Fitting with a mirrored landau

[vrbruening]

Hi all,

I’m trying to fit a reflected landau onto some output from a Geant4 simulation. My issue is that when I try to reflect the fit by setting the scale parameter to -1, the fit disappears entirely (but can be seen as a black line on the x-axis).

When the scale parameter is set to 1, a landau distribution fits over the dataset as expected. I have included my code below.

Thank-you,
Vincent

{
  gROOT->Reset();
  gROOT->SetStyle("Plain");
 
 TFile f("B4.root");
 
TCanvas* c1 = new TCanvas("c1", "", 20, 20, 1000, 1000);
 c1->Divide(1,1);
 c1->cd(1);
 TTree *B4 = (TTree*)f.Get("B4");

 TF1 f1("f1","[0]*TMath::Landau(x,[1],[2])",0,530);
 f1.SetParameters(1.59,480,0.3);
 B4->Draw("Ecore>>h1(50,300,530)");
 Ecore->Fit("f1");

_ROOT Version: 6.26
Platform: Fedora Linux
Compiler: GNU

[couet]

Welcome to the ROOT forum.

@vrbruening

by setting the scale parameter to -1

I do not see such scaling in your code. Your code does not seem complete.

Also, note you should not use gROOT->Reset() that’s not recommended anymore.

[Wile_E_Coyote]

If by the “scale parameter” you mean the parameter “[2]”, then it MUST be a positive value.

You can try a “mirrored” version (“long tail” on the left side):
TF1 f1("f1", "[0] * TMath::Landau([1] - x, 0., [2])", 300., 530.);

[vrbruening]

Hi couet,

My aplogies for the incomplete code, I accidentally copied the wrong version of the code.

My issue is now resolved, but I was wondering what is recommended instead of

gROOT->Reset()

When I remove the line of code, my mean changes, and the bins appear as crosses rather than rectangles.

Thank-you,
Vincent

[couet]

@vrbruening

My issue is now resolved,

good;

@vrbruening

My issue is now resolved, but I was wondering what is recommended instead of

gROOT->Reset()

nothing