Several of us have noticed that fits to point sources with LiFF+3ML appear to be having trouble with the Log_parabola model. The fits tend to be unstable and we find that in the case where there is no curvature the model does not reduce to a simple power law. The TS is also lower for Log_parabola than for powerlaw, which should not generally occur since the extra DoF should make Log_parabola a "better" fit most of the time.
giacomov
commented
7 years ago
Several of us have noticed that fits to point sources with LiFF+3ML appear to be having trouble with the Log_parabola model. The fits tend to be unstable and we find that in the case where there is no curvature the model does not reduce to a simple power law. The TS is also lower for Log_parabola than for powerlaw, which should not generally occur since the extra DoF should make Log_parabola a "better" fit most of the time.