Squared Pearson correlation coefficients as functions of the heliocentric distance guess for different samples (1-day cadence) of Makemake ephemeris, all covering 1 month in 2017. Despite almost identical time spans, numbers of observations and observation frequency, the ρ² functions differ greatly. For such data, the fitting range [d_min, d_max] must closely bracket the actual physical peak. Otherwise, the maximization routine may end up stuck at the borders of the fitting range or at one of the secondary peaks. While this could be avoided by checking linearity for small steps in heliocentric distance, this would void the speed of the fitting routine; so instead, the analysis is carried out with steps of d_max − d_min = 10 au.