Skype discussion with Ximo

[ytatus94]

[24.05.17, 3:51:13 PM] Yu-Ting: Hi Ximo, I want to ask a question.

[24.05.17, 3:52:23 PM] Ximo: go ahead

[24.05.17, 3:53:14 PM] Yu-Ting: As you know the NUHM2 is a complex model, it contains several final states. We applied 1L filter in the jobOption of the NUHM2 weak MC production. But the filter efficiency didn't include in the SUSYTools. I found the filter efficiency from AMI but it is mean value. i.e. AMI only gives me one value for the filter efficiency for each m12 mass point. [24.05.17, 3:54:08 PM] Yu-Ting: I think the filter efficiency should be different for different final state in one m12 mass point. But I am not sure. Do you think the filter efficiency is different between final states?

[24.05.17, 3:54:45 PM] Ximo: what would matter is if it is different for different production processes hmmm... I am not sure

[24.05.17, 3:55:58 PM] Yu-Ting: I think the filter efficiency should be different between FS. But the value should be closed to the mean value which AMI gives me.

[24.05.17, 3:58:27 PM] Ximo: I mena, between final states it is for sure that is different, but that would not affect how you have to normalize the MC

[24.05.17, 3:59:57 PM] Yu-Ting: Why? When we normalize MC, we should consider the filterEff. I don't understand why you said it would not affect normalization.

[24.05.17, 4:01:14 PM] Ximo: let's see, if you have a single production process, and have the filter efficiency, you simply use it, do we agree on that?

[24.05.17, 4:01:30 PM] Yu-Ting: yes

[24.05.17, 4:02:30 PM] Ximo: then, if you have two process (A and B), and they have a k-factor to scale from LO to NLO that is different for A and B, but we only have a single filter eff from AMI

[24.05.17, 4:04:15 PM] Yu-Ting: yes AMI only provides one value for the filter eff

[24.05.17, 4:05:00 PM] Ximo: well, I think you should probably close your eyes on that one, and simply use the filter eff like if it were a single process

[24.05.17, 4:07:13 PM] Yu-Ting: ok, although I don't quite understand, I will use the filter efficiency provided by AMI for different final states.

[24.05.17, 4:08:44 PM] Ximo: but I don't see why you are worried about final states if your selection is tighter that the filter, that has zero practical implications

[24.05.17, 4:10:33 PM] Yu-Ting: Yes, we required pT>7 GeV which is very loose requirement

[24.05.17, 4:11:04 PM] Ximo: ah, hold on: if your selection is looser than the trigger, then that is a problem and you cannot use that filtered sample

[24.05.17, 4:11:36 PM] Yu-Ting: Why?

[24.05.17, 4:11:53 PM] Ximo: because you would be missing events

[24.05.17, 4:12:04 PM] Yu-Ting: if the selection is looser than the trigger, the event cannot be triggered.

[24.05.17, 4:14:36 PM] Ximo: I am not talking about trigger it is about your event selection

[24.05.17, 4:18:10 PM] Yu-Ting: let me think about it.

[ytatus94]

• 20170531

[31.05.17, 4:48:24 PM] Yu-Ting: Hi Ximo, I have some root questions

h_yields = new TH1F("h_yields", " in SR", 5, 0, 5);
if (sample_type == "data") {
            h_yields->Fill(1);
        }
        else if (sample_type == "signals") {
            h_yields->Fill(2);
        }
        else if (sample_type == "backgrounds") {
            h_yields->Fill(3);
        }

When I try to get the BinContents, I notice the h_yields->GetBinContent(2)=0 for the signals and h_yields->GetBinContent(3)=0 for the backgrounds. But h_yields->GetBinContent(3) for the signals and h_yields->GetBinContent(4) for the backgrounds have values. I don't know why... ah~ I know the reason why now.

[31.05.17, 5:48:49 PM] Ximo: sorry I couldn't reply...

[ytatus94]

• 20170606

[06.06.17, 4:43:19 PM] Yu-Ting: Hi Ximo, I want to ask how to set the bottom margin of a histogram. I use SetTitleOffset() but the title is cut by the edge of the canvas. I google it but most of the results are teaching how to set left or right margin

[06.06.17, 4:44:28 PM] Ximo: you want the distance of the title to the axis, or of the axis to the bottom of the canvas?

[06.06.17, 4:44:45 PM] Yu-Ting: the axis to the bottom of the canvas?

[06.06.17, 4:45:12 PM] Ximo: try: gPad->SetBottomMargin(0.13);

[06.06.17, 4:45:38 PM] Yu-Ting: So that is the property of gPad not histogram thank you, i will try it

[06.06.17, 4:46:07 PM] Ximo: hope it works!

[06.06.17, 4:47:36 PM] Yu-Ting: yes, it works! Thank you very much

[06.06.17, 4:47:43 PM] Ximo: :)

[06.06.17, 4:52:26 PM] Yu-Ting: Another stupid question, you gave code snippet for making significant plot several weeks ago. Do you use TH2F or TGraph2D to make the plot? I tried, and I only can use TH2F to make plot but the the contour didn't show on the significant plot.

[06.06.17, 4:53:27 PM] Ximo: the workflow is like TH1 -> TGraph2D -> TH1 -> draw contour

[06.06.17, 4:54:17 PM] Yu-Ting: Yes, so you use hFinalSig and Ghisto to make significance and contour I can make hFinalSig plot but when I overlay Ghisto on it, no contour shows. For testing purpose, I filled hFinalSig using the same values the same as you did in your slides.

[ytatus94]

• 20170609

[09.06.17, 3:12:12 PM] Yu-Ting: Hi Ximo, because I couldn't get the TGraph2D works correctly, I try to type in the following codes

  1 {
  2 TGraph2D *g = new TGraph2D();
  3 g->SetPoint(0,305.0, 0.25,20);
  4 g->SetPoint(1,355.0, 0.25,20);
  5 g->SetPoint(2,405.0, 0.25,15.66);
  6 g->SetPoint(3,505.0, 0.25,6.98);
  7 g->SetPoint(4,605.0, 0.25,2.48);
  8 g->SetPoint(5,705.0, 0.25,0.01);
  9 g->SetPoint(6,805.0, 0.25,0.01);
 10 g->Draw("surf1");
 11 }

But the root doesn't show the plot. It only pop-up an X-window without anything. also root doesn't return the prompt. Do you know what is the possible reason? If the TGraph2D doesn't work correctly, then I couldn't get the 3sigma and 1.6sigma lines.

[09.06.17, 3:16:15 PM] Ximo: should we sit down and look into this together?

[09.06.17, 3:16:43 PM] Yu-Ting: sure, when? I have a meeting at 3:30 and I think it can be finish about 4pm

[09.06.17, 3:17:19 PM] Ximo: ok, let's meet at 4

[09.06.17, 3:17:37 PM] Yu-Ting: at b40 lobby?

[09.06.17, 3:19:50 PM] Ximo: yes

[09.06.17, 3:20:29 PM] Yu-Ting: thank you

[09.06.17, 4:31:01 PM] Ximo:

{
  TGraph2D *g = new TGraph2D();
  int np=0;
  g->SetPoint(np++,305.0, 0.25, 20);
  g->SetPoint(np++,355.0, 0.25, 20);
  g->SetPoint(np++,405.0, 0.25, 15.66);
  g->SetPoint(np++,505.0, 0.25, 6.98);
  g->SetPoint(np++,605.0, 0.25, 2.48);
  g->SetPoint(np++,705.0, 0.25, 0.01);
  g->SetPoint(np++,805.0, 0.25, 0.01);

  g->SetPoint(np++,305.0, 10.25, 20);
  g->SetPoint(np++,355.0, 10.25, 20);
  g->SetPoint(np++,405.0, 10.25, 15.66);
  g->SetPoint(np++,505.0, 10.25, 6.98);
  g->SetPoint(np++,605.0, 10.25, 2.48);
  g->SetPoint(np++,705.0, 10.25, 0.01);
  g->SetPoint(np++,805.0, 10.25, 0.01);

  g->SetPoint(np++,305.0, -10.25, 20);
  g->SetPoint(np++,355.0, -10.25, 20);
  g->SetPoint(np++,405.0, -10.25, 15.66);
  g->SetPoint(np++,505.0, -10.25, 6.98);
  g->SetPoint(np++,605.0, -10.25, 2.48);
  g->SetPoint(np++,705.0, -10.25, 0.01);
  g->SetPoint(np++,805.0, -10.25, 0.01);

  g->Draw("colz");

  TH2F* Ghisto=(TH2F*)g->GetHistogram();
  Double_t level[2]={1.64,3};
  Ghisto->SetContour(2,level);
  Ghisto->Draw("cont1,list ,same");

}

[ytatus94]

• 20170619

[19.06.17, 10:59:15 AM] Yu-Ting: Hi Ximo, I want to ask how to set the histogram title of a THStack. I tried

hs->GetHistogram()->SetTitle(var)

but it doesn't work.

[19.06.17, 10:59:53 AM] Ximo: hs->SetTitle(var) is not working?

[19.06.17, 11:00:41 AM] Yu-Ting: oh~ it works. Thank you!

[19.06.17, 11:00:52 AM] Ximo: :)

[19.06.17, 11:01:19 AM] Yu-Ting: I found some website and they all say about the cosmetic need to GetHistogram() first...

[ytatus94]

• 20170706

[06.07.17, 12:44:09 PM] Yu-Ting: Hi Ximo, I want to ask a question. Do you know which analysis need to use the electron isolation efficiencies or scale factors low pT < 25 GeV or high pT > 150 GeV? I am writing the annual department of energy report and want to explain why I had to extend the pT when I worked on isolation efficiency and SF.

[06.07.17, 12:44:49 PM] Ximo: high pT: many exotics analysis, such as W' and Z' low pT: for instance our SS/3L analysis, and many searches for compresses SUSY scenarios (some of them go as low as 4.5 GeV)

[06.07.17, 12:46:11 PM] Yu-Ting: yes, i knew the lower bound of pT is 4.5 GeV now. Thank you very much!

[06.07.17, 12:47:50 PM] Ximo: np!