What's difference between SUCCESSFUL and CONVERGED statuses

Sometimes I see the result status as MIGRAD ... CONVERGED:

FCN=56.902 FROM MIGRAD STATUS=CONVERGED 86 CALLS 87 TOTAL
EDM=1.03875e-08 STRATEGY= 1 ERROR MATRIX ACCURATE 
EXT PARAMETER STEP FIRST 
NO. NAME VALUE ERROR SIZE DERIVATIVE 
1 p0 4.76424e+04 1.62805e+03 1.71983e+00 5.51654e-08
2 p1 1.40637e+02 1.37586e-01 4.40840e-04 5.55002e-05
3 p2 1.07821e+01 1.97640e-01 3.67135e-04 1.73281e-04
4 p3 2.40070e+04 1.60053e+03 1.53556e+00 5.43990e-09
5 p4 1.37466e+02 3.53371e-01 1.08635e-03 -5.50099e-05
6 p5 2.19889e+01 4.60091e-01 7.85441e-04 6.30307e-05
TFitEditor::DoFit - using function PrevFitTMP 0x428ebe0

And sometimes the status is MINOS ... SUCCESSFUL:

FCN=56.902 FROM MINOS STATUS=SUCCESSFUL 101 CALLS 790 TOTAL
EDM=9.53237e-09 STRATEGY= 1 ERROR MATRIX ACCURATE 
EXT PARAMETER STEP FIRST 
NO. NAME VALUE ERROR SIZE DERIVATIVE 
1 p0 4.76424e+04 1.62750e+03 2.02907e+01 -1.92708e-07
2 p1 1.40637e+02 1.37577e-01 -5.01453e-04 2.21597e-03
3 p2 1.07821e+01 1.97597e-01 1.97794e-03 -6.89765e-04
4 p3 2.40070e+04 1.59994e+03 -2.01573e+01 -4.05387e-07
5 p4 1.37466e+02 3.53380e-01 -1.63387e-03 -4.98786e-04
6 p5 2.19889e+01 4.59895e-01 4.59895e-01 1.16305e-03
TFitEditor::DoFit - using function PrevFitTMP 0x428ebe0

I couldn’t find any explanation what’s a difference between them. Could you explain?

Thanks in advance.


ROOT Version: 6.12/06
Platform: Scientific Linux 6
Compiler: Not Provided


I would assume that MINOS and MIGRAD just call their status differently.
MIGRAD does a function minimisation, but not a proper estimation of errors, and it says that it “Converges”.
MINOS estimates errors, and it says that it was “successful” doing that. MINOS is what you should use if you are interested in the uncertainties of your fit.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.