Chi2 in Fit to TGraph

[Mustaphaa]

Dear Experts,

I’m trying to fit my distribution with a non-built-in ROOT function, but I have noticed that the parameter values I set are too large and the Chi-squared value I obtained is too small. I’m unsure if this makes sense or not. Could you please provide me with a solution for this? Here is my macro :
Fit_Resolution.cc (979 Bytes)

[couet]

Using errors bars on Y helps.

double Polynomial(double *x, double *p) {
    double pt = x[0];
    double r1 = p[0];
    double r2 = p[1];
    double r3 = p[2];
    return ((r1 / pt) + r2 + r3 * pt);
}

void Fit_Resolution() {
   double pt[]      = {20, 40, 60, 80, 100, 120, 150};
   double sigma_p[] = {0.0213361, 0.0225363, 0.0250548, 0.0280364, 0.0310096, 0.0295183, 0.0400554};
   double ptErr[7]  = {0};
   double sigma_pErr[7] = {0.001, 0.001, 0.001, 0.001, 0.001, 0.001, 0.001};

   auto graph = new TGraphErrors(7, pt, sigma_p, ptErr, sigma_pErr);

   auto fitDSCB = new TF1("fitDSCB", Polynomial, 20, 160, 3);
   graph->Fit(fitDSCB, "S");

   graph->SetMarkerStyle(20);
   graph->SetMarkerColor(kBlue);

   graph->Draw("AP");
   fitDSCB->Draw("same");
}

Processing Fit_Resolution.cc...
****************************************
Minimizer is Minuit2 / Migrad
Chi2                      =      20.2449
NDf                       =            4
Edm                       =   1.2801e-19
NCalls                    =           62
p0                        =    0.0958795   +/-   0.0501049   
p1                        =    0.0135572   +/-   0.00223064  
p2                        =  0.000158354   +/-   1.67148e-05 

[Mustaphaa]

Thanks a lot,

Best regards,
Mustapha