Purpose: The purpose of this study is to investigate the determinants of breakaway success in road cycling races. Design/methodology/approach: Descriptive statistics were computed, and a logit model of breakaway success was estimated based on a new kind of statistical data describing the development of each of the 268 breakaways that occurred in the 76 regular stages of the Tour de France 2013 to 2016. Findings: Breakaway success partly depends on the physics of cycling: breakaways are more successful when the stage is hilly or in mountain than flat. In addition, the likelihood of breakaway success depends on strategic moves such as attack timing and the percentage of riders with a teammate in the breakaway. Research limitations/implications: Understanding why certain breakaways succeed and others do not is useful to comprehend cycling performance and to help coopetitive temporary organizational forms such as breakaways optimize their strategic behavior. A limitation is the focus on the Tour de France only. Originality/value The present study adds to the literature on temporary organizational forms, coopetition and cycling performance by analyzing within-stage data in cycling and, as such, enabling to capture its strategic dimension.