The part describing lepton isolation requirement rewritten. (Kirill, 16/06/2017)
• L117: medium operating point of the CSVv2
  Modified (Kirill, 16/06/2017)
• L124: are referenced as -> are referenced to as
  Done (Kirill, 16/06/2017)
• L188: separately per event category, as well as combined -> separately per each event category, as well as for their combination
  Done (Kirill, 16/06/2017)

ARC comments on AN v7

Javier (15/06/2017)

• L48: asocciative -> associative
  corrected (Kevin, 15/06/2017)
• L85: i -> in
  corrected (Kevin, 15/06/2017)
• L57-60: It would be nice to quote signal samples
  implemented (Kevin, 15/06/2017)
• L130: As HLT was 2.1, this should be|eta|<2.1 right?
  Yes, corrected (Kevin, 15/06/2017)
• L149: at least 3 jets, right?
  The two constituents refers to the number of PF particles that need to be present in a jet. But the minimal number of selected jets is missing in this section. It's been added. (Kevin, 15/06/2017)
• L251: both single top AND tt
  corrected (Kevin, 15/06/2017)
• L252: missing . after TAble 11 and missing reference to figs 14 and 15
  corrected (Kevin, 15/06/2017)
• Fig 11: I find strange that the BDT does not find an improvement on efficiency for the "All" category in TOPTOPLEPHAD events
  bMVA uses only four most important variables from the kinematic reconstruction in the training. In case of ttbar hypothesis (TOPTOPLEPHAD) there is a larger number of light jets present in event, as well as the hadronic W has quite bad mass resolution. This fact contributes to no improvement seen in the result for jet assignment efficiency. This can be improved by adding more variables in bMVA. However, as in the main analysis we are only considering events already containing at least two b jets, this particular case of all jets treatment in bMVA is not relevant for us. This case was just considered to have a broader coverage of various jet selections to have a better understanding of the impact on the final assignment efficiency. (Kirill, 16/06/2017)
• L272: numbber->number
  corrected (Kevin, 15/06/2017)
• L383: Lumi reported here is greater than in the rest of the AN/paper
  corrected (Kevin, 15/06/2017)
• L437: typo in arrow
  corrected (Kevin, 15/06/2017)

• L208: what does “several times” mean? You do a minimisation fit or just a manual scan? With which step?
  We create toy events for which the 4-momenta are varied according to the probability density functions constructed from the observed differences between objects at reco-level and generator level. The text has been adapted to make this a bit clearer. (Kevin, 15/06/2017)
• Fig 4 to 7: why only for TT signal hypothesis?
  This was a choice to reduce the number of plots in the AN. As these plots look similar for the other two hypotheses, we esteemed it sufficient to show it only for one case. (Kevin, 15/06/2017)
• L243: which FOM do you use to select the best compromise?
  This choice is mainly motivated by the number of events that remain in the high b-tag multiplicity categories. In principle the CSVv2 Tight WP gives the best event reconstruction efficiency, but in b3 and b4 categories the number of selected events was not sufficient to perform a discriminating BDT training, bringing down the separation power between signal and background. (Kevin, 15/06/2017)
• Fig13: b2j4 maybe better grouped with other TT signal bins?
  Background composition is indeed very similar in b2j3 and b2j4 categories. The reason to split in these two categories was in order to have them individually defined for ST and TT signals which have different total number of reconstructed jets. However, the gain is indeed small, as most of sensitivity comes from b3j3, b3j4 and b4j4. (Kirill, 16/06/2017)
• Table 11: can you explicitly say what bMVA is? Is it related to the output of the BDT used for event reconstruction?
  It is. It has been added to the text (Kevin, 15/06/2017)
• L305: To which some of weights are you normalising to? The one from the default scales or the one from the alternative scale?
  The reweighting was done according to the sum of weights from the alternative scale. However, in the next version, this sum-of-weight reweighting is not applied in order to take the normalisation of the ME QCD scale uncertainty into account as well. This is in attempt to understand the issue discussed in this hypernews thread. (Kevin, 15/06/2017)
• L353: is it a TOP PAG recommendation to consider only the effect on the shape for top pt reweighting?
  This is implemented as recommended in this twiki: https://twiki.cern.ch/twiki/bin/view/CMS/TopPtReweighting#MC_SFs_Reweighting (Kevin, 15/06/2017)

ARC comments on PAS v1

Physics comments

• Javier (07/03/2017): Consider changing the electron trigger from HLT_Ele32_eta2p1_WPTight_Gsf to HLT_Ele25_eta2p1_WPTight_Gsf and HLT_Ele27_WPTight_Gsf, for which the scale factors have been measured and approved as presented in https://indico.cern.ch/event/604912/ by Andrey Popov. Take into account/think about the following differences between your current electron selection and the selection used for the trigger SF measurements: 1) Veto on loose electrons, you use pT>10, Andrey pT>20: OK this does not matter provided that the probability of an additional loose electron to fire the trigger is negligible (which is true with the threshold of 20 GeV and even more so for 10 GeV). 2) It looks like you don't apply the cut on impact parameters, but Andrey does, this might be a small effect, but it has not been investigated it. So, strictly speaking, the ID is different.
  We agree with this proposal and will change the electron trigger for the next iteration of the PAS and AN. For 1) we won't take any action, but will synchronize on the IP cuts mentioned in 2). (24/03/2017, Kevin)
  

• Luca (14/03/2017): PAS line 54: is this particular choice of the chiral parameters a natural consequence of the chiral properties of the quarks? Maybe it’s worth to put a comment on it, and also repeat in the summary the values of these parameters used to set the limits.
  This choice is arbitrary. It was investigated that the chirality choice did not have any influence on the kinematics. (24/03/2017, Kevin)
  
• Luca (14/03/2017): PAS line 123: you mention here that the reconstruction is done in three different hypotheses, but you don’t explain how they are used. Also in the note I cannot find many details. My understanding looking at the tables of variables used in the MVA (Table 11 AN-15-097) is that you use different hypotheses for different variables. Is it correct? I think that should be mentioned in the PAS, and some comments on the choices done should be given in the AN.
  Yes, we agree, and we have added the following sentence: All kinematic variables reconstructed under \ttbar semileptonic background and different signal hypotheses are considered, and the most powerful ones are used in the BDT. (11/04/2017, Kirill)
• Luca (14/03/2017): The ST production is an additional process for producing top quarks, but the TT is not, in the sense that a significant BR of t->Hc(u) would decrease the background from ttbar. How is this taken into account in the background prediction?
  The best limits for the BR of t->Hc(u) are currently around 0.5%. As this is lower than the theoretical uncertainty of about 5% on the ttbar cross section, we considered the need to subtract the Signal TT cross section from the ttbar cross-section by an unnecessary complication. We can try to estimate in the limit setting procedure whether this assumption is valid. (24/03/2017, Kevin)
  
• Benedikt (14/03/2017): Replace the prefit plots with post fit plots in the paper.
  Will be done as soon as we fully understand all the systematics. (24/03/2017, Kevin)
  
• Benedikt (14/03/2017): I would like to understand a bit more where your sensitivity mostly comes from. Is it ttbarFCNC or singletop FCNC? --> what are the limits on either of these instead on the combined signal? - How often, for the previous iteration of the analysis, was MVA ST>MVA TT? (slide 32 of the pre approval talk)
  From the limit plots these and these we can see which of the signals contributes the most. For the Hut coupling, both single top and ttbar have comparable contributions to the limits, while it's clearly for the Hct that is mainly driven by the ttbar contribution. There is no unexpected things in here.
• Javier (14/03/2017): Line 148: From table 1, the dominant systematic seems to be scale variations.... followed by Lepton SFs and UE...
  The tables were buggy. This is fixed now and the dominant systematic is the CSVv2 shape reweighting (21/04/2017,Kevin)
• Javier (14/03/2017): Table 2 and 3: I'm not sure if understand the trend in ST signal yields for b2j3 -> b2j4 and b3j3 -> b3j4 : we expect 3bs and hence ~3js for ST. However, ST decreases from b2j3 to b2j4 (OK...) but increases from b3j3 to b3j4 (?)... I would have expected both increasing since b2j4 includes jet>=4 multiplicity bins, not only j=4... Perhaps I'm missing something... In another words, why the probability to see an extra jet seems to depend on how many b-jets have been tagged?
  This trend looks indeed a bit counter-intuitive. We can try to make sense of it as follows. In 3-jet categories, the final state for this signal will most likely be made up of either the 3 b-quarks from the generator, or of 2 b-quarks from the generator and a light quark from PS. In 4-jet categories, the final state will be induced by the 3 b-quarks from the generator and 1 light quark from PS. Requiring exactly 2 b-tags in the 4-jet category, means we force one b-quark jet to be non-b-tagged, reducing the selection probability much more than if we'd require exactly 3 b-tags. Requiring exactly 2 b-tags in the 3-jet category, means we will relatively select more of the 2b+1light quark final state, whereas in the 3 b-tag scenario we will mainly select the 3b-quark final state. So the step of going from 3 jets to 4 jets has different underlying final states contributing if we require 2 or 3 b-tags. (24/03/2017, Kevin)
  
• Javier (24/03/2017): recently, inside the TMG we have seen several plots concerning the ttbar njets multiplicity which show a poor Data/MC agreement starting from njets=4 bin. Since your analysis is based on njet/nBquark categories, could you please share your plots for both distributions to see the real modelling, which might afecting your BDT sensitivity? Besides, your jet selection is pT>30GeV, and some analysis are seeing a better agreement with pT>20GeV, which points out to a possible issue in calibration, so I would be also interested (if possible in your FW) in plots where the jet has a different threshold.
  The number of jets distribution (after requiring at least 3 jets and 2 Medium b-tagged jets) can be found here. To investigate the effects of a lower pt-cut, I would have to remake the ntuples. (28/03/2017, Kevin)
• Benjamin (24/03/2017): Why don’t you drop the b4j4 category for the k_Hut limits? You don’t really gain anything in sensitivity from it, right? On the other hand in k_Hct it adds a bit, I’m guessing from b-tagging the c quark jet?
  The b4j4 category is left out of the Hut limit calculation from now on (21/04/2017,Kevin)
• Benjamin (24/03/2017): Do I understand correctly that the TT component of your signal is ttbar where exactly one of the tops decays to Hu or Hc. Then you float the signal strength of that and the ST component and interpret it as a branching fraction of t->Hu/Hc and then in turn as a coupling strenght? Do you float both ST and TT with a single signal strength parameter? Do both ST and TT scale linearly in k_Hut or k_Hct?
  Yes, this is indeed what the ttbar signal represents. In order to calculate the 1D limits (slides 3 and 4 from here), we fit the signal strength on the combination of the ttbar and single top component. Indeed, they both scale the same way, but the cross section is not scaled linearly, but quadratically to the coupling.

Textual comments

• Luca (14/03/2017): Line 10: “to be predicted” -> “predicted to be”?
• Luca (14/03/2017): Line 29: “the same for each of the two coupling” -> “the same for the two couplings”
• Luca (14/03/2017): Line 52: missing reference
• Luca (14/03/2017): Line 120: “identified a b jets” -> “identified as b jets”
• Luca (14/03/2017): Line 147: Fig 1 is Table 1.
• Luca (14/03/2017): Line 178: “[the limits are] shown as a function of the FCNC coupling strength (Fig 6) -> “compared to the prediction from theory as a function of […]”.
• Luca (14/03/2017): Line 12: there should not be also the list of institutes in the acknowledgements? https://twiki.cern.ch/twiki/bin/viewauth/CMS/Internal/PubAcknow
• Javier (14/03/2017): Line 61: m -> \mu
• Javier (14/03/2017): Line 68: missing "and"
• Javier (14/03/2017): Line 72-73: the sentence "In ttbar generation..." seems to be orphan, it should be connected with the use of Powheg for ttbar
• Javier (14/03/2017): Line 102: if you used HLT_Ele32_eta2p1_WPTight, as declared in the AN, the |eta| for electrons is 2.1 too
• Javier (14/03/2017): Line 106: how large is the cone radius used?
• Javier (14/03/2017): Line 120: a -> as b jets
• Javier (14/03/2017): Line 120: Simiarly -> Similarly
• Javier (14/03/2017): Line 147: Fig 1. -> Table 1
• Javier (14/03/2017):Lines 171 & 179: Tabs -> Tables?
• Javier (14/03/2017): Lines L269, L273, L276, L289, incomplete references?
• Javier (14/03/2017): Line 49-50: despite the TT_FCNC samples names do not reflect the algorithm used for matching, it should be quoted that MLM (?) has been used to take into account the additional partons
• Javier (14/03/2017): Line 137: Fig 2 does not show distributions for the reconstructed mtop
• Javier (14/03/2017): Figure 2: Top right plot, not sure if the x scale should be zoomed to the real range where the distribution has entries... Also, bMVA is shown in the bottom right plot, but not referenced in the text. Missing "the" in the caption, one of "the" reconstructed b jets....
• Benjamin (24/03/2017):L9: the Glashow-
  L23: *The analysis
  L25: on *the respective branching
  L31: in *the presented analysis.
  L50: simulated in *the initial hard
  L52: fix reference
  L65: "W boson" -> "W bosons"
  L68: "W boson" -> "W bosons"
  L68: (serial comma) "W, Z and Higgs" -> "W, Z, and Higgs"
  L70: "In signal and … processes simulations the initial" -> "In the
  simulations of signal and … processes, the initial"
  L76: “vertex distribution” -> “distribution of number of vertices”
  L102f: This is a bit confusing, as it is formulated now. I assume you want
  to write that you *select (not “reconstruct”) electron and muon candidates
  with |eta|<2.1 and pt > 35 (27) GeV.
  L112: present in *the event
  L115: provides *a b jet efficiency
  L115: at *a misidentification rate
  L116: “jets originated” -> “jets originating”
  L120: identified *as b jets
  L123: (serial comma): “ST, TT and ttbar” -> “ST, TT, and ttbar”
  L124: Rephrase this, I’d suggest: “The reconstruction is done for all
  possible permutations of the b jets to be associated with the decay
  products of the Higgs boson or the top quark.”
  L125: each *permutation
  L126: fed into *a multivariate … that uses *a Boosted
  L129: in *the event … *the permutation *with the highest
  L133: “type of FCNC coupling” is a bit unclear. Why not “For each jet
  category and separately for k_Hut and k_Hct couplings, a dedicated BDT is
  defined.”
  L139: as *the BDT score
  L144: *backgrounds
  L145: of *the exclusion limit. *A summary of
  L146: on *the predicted
  L147: “Fig.” -> “Tab.”
  Table 1 caption: Summary *of relative
  L148: from *the application of *the b tagging
  L149: and *the b tagging
  L150: with *these correction
  L151: namely, *the statistical
  L152: derived, *a systematic
  L152: (serial comma) “…as predicted by MC and jet…” -> “…as predicted by
  MC, and the jet…"
  L153: and *the jet energy scale
  L159: (serial comma) “Identification, isolation and trigger” ->
  “Identification, isolation, and trigger”
  L163ff: You already described this in L76. I would mention the number of
  69.2 mb there (“and the total inelastic cross section (69.2 mb).”), and
  then just write that an uncertainty of 5% on that leads to a negligible
  contribution to the overall systematic uncertainty.
  L167: You’ll need to introduce the top pt correction with a sentence about
  the disagreement found between data/MC.
  L171: *A comparison
  Tables 1, 2: Please add a row with total background or total
  background+signal. Something that makes it easy to compare to the data
  yield. I would also suggest to remove all of the vertical and most of the
  horizontal lines to make the tables look a bit cleaner.
  L184: *a Higgs boson
  L187: “…in the top quark decays. The observed (expected) …” -> “…in top
  quark decays. Observed (expected)…”
  Figures: Remove “Preliminary” before the final draft. Add “CMS” in Fig. 6.

ARC comments based on AN-15-097 v4

• Luca (14/03/2017): In Figs 5 to 20 you mention “fully reconstructed” and partially reconstructed” events, can you please explain what these categories are (sorry if I miss it in the text)?
  Fully reconstructed events are events where the longitudinal momentum of the neutrino has been reconstructed in the kinematic fit, whereas partially reconstructed events don't. (28/03/2017, Kevin).
  Is updated (06/04/2017, Kevin)
  
• Luca (14/03/2017): In Figs 13 to 20 you have signal and background, shouldn’t it be wrong and right assignment? This is the result of the BDT training for event reconstruction, if I am not wrong.
  Is updated (Kevin)
• Luca (14/03/2017): In Figs 21 to 24, which jet assignment is used in the top plots? The one which minimizes the likelihood used to reconstruct the neutrinos?
  The top plots use the likelihood based kinematic fit while the bottom plots show results when an additional BDT is used on top of this likelihood based kinematic reconstruction. (Kirill)
• Luca (14/03/2017): In Figs 21 to 25: maybe you can add some comments in the text on your final choice to use the “higher CSVv2 jets” strategy for b-tagging, as the comparison of the different strategies in Figs 21 to 24 is misleading.
  We have added the following sentence after L238: As a compromise between the amount of selected statistics and the quality of reconstructed kinematic variables, the b tagging selection that corresponds to CSVv2M requirement is used in the main analysis. (Kirill)
• Luca (14/03/2017): Line 372: the procedure used to derive limits from just one production mode is not clear to me: what do you mean by “the limits are accordingly scaled”?
  "The limits are accordingly scaled" means that the limits on the signal strength, calculated using combine tool, have been multiplied by the predicted cross section (ST+TT) for each of the production modes (Hut=50.82 and Hct=38.88). From now on however, we will provide 2D limit plots, deriving the limits for both production modes simultaneously. The 2D limit BR plot is going to be included eventually in the publication.
• Benedikt (14/03/2017): The pull plots in the AN (Fig. 44-54) need a better explanation and understanding. Some nuisances are pulled by more than 3 (!) sigma (Fig. 49), in Fig.52 the UE nuisance is pulled to 1sigma and has an uncertainty of 0. This sort of means that entire CMS could just get rid of the nominal TTbar sample and one of the TTbar samples with varied underlying event, because according to this pull plot we seem to know what the underlying event looks like (namely like in the other systematic sample that represents the +1sigma variation). I suspect there is significant tension in the fit due to assigned uncertainties that are not conservative enough or similar. You could compare B-only to S+B fit as a start. Alternatively, you could try to understand why Fig.45 (hut,b3j3) looks more pathological than Fig.51(same region but hct). I understand it's a different training, i.e. a different BDT output shape, but I wouldn't expect it to be different enough to cause problems for the one but not for the other.
  AND
• Benjamin (24/03/2017): Benedikt already commented on the pull plots, but just to re-iterate: I think we should try to understand these before we unblind. Did you check if the fit is converging properly?
  We understand your worries about the systematic uncertainties, that obviously behave strange in a way you stated, and therefore we are performing additional tests of the MLfit in order to investigate the cause. Besides the comparison of the B-only to S+B fit (slides 5-8 from here), we also check if the covariance matrices in the maximum likelihood fit converge. We are going to add additional rate uncertainties on the ttbb and ttcc background contributions separately (on top of the 5.5% ttbar cross section uncertainty). We will assign a conservative 50% uncertainty on these contributions.
• Benedikt (14/03/2017): Show the impact plot for the combined (i.e. all categories) fits to understand which are your dominating uncertainties.
  The plots this and this are showing the impacts (+/- 1 and 2 sigma) of various systematic uncertainties on our parameter of interest (r parameter of the maximum likelihood fit) for the case where we look into data, and this one and this one the impacts obtained without looking at the data (using the option '-t -1' of the multidim fit).
• Benedikt (14/03/2017): Add postfit plots to the AN, and ideally replace the prefit ones in the paper.Examples of the post-fit MVA output plots for each of the categories are shown in the right column on the slides 13-18 from here. The scale factors (slides 19-21 from here) extracted from the output of the performed ML fit can be used to derive the complete set of the post-fit plots from the pre-fit ones in the paper. These scale factors should be accounted in order to produce complete set of the post-fit plots.
• Benedikt (14/03/2017): Investigate whether the ttbar FxFx sample would cure the problem in the ttbar rate prediction.
  Seems like powheg gives a better description for our analysis. See: https://indico.cern.ch/event/625869/ (24/03/2017, Kevin)
  
  • Benedikt (25/03/2017): For some regions FxFx predicts much less ttbar events than powheg (b3j3) but then predicts much more in neighboring phase spaces (b3j4). I do not remember such a drastic behavior in any comparison I have seen between the two samples; if anything, I would think the jet multiplicity should be better modelled in the FxFx sample. I would therefore just ask you to crosscheck it once more (or show the nJet plots for different numbers of btags).
    The cross-check with number of Jets and b-tagged jets can be found here. The previous comparison didn't have the amc@nlo negative weights taken into account for the FxFx sample. This new comparison does however. In this comparison the FxFx sample doesn't seem to really improve the number of jets description wrt to the Powheg sample. For this and convenience reasons, we suggest to stick to the Powheg sample. (28/03/2017, Kevin)
  • Benedikt (25/03/2017): I am a bit skeptical also because of the second observation: I cannot reconcile the Powheg plots in the slides with the plots of Fig.29/30 in the ANv4. Can you say what has changed? The overprediction of Powheg, which was there in the Analysis Note, is gone in the slides.
    In this comparison a couple of things changed: The electron trigger changed and we have moved to using the February reprocessed data samples. The reweighting of the MC samples to the integrated luminosity has been kept closer track of as well. (28/03/2017, Kevin)
• Benedikt (14/03/2017): How does the assumption of SM-like top quark width in your samples relate to quoting limits on the decay t->hu/c? (This question has been answered nicely by one of the theorists in a separate thread, but I think it's nice to have it documented in the TWiki or on the hypernews).
  In the production and in the interpretation, we have made the assumption that the anomalous coupling is small and therefore its contribution to the top width could be neglected. First, this assumption is convenient: you produce only sample a coupling and you rescale the cross section. Otherwise, we must have to produce for one coupling several samples related to different values of the couplings. But this assumption is also realistic. The coupling values must be small, otherwise we would have discovered the FCNC. If you take the current experimental limits, the corresponding FCNC partial-width of the top is totally negligible. (Kirill)
• Javier (14/03/2017): Figures 13-20: The BDT shape for (specially) background and (also) signal in CSVv2L, CSVv2M and CSVv2T selections is quite weird, is this effect understood?
  Our understanding is that it happens because of strong correlations present between the input variables computed for several jet combinations per event. The same jet enters several times in the BDT per event. (Kirill)
• Javier (14/03/2017): Figures 21-24: it is hard to compare bottom plots w.r.t. top plots to see the improvement of the BDT. Suggestion: superimpose bars from top and bottom plots to appreciate the gain.
  The corresponding plots have been replaced with the ones with the superimposed efficiencies. (Kirill)
• Javier (14/03/2017): Line 247: since you are using Powheg, I would consider to check the systematic due to "matching" through the hdamp variation samples: https://twiki.cern.ch/twiki/bin/viewauth/CMS/TopSystematics#Matrix_element_Parton_Shower_ME
  The ME-PS matching has been added as uncertainty (06/04/2017, Kevin)
  
• Javier (14/03/2017): Line 76-77: nine categories are defined here, but only 5 are used at the end. Indeed 0b* and 1b* categories are not mentioned anywhere else in the text, are they really used? For what?
  The superfluous categories (b0 and b1) are left out. (24/03/2017, Kevin)
  
• Benjamin (24/03/2017): AN Table1: Why is ttbar not split into Hut and Hct? These numbers are just illustrative for a coupling strength of 1, right?
  It are indeed numbers for a coupling strength of 1. Therefore the ttbar signal cross section is the same for Hut and Hct. (28/03/2017, Kevin)
• Benjamin (24/03/2017): AN figures 7/9 and 8/10 are identical. Is that a bug or do the BDT inputs not depend on the Hut or Hct hypothesis?
  Thanks. The typo in figures naming is fixed and the correct plots are now visible in AN. (Kirill)
• Benjamin (24/03/2017):In the jet permutation selection BDT, why only these four input variables? Have you tried others? What settings for the BDT have you used, and have you tried others?
  These four variables are the most powerful ones to distinguish between the correct and the wrong combinations. Other kinematic variables are either strongly correlated with these four or add minor improvement to the final separation. We've tried other BDT options (e.g. BDT vs BDTG, different number of trees), but again the estimated change in the power of separation was minor. (Kirill)
• Benjamin (24/03/2017): AN figures 21–24 and 25: which of the jet assignment algorithms did you end up using? Maybe you write it somewhere, but I missed it. And the follow-up: how did you choose it?
  Yes, we agree. We have added the following sentence after L238: As a compromise between the amount of selected statistics and the quality of reconstructed kinematic variables, the b tagging selection that corresponds to CSVv2M requirement is used in the main analysis. (Kirill)

-- KevinDeroover - 2017-03-23

Event counts (PART 2)

A. Muon-channel

B. Electron channel

Event counts (PART 1)

A. Muon-channel

cutFlow step	object specifications	counts KIRILL	counts KEVIN	counts DESY	counts BELGRADE	counts HANYANG	event list KIRILL	event list KEVIN	event list DESY	event list BELGRADE	event list HANYANG
1) Initial	-		45 000		45 000	45 000
2) Good PV	vertex not fake; ndof > 4; z < 24; rho < 2		44 986		44 986
3) Trigger	HLT_Iso(Tk)Mu24_v2		5 919		5 919
4) Exactly one tight muon	Pt > 27 GeV; abs(eta) < 2.1; RelIso < 0.15 (DeltaR < 0.4); Tight Muon ID: SWGuideMuonIdRun2# Tight_MuonID		4 872		4 872
5) No loose electrons	Pt > 10 GeV; abs(eta) < 2.5; Loose Candidate Based Electron ID: CutBasedElectronIdentificationRun2#Working_points_for_2016_data_for		4 440		4 437
6) veto on extra loose muons	Pt > 10 GeV abs(eta) < 2.4; RelIso < 0.25(DeltaR < 0.4); Loose Muon ID: SWGuideMuonIdRun2# Loose_Muon		4 249		4 246
7) At least 3 Jets	Pt > 30 GeV abs(eta) < 2.4; Loose PFJetID: JetID#Recommendations_for_13_TeV_data; JEC: applied (Spring16_25nsV6_*_AK4PFchs)		3 076		3 071			http://mon.iihe.ac.be/~kderoove/FCNC_SynchExercise/EventInfo_mu.txt		http://test-cirkovic.web.cern.ch/test-cirkovic/09-11-2016/FCNC_sync_2/EventInfo_mu.txt

B. Electron channel

cutFlow step	object specifications	counts KIRILL	counts KEVIN	counts DESY	counts BELGRADE	counts HANYANG	event list KIRILL	event list KEVIN	event list DESY	event list BELGRADE	event list HANYANG
1) Initial	-		45 000		45 000	45 000
2) Good PV	vertex not fake; ndof > 4; z < 24; rho < 2		44 986		44 986
3) Trigger	HLT_Ele32_eta2p1_WPTight_Gsf_v3		4 180		4 180
4) Exactly one medium electron	Pt > 35 GeV; abs(eta) < 2.1; Medium Candidate Based Electron ID: CutBasedElectronIdentificationRun2#Working_points_for_2016_data_for		3 482		3 495
5) Veto on loose muons	Pt > 10 GeV; abs(eta) < 2.4; RelIso < 0.25 (DeltaR < 0.4); Loose Muon ID: SWGuideMuonIdRun2#Loose_Muon		3 117		3 121
6) veto on extra loose electrons	Pt > 10 GeV; abs(eta) < 2.5; Loose Candidate Based Electron ID: CutBasedElectronIdentificationRun2#Working_points_for_2016_data_for		2 980		2 988
7) At least 3 Jets	Pt > 30 GeV; abs(eta) < 2.4; Loose PFJetID: JetID#Recommendations_for_13_TeV_data; JEC: applied (Spring16_25nsV6_*_AK4PFchs)		2 177		2 190			http://mon.iihe.ac.be/~kderoove/FCNC_SynchExercise/EventInfo_El.txt		http://test-cirkovic.web.cern.ch/test-cirkovic/09-11-2016/FCNC_sync_2/EventInfo_el.txt

Spring15 Synchronization Iteration 1.0

For the initial part of this exercise, we will start by syncing on the preselected objects. The object definitions for muons, electrons, hadronic taus, jets and MET can be found on the main twiki. Note that the lepton effective areas have been updated for Spring15 -- these are given in the tables below.

• Sync sample: /store/mc/RunIISpring15MiniAODv2/ttHToNonbb_M125_13TeV_powheg_pythia8/MINIAODSIM/74X_mcRun2_asymptotic_v2-v1/10000/02FE2DB6-D06D-E511-8BC7-0025905C431C.root
• Release: CMSSW_7_4_12_patch4

Please upload tables separately for the different objects in the requested format below. Also, include the number of (unweighted) events passing the indicated selection.

Muons

• Require >= 1 muon passing preselection. Print quantities of leading muon.
• Format: event#, pt, eta, phi, E, pdgID, charge, miniIso, miniIsoCharged, miniIsoNeutral, jetPtRel, jetCSV, jetPtRatio, sip3D, dxy, dz, segmentCompatibility

	CERN	ND/Nebraska	IPHC	IHEP	TIFR
events	15686	15686	15684	15686
table	download	download	download	download

Electrons

• Require >= 1 electron passing preselection (including cleaning). Print quantities of leading electron.
• Format: event#, pt, eta, phi, E, pdgID, charge, miniIso, miniIsoCharged, miniIsoNeutral, jetPtRel, jetCSV, jetPtRatio, sip3D, dxy, dz, eleMVA

	CERN	ND/Nebraska	IPHC	IHEP	TIFR
events	15301	15301	15299	15798
table	download	download	download	download

Taus

• Require >= 1 tau passing selection (including cleaning). Print quantities of leading tau.
• Format: event#, pt, eta, phi, E, dxy, dz, decayModeFinding, byLooseCombinedIsolationDeltaBetaCorr3Hits

	CERN	ND/Nebraska	IPHC	IHEP	TIFR
events	4504	3379	4847	4844
table	download		download	download

Jets and MET

• Require >= 1 jet passing preselection (including cleaning). Print quantities of MET and leading jet.
• Format: event#, jet pt, jet eta, jet phi, jet E, jet CSV, MET pt, MET phi

	CERN	ND/Nebraska	IPHC	IHEP	TIFR
events	48154	48156	48157	48157
table	download	download	download	download

Installation instructions for CMSSW_7_4_12

Show GIT recipe Hide

cmsrel CMSSW_7_4_12_patch4
cd CMSSW_7_4_12_patch4/src
cmsenv

# create empty repository (with the cmssw trick to keep the repository small)
git cms-init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git
git remote add backup git@github.com:cirkovic/cmg-cmssw-backup.git

# configure the sparse checkout
cp /afs/cern.ch/user/c/cmgtools/public/sparse-checkout_7412_heppy .git/info/sparse-checkout

# checkout the CMGTools branch of the release, and push it to your CMG repository
DATE=`date +'%d-%m-%Y'`
git checkout -b CMGTools-from-CMSSW_7_4_12_$DATE cmg-central/CMGTools-from-CMSSW_7_4_12
git push -u origin CMGTools-from-CMSSW_7_4_12_$DATE

# create also the heppy branch
git branch heppy_74X cmg-central/heppy_74X

#compile
scram b -j 8

You can then merge or rebase the branch containing your existing developments on top of this release.

To get updates

git fetch cmg-central
git merge cmg-central/CMGTools-from-CMSSW_7_4_12

To make an update that requires changes also to Heppy (tidy version)
Note: this will temporarily make your CMGTools area disappear, so don't do it when running jobs
1) get on the heppy branch and make sure your heppy branch is updated

git checkout heppy_74X
git fetch cmg-central
git merge cmg-central/heppy_74X

2) create a branch for your development, commit your code there, and push on your repository

git checkout -b my-heppy-development
git add ...
git commit -m ...
git push origin my-heppy-development

3) make a pull request of your branch into the Heppy of the CMG repository
4) merge your heppy into your CMGTools branch

git checkout CMGTools-from-CMSSW_7_4_12 # now your heppy development disappears, but cmgtools reappears
git fetch cmg-central
git merge my-heppy-development  # now you have everything

5) push it to your repository (e.g. to branch my-cmg-development), and do pull request for the CMGTools branch. In the description of the PR, refer also the Heppy one of point (3). You just have to put "#" plus the PR number, and github will make a link for you.

git push origin CMGTools-from-CMSSW_7_4_12:my-cmg-development

To make an update that requires changes also to Heppy (for less tidy people)
Do your development on the CMGTools branch, but never include in the same commit CMGTools and Heppy changes, and always start the commit message of your heppy commits with Heppy (so that you can find them later)
When done, do the following (this will temporarily make your CMGTools area disappear, so don't do it when running jobs)
0) get the list of the ids of the heppy commits inside your branch, in chronological order starting from the oldest one
1) create a branch for putting your heppy commits inside, starting from the HEAD of the central branch

git fetch cmg-central
git checkout -b my-heppy-developments cmg-central/heppy_74X

2) cherry-pick all your heppy commits, in chronological order starting from the oldest one

git cherry-pick 

3) push that branch to github.
4) now get back into your CMGTools branch, and merge the heppy branch into it (it should do nothing)

git checkout CMGTools-from-CMSSW_7_4_12
git merge my-heppy-development  
git push origin CMGTools-from-CMSSW_7_4_12:my-cmg-development

5) if everything works correcly, push your CMG and Heppy branches to your github, and make pull requests for them

git push origin my-heppy-development
git push origin CMGTools-from-CMSSW_7_4_12:my-cmg-development

Installation instructions for CMSSW_7_4_7

Show GIT recipe Hide

cmsrel CMSSW_7_4_7
cd CMSSW_7_4_7/src
cmsenv

# create empty repository (with the cmssw trick to keep the repository small)
git cms-init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git
git remote add backup git@github.com:cirkovic/cmg-cmssw-backup.git

# configure the sparse checkout
cp /afs/cern.ch/user/c/cmgtools/public/sparse-checkout_74X_heppy .git/info/sparse-checkout

# checkout the CMGTools branch of the release, and push it to your CMG repository
DATE=`date +'%d-%m-%Y'`
git checkout -b CMGTools-from-CMSSW_7_4_7_$DATE cmg-central/CMGTools-from-CMSSW_7_4_7
#git push -u origin CMGTools-from-CMSSW_7_4_7_$DATE
#git push -u backup CMGTools-from-CMSSW_7_4_7_$DATE

# create also the heppy branch
git branch heppy_74X cmg-central/heppy_74X

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMGTools-from-CMSSW_7_4_7

Installation instructions for CMSSW_7_4_3

Show GIT recipe Hide

cmsrel CMSSW_7_4_3
cd CMSSW_7_4_3/src
cmsenv

# create empty repository
git cms-init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git
git remote add backup git@github.com:cirkovic/cmg-cmssw-backup.git

# configure the sparse checkout
cp /afs/cern.ch/user/c/cmgtools/public/sparse-checkout_74X_heppy .git/info/sparse-checkout

# checkout the CMGTools branch of the release, and push it to your CMG repository
DATE=`date +'%d-%m-%Y'`
git checkout -b CMGTools-from-CMSSW_7_4_3_$DATE cmg-central/CMGTools-from-CMSSW_7_4_3
#git push -u origin CMGTools-from-CMSSW_7_4_3_$DATE
#git push -u backup CMGTools-from-CMSSW_7_4_3_$DATE

# create also the heppy branch
git branch heppy_743 cmg-central/heppy_743

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMGTools-from-CMSSW_7_4_3

Installation instructions for CMSSW_7_4_0

Show GIT recipe Hide

cmsrel CMSSW_7_4_0
cd CMSSW_7_4_0/src
cmsenv

# create empty repository
git init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git
git remote add backup git@github.com:cirkovic/cmg-cmssw-backup.git

# configure the sparse checkout
git config core.sparsecheckout true
cp /afs/cern.ch/user/c/cmgtools/public/sparse-checkout_74X_heppy .git/info/sparse-checkout

# checkout the CMGTools branch of the release, and push it to your CMG repository
DATE=`date +'%d-%m-%Y'`
git checkout -b CMGTools-from-CMSSW_7_4_0_$DATE cmg-central/CMGTools-from-CMSSW_7_4_0
#git push -u origin CMGTools-from-CMSSW_7_4_0_$DATE
#git push -u backup CMGTools-from-CMSSW_7_4_0_$DATE

# create also the heppy branch
git branch heppy_740 cmg-central/heppy_740

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMGTools-from-CMSSW_7_4_0

Installation instructions for CMSSW_7_2_3

Show GIT recipe Hide

cmsrel CMSSW_7_2_3
cd CMSSW_7_2_3/src
cmsenv

# create empty repository
git init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git
git remote add backup git@github.com:cirkovic/cmg-cmssw-backup.git

# configure the sparse checkout
git config core.sparsecheckout true
cp /afs/cern.ch/user/c/cmgtools/public/sparse-checkout_72X_heppy .git/info/sparse-checkout

# checkout the CMGTools branch of the release, and push it to your CMG repository
DATE=`date +'%d-%m-%Y'`
git checkout -b CMGTools-from-CMSSW_7_2_3_$DATE cmg-central/CMGTools-from-CMSSW_7_2_3
#git push -u origin CMGTools-from-CMSSW_7_2_3_$DATE

# create also the heppy branch
git branch heppy_7_2_2_patch2 cmg-central/heppy_7_2_2_patch2

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMGTools-from-CMSSW_7_2_3

Installation instructions for CMSSW_7_0_6_patch1

Show recipe Hide

cmsrel CMSSW_7_0_6_patch1
cd CMSSW_7_0_6_patch1/src
cmsenv

# create empty repository
git init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git

# configure the sparse checkout
git config core.sparsecheckout true
cp /afs/cern.ch/user/c/cmgtools/public/sparse-checkout_706_miniAOD_lite .git/info/sparse-checkout

# checkout the release, make a branch for it, and push it to your CMG repository
git checkout cmg-central/CMG_MiniAOD_Lite_V6_0_from-CMSSW_7_0_6
CURRDATE=`date +'%d-%m-%Y'`
git checkout -b CMG_MiniAOD_Lite_V6_0_from-CMSSW_7_0_6_$CURRDATE
git push -u origin CMG_MiniAOD_Lite_V6_0_from-CMSSW_7_0_6_$CURRDATE

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMG_MiniAOD_Lite_V6_0_from-CMSSW_7_0_6

Installation instructions for CMSSW_5_3_19

Show recipe Hide

cmsrel CMSSW_5_3_19
cd CMSSW_5_3_19/src
cmsenv

# create empty repository
git init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git

# configure the sparse checkout
git config core.sparsecheckout true
cp /afs/cern.ch/user/g/gpetrucc/public/CMG_PAT_V5_18_from-CMSSW_5_3_14.sparse-checkout .git/info/sparse-checkout

# checkout the release, make a branch for it, and push it to your CMG repository
git checkout cmg-central/CMG_PAT_V5_18_from-CMSSW_5_3_14
CURRDATE=`date +'%d-%m-%Y'`
git checkout -b CMG_PAT_V5_18_from-CMSSW_5_3_14_$CURRDATE
git push -u origin CMG_PAT_V5_18_from-CMSSW_5_3_14_$CURRDATE

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMG_PAT_V5_18_from-CMSSW_5_3_14

Installation instructions for CMSSW_5_3_14

Show recipe Hide

cmsrel CMSSW_5_3_14
cd CMSSW_5_3_14/src
cmsenv

# create empty repository
git init

# add the central CMG repository, and fetch it
git remote add cmg-central https://github.com/CERN-PH-CMG/cmg-cmssw.git
git fetch cmg-central

# add your mirror (see https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsGitMigration#Prerequisites )
git remote add origin git@github.com:cirkovic/cmg-cmssw.git

# configure the sparse checkout
git config core.sparsecheckout true
cp /afs/cern.ch/user/g/gpetrucc/public/CMG_PAT_V5_18_from-CMSSW_5_3_14.sparse-checkout .git/info/sparse-checkout

# checkout the release, make a branch for it, and push it to your CMG repository
git checkout cmg-central/CMG_PAT_V5_18_from-CMSSW_5_3_14
CURRDATE=`date +'%d-%m-%Y'`
git checkout -b CMG_PAT_V5_18_from-CMSSW_5_3_14_$CURRDATE
git push -u origin CMG_PAT_V5_18_from-CMSSW_5_3_14_$CURRDATE

#compile
scram b -j 8

git fetch cmg-central
git merge cmg-central/CMG_PAT_V5_18_from-CMSSW_5_3_14

Additional steps:

Show recipe Hide

cd CMGTools/TTHAnalysis/python/plotter
root.exe -b -l -q smearer.cc++ mcCorrections.cc++
root.exe -b -l -q functions.cc++
root.exe -b -l -q fakeRate.cc++
echo '{ gSystem->SetIncludePath("-I$ROOFITSYS/include"); gROOT->ProcessLine(".L TH1Keys.cc++"); }' > rootScript.C
root.exe -b -l -q rootScript.C
rm rootScript.C
chmod 644 *_cc.d # mcCorrections_cc.d

Useful commands

Open the file from the last revision:

Show recipe Hide

git show REVISION:FILE | vim - "+set filetype=cpp"

git show HEAD~:CMGTools/TTHAnalysis/macros/finalMVA/2lss/trainMVA_2lss.cxx | vim - "+set filetype=cpp"

find . | xargs grep 'string' -sl

find $CMSSW_BASE -name "*.py" | xargs grep 'string' -sil

find -type f -exec grep -H 'string' {} +

eval $(sed -n -e 1p run_script.sh)

source /afs/cern.ch/user/g/gpetrucc/sh/init_root528_afs

CURRDIR=`pwd`; for i in `find . -name "*.root"`; do cd `dirname $i`; root -l '/afs/cern.ch/sw/lcg/app/releases/ROOT/5.28.00g/x86_64-slc5-gcc43-opt/root/tmva/test/TMVAGui.C("'`basename $i`'")'; cd $CURRDIR; done

for i in `find . -name "plots"`; do cp index.php $i; done

ps aux | grep python | grep -v "grep python" | awk '{print $2}' | xargs kill -9

Links:

https://resources.web.cern.ch/resources/Manage/AFS/Settings.aspx

https://twiki.cern.ch/twiki/bin/viewauth/CMS/SUSYCMGfwk

https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGTTHPythonAnalysisCode

https://twiki.cern.ch/twiki/bin/viewauth/CMS/SWGuideHiggsAnalysisCombinedLimit

https://twiki.cern.ch/twiki/bin/viewauth/CMS/TTHmultileptonsFor13TeV

https://twiki.cern.ch/twiki/bin/view/CMS/TTH-Multileptons

CMS DAS (Data Aggregation System)

https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookMiniAOD

https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookGenParticleCandidate

https://twiki.cern.ch/twiki/bin/view/Main/PdgId

http://pdg.lbl.gov/2005/reviews/montecarlorpp.pdf

https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuidePythonAnalysis

https://twiki.cern.ch/twiki/bin/viewauth/CMS/CMGToolsReleasesExperimental

https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideHeppy

https://twiki.cern.ch/twiki/bin/view/CMS/OnlineWBCentralDCSInstructions

https://twiki.cern.ch/twiki/pub/CMS/OnlineWBCentralDCSInstructions/centralDCStutorial.pdf

https://twiki.cern.ch/twiki/bin/viewauth/CMS/EcalPFGShift

http://test-cirkovic.web.cern.ch/test-cirkovic/01-06-2015/ttHCombinatorics/without_bb_gg/OUTPUT.txt

Related to ATLAS

My Links (TWiki, Tutorials, Tools, Templates)

• The ATLAS Computing Workbook (pdf v16.6.3)
• Physics Analysis Workbook - Release 17
  • Introduction
• ROOT (Users Guide pdf v5.26)
• ROOT grid analysis
• Mini ROOT Tutorial
• GANGA (Introduction pdf)
• PATHENA
• How to submit ROOT/general jobs to Panda
• PANDA Monitor, (my jobs)
• Bookkeeping for Panda analysis jobs
• ATLAS Storage at CERN
  • EOS Manual, V 0.1.0 (EOS - CERN Disk Storage)
• ATLAS Flavour Tagging Working Group (TODO)
• Shifts
  • Comp@P1 Shift twiki
    • EDH Access Request (ACRQ)
  • General Data Quality Shifts
    • Offline Shifter's Manual
    • Online Shifter's Manual
      • EDH Access Request (ACRQ)
• Offline Flavour Tagging DQ shifter's Manual (B-tagging DQ shifts)
  • List of reprocessed runs (10/24/2012)
  • List of reprocessed runs (11/28/2012)
• IT Department
• BTaggingJetTagNtuple documentation
• JetTagD3PD documentation
• The Calibration for EPS 2011
• Analysis17: Supported Taggers and Operating Points; Data/MC Scale Factors
• Map Between JetTagNtuple and JetTagD3PD Names
• D3PDContent (PDF)
• JetTag class documentation
• LaTeX template for ATLAS Notes

ATLAS Links

• ATLAS Experiment site
• ATLAS Collaboration site
  • ATLAS Experiment - Public Results
• ATLAS Grey Book
• European Organization for Nuclear Research - CERN

Conferences

• BsJpsiphi and B-flavour charge tagging Workshop
• Physics and Performance week
• ATLAS offline software tutorial (3-7 December 2012 CERN)
  • Timetable
• Trans-European School of High Energy Physics (July 13th-20th, 2012, Petnica Science Center, Serbia)
• AIDA Student Tutorial - Solid State Detectors (March 27 2012, DESY, Hamburg)
• Informal B-Tagging Meeting (October 20 2011, EVO only)
• ATLAS Offline Software tutorial (December 14-17 2010, CERN), (presentation pdf)

My Personal Preferences

• LxplusProblem
• ATLAS OTP
• CERN Mail Services
  • Outlook Web Access (Webmail)
• CERN WEB Services
  • http://cern.ch/cirkovic
  • http://cern.ch/test-cirkovic
• Account Management
• CERN Central SVN Service
  • CERN SVN
  • SVN Trac
• The Glance Project
• CERN Phonebook
• ATLAS webcams
  • ATLAS Control Room (ACR)
• Shifters at the ACR
• LHC Announcer
• OP Webtools
• inSPIRE HEP (High-Energy Physics Literature Database)
  • My account
    • Publications
• CERN Shuttle Service
• Map of buldings at CERN

Physics Links

• My task
  • MC request
  • TaskList
• Particle Data Group (PDG)
• ATLAS B-Physics Working Group
  • BJpsi sub-group page
    • Page of Bd Jpsi K0* and BsJpsi phi physics
      • Searching for New Physics in Bs J/psi phi at the ATLAS experiment (pdf)
      • Measurement of the angular and lifetime parameters of the decays $B_{d}^{0}\to J/ψK^{*0}$ and $B_{s}^{0}\to J/ψφ$
      • Flavor symmetry for strong phases and determination of $β_s, ΔΓ$ in $B_s \to J/ψφ$
      • Measurement that uses only one angle to extract one helicity amplitude (LHCb)
    • B Physics Flavour Tagging
      • BsJpsiphi and B-flavour charge tagging Workshop
      • BsJpsiPhi SVN
      • BPhysWorkingGroupAnalysisAndVertexing
        • BPhysWorkingGroupStatTools
        • Skimmed di-muon AOD (DAOD_ONIAMUMU)
      • MLfit fitter code (SVN)

BTaggingBenchmarks

• Benchmarks/Workingpoints
• Scale Factors

BTagging

• BTagging Tutorial For Release 17
  • b-tagging code (JetTagging package on CVS)
• Track Jets In Athena
• b-Jet substructure
• Boosted BTagging To Do

Related Topics

• ChangePassword for changing your password
• ChangeEmailAddress for changing your email address
• UserList has a list of other TWiki users
• UserDocumentationCategory is a list of TWiki user documentation
• UserToolsCategory lists all TWiki user tools
• OrganisationName: CERN
• OrganisationURL: http://www.cern.ch
• Comment: