How do i know when the fit is good?

Hi, I am a PhD student who is new to SPR without anyone to really ask questions to so i was hoping i could perhaps get an answer here.

I have these two ligands that i am comparing and running analyte over and i want to compare the binding.
As the fit is better in the lower concentrations does that mean that i am mass transfer limited in the higher concentrations?
Is the fit still good in the association as i do have curvature?
During dissociation the baselines does not return to zero, how much will this affect the Koff and is this expected to have some binding still left?

From the inset I deduce that the sensorgram is double referenced. Despite this, there is clearly some non-specific binding which proportional to the analyte concentration used. Also in the association there is a sign of non-specific binding where the measured curve crosses the fitted curve. The dissociation fit is not following the curves and therefore the calculated values are not very reliable. I don't think you have to worry about mass transfer here. On the other hand it seems that the analyte concentration is still below K_D. Please study www.sprpages.nl/sensorgram-tutorial/a-curve and look at the curve spacing.