The evaluation of probe measurements in the swept regime can be performed via a fit to the ion branch of the current-voltage (I–V) characteristic, providing the most relevant plasma parameters such as the ion saturation current, the electron temperature and the floating potential. In this paper, we present a parametric study based upon the existing flush-mounted probe setup briefly operated at the COMPASS tokamak, aiming to compare relevant I–V formulas widely used for such analysis. Selected four-parametric fit descriptions were applied to synthetic data obtained by simulation of the probe in a particle-in-cell model, SPICE, both in simplified 2D and more complex 3D geometries. Plasma parameters recovered by the fit were compared to the input from the model and the precision of this recovery was mapped to a wide range of parameters, especially with respect to the ratio of the probe size dpin to the Larmor radius rL. Results show that while the electron temperature and the floating potential can be obtained quite precisely regardless of the method, the ion saturation current can be identified incorrectly, mainly due to the fact that the sheath expansion description by optical approximation is not sufficient for almost grazing field line angles of incidence, especially when dpin ∼ rL. We show, however, that operation of flush-mounted probes can be fruitful especially in high-field devices. Additionally, we have explored the sensitivity on the available bias potential range, showing that for proper analysis of I–V characteristic using four-parametric fit, the lower bound of the fit range should be at least below the floating potential. If this condition cannot be satisfied, three-parametric fit in the range close to the floating potential can still produce results with acceptable precision.