Pacing strategy patterns and performance outcomes in marathon running: A large-scale analysis of split time data

Abstract

Although pacing strategy is widely recognized as critical for marathon performance, the actual distribution of pacing patterns among recreational runners and their associations with finish time across demographic groups have not been systematically characterized in large samples. This study analyzed split time data (i.e., cumulative elapsed times recorded at intermediate checkpoints including 5K, 10K, 15K, 20K, half-marathon, 25K, 30K, 35K, and 40K) from 78,912 finishers of the Boston Marathon (2015–2017), characterizing pacing profiles using the ratio of first-half to second-half completion times and segment-by-segment pace variations across these nine timing points. Cluster analysis was employed to identify distinct pacing patterns, and the relationships between pacing strategy and finish time were examined across age groups, sex, and performance levels using analysis of variance and multiple regression analysis. Four distinct pacing patterns were identified: even-pacing (maintaining consistent pace throughout; 47.9%), mild positive-splitting (controlled late-race slowing; 31.9%), strong positive-splitting (pronounced late-race slowing; 12.2%), and variable-pacing (highly inconsistent pace fluctuations; 8.0%). Even-pacing was consistently associated with the fastest finish times across age and performance strata ( p  < 0.001), with mean finish times markedly lower than those observed for strong positive-splitting (217 vs. 276 minutes). Variable-pacing was associated with the highest incidence of “hitting the wall” (pace decline >20% in the final segment; 85.4%). High-performance runners (<3:00 finish time) demonstrated significantly higher rates of even-pacing (82.0%) compared to casual participants (19.6%). Age and sex significantly influenced pacing strategy selection ( p  < 0.001). These findings may help inform marathon pacing guidance by indicating that age, sex, and performance level should be considered when interpreting pacing patterns and developing individualized race plans.