Abstract Background Challenging behaviors like aggression and self-injury are dangerous for clients and staff in residential care. These behaviors are not well understood and therefore often labeled as “complex”. Yet it remains vague what this supposed complexity exactly entails at the individual level. The current case-study used a mixed-methods analytical strategy, inspired by complex systems theory, that consisted of three steps. First, we construed a holistic summary of the client’s relevant factors in her daily life. Second, we described her challenging behavioral trajectory by identifying stable phases. Third, instability and extraordinary events in her environment were evaluated as potential mechanisms for changes between different phases. Case presentation A woman, living at a residential facility, with mild intellectual disability and borderline personality disorder with a chronic pattern of aggressive and self-injurious incidents. She used ecological momentary assessments to daily self-rate challenging behaviors for 560 days. Conclusions A qualitative summary of caretaker records revealed many internal and environmental factors relevant to her daily life. Her clinician narrowed these down to 11 staff hypothesized risk- and protective factors, such as reliving trauma, experiencing pain, receiving medical care, compliments or psychological therapy. The vast multitude of bivariate associations between these 11 factors and self-reported challenging behaviors were non-significant. These null-results indicate that challenging behaviors are not governed by mono-causal if-then relations, speaking to its complex nature. Despite this complexity there were patterns in the temporal ordering of incidents. Aggression and self-injury occurred on respectively 13% and 50% of the 560 days. On this timeline, we identified 11 distinct stable phases, that were alternating between four unique states: high levels of aggression and self-injury, average levels of aggression and self-injury, low levels of aggression and self-injury, and low aggression with high self-injury. Eight out of ten transitions between phases were either triggered by extraordinary events in her environment (event-induced) or preceded by increased fluctuations in her self-ratings (instability-induced), or a combination of these two. Desirable patterns emerged more often and were less easily malleable, indicating that when the participant experiences bad times, keeping in mind better times are ahead is as hopeful as it is realistic.