Among the different experimental strategies available in capillary electrophoresis (CE) to determine binding parameters, affinity capillary electrophoresis (ACE) has been the most widely embraced due to its easiness of implementation and of data handling. Ligand-substrate binding constants are thus directly derived from the substrate migration time shifts resulting from the variation of ligand concentration introduced in a background electrolyte. Classically, the substrate migration time is measured on top of the electrophoretic peak, assuming symmetrical peak shape. Depending on both substrate and ligand concentrations that may be required to meet detection sensitivity or complexation conditions, zonal migrations in ACE may, however, produce triangular peak shape, most often due to pronounced electromigration dispersion (EMD), and this may result in positively or negatively erroneous migration time assessments. In this work, EMD distorted triangular peak shapes obtained in the course of host-guest complexation studies were fitted with the Haarhoff-Van der Linde function, allowing better estimation of migration time. The model systems studied were those of beta-cycloclextrin and naproxen, 2-naphthalenesulfonate, or 1-adamantanecarboxylate. The impact of this correction on binding isotherms and binding constant evaluation was exemplified. Furthermore, in situations where the substrate concentration injected by far overtakes that of the ligand in the electrolyte, the interest in this peak shape correction was discussed in connection with the question of whether the free ligand concentration can be still considered equal to the ligand concentration introduced, a question that still remains under debate nowadays.