Law enforcement is a profession of sedentary physical activity interspersed with physically demanding activity that requires high levels of fitness. It is imperative that agencies training law enforcement recruits maximise their fitness during their time at the academy. The aim of this study was to investigate changes in physical fitness during academy training. Retrospective data for 10 academy recruit classes, totalling 715 participants, were collected from a US law enforcement agency. The change in performance on two standardised tests were used as outcome measures. Comparisons were made between percentiles utilising one-way ANOVA and a linear mixed model (LMM). Overall, higher percentiles were found to have smaller improvements in physical fitness than lower percentiles. The results of the LMM support this supposition, showing that lower physical fitness scores resulted in greater improvements in a generalised fitness assessment (value = -0.45, standard error 0.02, p < 0.001) and an occupational assessment (value = -0.49, standard error = 0.02, p < 0.001). The results of this study suggest that recruits with lower physical fitness will see greater improvements during academy training. This could be due to a ceiling effect for the more fit but may also be due to recruits of higher physical fitness being under trained during academy. Utilising ability-based training and prescribing an appropriate workload to recruits of higher fitness may improve overall recruit fitness upon graduation.