During early drug discovery the initial in vivo efficacy testing is often performed in rodent models optimized to screen and select lead compounds rapidly, before progressing them to in vivo models that reflect the human form of the disease more closely. The way such models are frequently run can risk overestimating the efficacy of new compounds when using pre- and co-administration, as shown in three examples from different central nervous system research areas. This is undesirable for reasons ranging from good decision-making, cost efficiency and time management to the ethics of animal use. Abandoning the use of pre-treatment, monitoring crucial physiological parameters in (satellite) animals and systematically applying simple pharmacokinetic-pharmacodynamic analysis could reduce the number of false positive results.