Goueta et al. (Journal of Child Psychology and Psychiatry, 2025, 66, 1209) utilized a Random Intercept Cross-Lagged Panel Model (RI-CLPM) to investigate reciprocal links between ADHD symptoms and adolescent risky behavior. Their finding that stable between-person differences, rather than within-person fluctuations, primarily drive this association challenges common clinical assumptions. While commending their methodological rigor, this commentary proposes five refinements to better capture the dynamic ADHD-risk nexus. First, aggregating diverse risk behaviors may mask distinct symptom-coupled fluctuations, requiring multivariate models to separate impulsivity from social deviance. Second, relying exclusively on parent reports introduces bias; future studies should incorporate multi-informant designs, ecological momentary assessment (EMA), and passive sensing. Third, standard time metrics overlook critical developmental milestones; event-contingent sampling around salient transitions can address process non-stationarity. Fourth, integrating time-varying mediators, such as sleep and parental monitoring, can reveal precise windows of heightened risk for targeted interventions. Finally, dimensional scoring might obscure non-linear threshold effects and pharmacological impacts. By addressing behavioral heterogeneity, reporter variance, developmental contexts, and non-linearities, future research can clarify exactly when, for whom, and under what conditions ADHD symptom fluctuations forecast adolescent risk.