I have done a small systematic study on the possible bias that GENFIT could introduce on reconstructed barrel tracks, and I attach here a presentation with the results I obtained.
I copy here my final conclusions:
* GENFIT is stable for reconstruction of barrel tracks
* GENFIT loss in efficiency is negligible
* The minimum momentum resolution is ~1.5%, not possible to go below this number (material effects? field? hit resolution?)
* By using “realistic” momentum/position smearing the performance is stable, no big effects
* No need to increase the number of iterations to obtain better results
* No systematic problems at low momentum
We could discuss about at the next SeeVogh meeting, of course.