In an era where violent extremism poses increasing global threats, identifying the key predictors that drive individuals towards ideological violence is paramount. This study systematically compares 12 critical predictors of violent extremism identified through recent meta-analyses, including past activism, collective relative deprivation, obsessive passion, and identity fusion. Data from over 40 nations (N=10,659), covering left-wing, right-wing, and religious extremism, are analyzed in what we term a "predictor tournament", where each variable's influence is competitively assessed. We employ Random Forest regressions to rank these predictors based on their association with support for ideological violence. Our methodology includes not only broad population surveys but also unique data from two distinct groups of terrorists (N=160): one incarcerated and one evading capture, thereby providing a dual lens on radicalization pathways. Our findings reveal a striking consistency in predictors across both general populations and terrorist groups, underscoring shared psychological and contextual mechanisms. This convergence suggests that radicalization drivers transcend ideological boundaries, pointing to universal motivational patterns in the pathway to ideological violence and offering crucial insights for counter-radicalization strategies.