Dear experts,
I have a general question about the EDM value of the roofit.
Is it necessary to be less than <1e-6 to indicate a good fit? I know it is used to control the stop of the minimization. By default setting, I usually get an EDM value like 1e-4 ~1e-5, do I need to change the tolerance?
e.g I am performing a UML fit, the output is like following and the fit likes fine.
the output from the fitting
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=-2.48097e+07 FROM HESSE STATUS=OK 40 CALLS 324 TOTAL
EDM=0.000311126 STRATEGY= 1 ERROR MATRIX ACCURATE
EXT PARAMETER INTERNAL INTERNAL
NO. NAME VALUE ERROR STEP SIZE VALUE
1 lambda_Bu -3.21855e-03 3.08810e-05 7.78062e-07 -6.43711e-04
2 mean_Bu 5.27866e+03 3.47285e-02 2.52633e-04 1.55008e-01
3 nbkg 2.85245e+05 8.92294e+02 1.25842e-04 -9.24946e-01
4 nsig 2.45351e+06 1.72176e+03 2.30789e-04 8.21959e-01
5 sigmaL_Bu 1.69468e+01 2.71322e-02 2.09937e-04 -3.10274e-01
6 sigmaR_Bu 1.82945e+01 2.55485e-02 1.88080e-04 -1.71392e-01
ERR DEF= 0.5
EXTERNAL ERROR MATRIX. NDIM= 25 NPAR= 6 ERR DEF=0.5
9.536e-10 3.269e-07 -1.098e-03 1.099e-03 2.833e-07 -2.605e-07
3.269e-07 1.206e-03 -3.079e+00 3.078e+00 7.430e-04 -7.202e-04
-1.098e-03 -3.079e+00 7.962e+05 -5.109e+05 -7.400e+00 -1.943e+00
1.099e-03 3.078e+00 -5.109e+05 2.964e+06 7.400e+00 1.943e+00
2.833e-07 7.430e-04 -7.400e+00 7.400e+00 7.362e-04 -3.593e-04
-2.605e-07 -7.202e-04 -1.943e+00 1.943e+00 -3.593e-04 6.527e-04
PARAMETER CORRELATION COEFFICIENTS
NO. GLOBAL 1 2 3 4 5 6
1 0.40200 1.000 0.305 -0.040 0.021 0.338 -0.330
2 0.92004 0.305 1.000 -0.099 0.051 0.788 -0.812
3 0.48637 -0.040 -0.099 1.000 -0.333 -0.306 -0.085
4 0.34337 0.021 0.051 -0.333 1.000 0.158 0.044
5 0.84451 0.338 0.788 -0.306 0.158 1.000 -0.518
6 0.84970 -0.330 -0.812 -0.085 0.044 -0.518 1.000
Best,
Quan