We created a complex, interactive environment and a task to track the behavior of rats for morethan 230 testing days (Jankowski et al., 2023, PMID: 37568111). The experiment allowed ratsto freely adapt their behavior. The task involved a set of predefined rules and contingencies thatenabled them to predict reward and punishment locations, allowing them to satisfy their foodrequirements while expressing their individual preferences. Rats could freely choose a level ofdifficulty, which was linked to the size of the reward. We challenged them with changes in thelogic of the environment and observed how they adapted to the changing contingencies. Rats,as a group, quickly learned the simplest part of the task. Only after this option was disabled didthey learn one of the harder, but higher-rewarded, behaviors. When the simpler option becameavailable again, they combined both routines. Subsequently, they all learned to identify warningsounds and to avoid an aversive stimulus. While the group generally showed similar overallbehavior patterns in the task, each individual showed its own set of preferences for rewards,distinct strategies for achieving them, and different ways of resuming chosen behaviors afterbeing disrupted by internal or external distractors. Interestingly, for most subjects, thesestrategies changed over time. Our study reveals that in favorable conditions, rats exhibitcomplex, individual, time-varying behavioral policies. These results suggest that quick, simplebehavioral tests may not measure stable behavioral traits.