I am trying to fit a crystal ball function to Z peak. The fit finishes successfully and fit result printout (RooFitResult::Print()) has reasonable error values. However, if you look at the table for the last fitting step (HESSE), error is much larger than expected. At the same time, the parameters themselves have the same values as the results.

** 11 **HESSE 2000

COVARIANCE MATRIX CALCULATED SUCCESSFULLY

FCN=8.72292e-05 FROM HESSE STATUS=OK 31 CALLS 232 TOTAL

EDM=1.34832e-08 STRATEGY= 1 ERROR MATRIX ACCURATE

EXT PARAMETER INTERNAL INTERNAL

NO. NAME VALUE ERROR STEP SIZE VALUE

1 cb_alpha -1.23559e+00 **1.13083e+01** 7.08829e-04 -1.23876e-01

2 cb_mass 9.07937e+01 **1.52140e+01** 1.94097e-03 7.94558e-02

3 cb_power 1.30459e+00 **8.42296e+00** 3.54643e-03 -8.31706e-01

4 cb_sigma 2.69402e+00 **9.69808e+01** 1.18205e-03 -1.24710e+00

ERR DEF= 0.5

EXTERNAL ERROR MATRIX. NDIM= 25 NPAR= 4 ERR DEF=0.5

4.741e+05 -2.869e+05 5.965e+05 -3.063e+05

-2.869e+05 6.945e+05 -3.088e+05 3.332e+05

5.965e+05 -3.088e+05 9.262e+05 -2.936e+05

-3.063e+05 3.332e+05 -2.936e+05 7.386e+05

PARAMETER CORRELATION COEFFICIENTS

NO. GLOBAL 1 2 3 4

1 0.92946 1.000 -0.500 0.900 -0.518

2 0.56102 -0.500 1.000 -0.385 0.465

3 0.91046 0.900 -0.385 1.000 -0.355

4 0.60990 -0.518 0.465 -0.355 1.000

[#1] INFO:Minization – RooMinimizer::optimizeConst: deactivating const optimization`

RooFitResult: minimized FCN value: 8.72292e-05, estimated distance to minimum: 1.34832e-08

covariance matrix quality: Unknown, matrix was externally provided

Status : MINIMIZE=0 HESSE=0 HESSE=0

```
Floating Parameter FinalValue +/- Error
```

```
cb_alpha -1.2356e+00 +/- 2.21e-02
cb_mass 9.0794e+01 +/- 2.53e-02
cb_power 1.3046e+00 +/- 3.11e-02
cb_sigma 2.6940e+00 +/- 2.60e-02
```