i am a beginner in funnel metadynamics and I was following the paper published in nature protocols about FMAP titled: Ligand binding free-energy calculations with funnel metadynamics. i started doing a well-tempered funnel metadynamics for just 10 ns for a complex different than the one in the paper ( I know it is a very short duration but I was just trying the small trajectory with the analysis implemented in the FMAP tcl code). I was trying to plot a scatterplot between fps.lp and fps.ld with points color as funne.bias as shown in the attached image. As far as I understood from the paper, the ligand should not feel the potential bias when inside the funnel-shaped part, however, as can be seen in the image there are some points at the edge of the funnel with a funnel.bias not equal to zero. is this ok or should I increase alpha parameter a little bit?

Note: the binding site is at the surface of the protein and it is partly exposed to solvent.

the funnel shape with the protein is in the second image.