The problem of power exhaust is one of the grand challenges of nuclear fusion research today. In order to understand the physics phenomena occurring in the scrape-off layer and the divertor regions of tokamaks, it is essential to correctly determine the divertor plasma parameters, which are often measured by swept Langmuir probes (LPs). While the construction and operation of this diagnostic can be straightforward, the data analysis using three- or four-parameter fits presents a challenge and can potentially lead to erroneous values of electron temperature and ion saturation current. In this work, we present modelling and experiments aimed at determination of conditions for proper analysis of swept LPs using these two fitting models. Particle-in-cell modelling was employed to evaluate the sheath-expansion effects for particular probe geometry and plasma conditions, yielding a semi-empirical rule capable of predicting its magnitude. Experiments with unusually wide range of swept voltage in the divertor of the COMPASS tokamak explored the magnitude of voltage range required for successful analysis with either three or four-parameter fitting. With the use of our new semi-empirical rule, it is possible to improve the four-parameter fit reliability in situations where the available voltage range is limited. In addition, we introduce the tangent method—an independent and fast method of electron temperature estimation, which allows to reliably determine the available voltage range and as such assist more complex methods of probe analysis.