Determining the relative importance of criteria is a critical aspect of Multi-Criteria Decision Analysis (MCDA), directly influencing decision outcomes. Weighting methods in MCDA are generally divided into subjective approaches, based on expert opinions, and objective approaches, which derive weights from statistical data properties. Among subjective weighting methods, the RANking COMparison (RANCOM) approach has gained recognition for its simplicity and effectiveness in determining criteria importance. However, its standard formulation does not provide mechanisms for expert-driven adjustments after the initial weight computation, limiting its adaptability. To address this limitation, this study proposes an adaptive RANCOM-ST method, which systematically refines expert weight adjustments through the use of Statistical Thresholds—calculated based on the mean and standard deviation of errors observed in simulation results. In the proposed approach, experts assess the correctness of the generated weights and indicate whether they should be increased or decreased using a three-level scale. This additional step enhances the robustness of the weighting process by incorporating statistical measures to reduce expert bias and improve consistency. Synthetic experiments demonstrate that RANCOM-ST leads to more stable and reliable results compared to the traditional RANCOM method. The findings highlight the potential of RANCOM-ST as an effective refinement for subjective weighting methods, making expert-based MCDA models more resilient to inconsistencies.