Aims: When conducting computational electrophysiological simulations aimed at assessing pharmacological cardiotoxicity, several parameters play crucial roles. Among theme, the size of the population under scrutiny emerges as potentially critical, negatively impacting on the expected final outcomes. The process of pharmacological pro-arrhythmia classification entails an examination of virtual cardiomyocytes samples, in which at each cell is associated with an action potential. This process involves then the selection of a specific electrophysiological model, followed by the replication of cellular responses subsequent to electrical stimulation. These simulations encompass the assessment of morphological alterations in action potentials as well as intracellular concentration dynamics. Currently, in literature there's not a standardization concerning parameters and their associated values within pharmacological cardiotoxicity simulations. Particularly noteworthy is the lack of a definitive standards specifying the optimal sample size to be employed. This study attempts to shed light on the influence exerted by population size on the ultimate classification of drugs, pointing out the potential for conflicting results to arise. Methods: Initially, a population of 100,000 stable virtual cardiomyocytes was generated. From this pool, a further extraction process yielded 2989 stable cells, which were subjected to cardiotoxicity analysis. Subsequently, a bootstrapping process was performed, involving the random extraction of 107 cells from the 2989 and iterating the operation 200 times for statistical significance. Similarly, the same methodology was replicated to select the initial 2989 cells, iterating the process 10 times due to high computational costs. Results and Conclusions: The study reveals a greater variability within the population of 107 cells as opposed to the larger cohort of 2989 cells. This discrepancy is to highlight the potential for considerable errors in pharmacological cardiotoxicity classification deriving from the utilization of inadequate sample sizes.