you write:
> As you say, this is pretty simple to implement.
> Question; Do you want asymmetric errors in both X and Y or only along Y?
in both x and y.
btw: there is a *little* problem: having a look at
TGraphErrors.cxx, there is a comment in GetX() rsp. GetY():
Float_t TGraphErrors::GetErrorY(Int_t i)
{
// This function is called by GraphFitChisquare.
// It returns the error along Y at point i.
i overlooked that before my 1st posting.
certainly values in a TGraphAsymmErrors context
can hardly be fitted by a simple, general chi2-fit.
probably one could simply type
# include <math>
Float_t TGraphAsymmErrors::GetErrorY(Int_t i)
{
// This function is called by GraphFitChisquare.
// It returns the error along Y at point i.
if (i < 0 || i >= fNpoints) return -1;
if (fEYLow && fEYHigh)
return sqrt (fEYLow[i]*fEYLow[i]+fEYHigh[i]*fEYHigh[i]);
else
return -1;
}
and keep in mind that the results of a chi-2 fit cannot be taken
too literally.
j.
-------------------------------------------------------------------
Dr. J. Rieling Tel: +49 421 539 5643
DaimlerChrysler Aerospace eMail: joerg.rieling@ri.dasa.de
Bremen / Germany PGP: www.desy.de/~joerg/mypubkey.html
-------------------------------------------------------------------