Even experts cannot detect faking above chance (Fiedler & Bluemke, 2005). Recent studies (Boldt et al., 2018; Calanna et al., 2020) have suggested that machine learning may help in this endeavor. The ability of classifiers to detect faking depends on which classifiers are implemented (logistic regression vs. random forest vs. XGBoost; Calanna et al., 2020). The ability of classifiers to detect faking also depends on the type of input data (response patterns vs. scores; Calanna et al., 2020). However, faking differs with respect to faking conditions, and previous efforts have not taken these differences into account.