The phase diagram of the strongly interacting matter is the main research subject for different current and future experiments in high-energy physics. System size and energy scan programs aim to find a possible critical point. One of such programs was accomplished by the fixed-target NA61/SHINE experiment in 2018. It includes six beam energies and six colliding systems: p + p, Be + Be, Ar + Sc, Xe + La, Pb + Pb and p + Pb. In this study, we discuss how the efficiency of centrality selection by forward spectators influences multiplicity and fluctuation measures and how this influence depends on the size of colliding systems. We use SHIELD and EPOS Monte-Carlo (MC) generators along with the wounded nucleon model, introduce a probability to lose a forward spectator and spectator energy loss. We show that for light colliding systems such as Be or Li even a small inefficiency in centrality selection has a dramatic impact on multiplicity scaled variance. Conversely, heavy systems such as Ar + Sc are much less prone to the effect.