we should check different schemes of creating RFs for joint analysis, and comparing their performance with the up-to-date MCs:

1. M1+LST1, M2+LST1, M1+M2, M1+M2+LST1 (all combinations) - current scheme
2. M1, M2, LST1 - each telescope separately + common (e.g. stereo) parameters - like CTA scheme
3. M1+M2+LST1 all in one RF, zeroing the parameters if the telescope does not surviving cleaning - similar to MAGIC scheme
   ... (any other scheme?)

comparison of g/h separation performance on MC and on data