The use of a pitch count estimator to calculate exposure in collegiate baseball pitchers

Objective: Excessive pitch counts have been associated with arm pain in pitchers. Tracking of exposure is difficult based on participation on multiple teams and variability in organizational rules. Statisticians have estimated exposure for professional pitchers using pitch count estimators. Our objective was to determine the utility of pitch count estimators at the collegiate level. Design and participants: Cohort; 29 collegiate pitchers. Methods: The team athletic trainer collected game pitch count data. The total number of batters faced (PA), strike outs (SO), and walks (BB) were recorded from the box score and entered into the pitch count formula (3.3*PA + 1.5*SO + 2.2*BB) to estimate pitch counts. Intraclass correlation coefficients (ICCs_(2,1)) and standard error of measurement (SEM), were used to examine the agreement between actual and estimated pitch counts. Results: The mean pitch count was correlated with the estimated pitch count (628 ± 476vs.603 ± 426; r = .99, p < .001). The actual and estimated pitch counts per season demonstrated excellent agreement (ICC_(2,1) = 0.99; SEM = 56 pitches). The ICC_(2,1) calculated to compare actual and estimated pitch counts for starters and relievers (ICC_(2,1) = 0.98; SEM = 77; 0.98; SEM = 39) reflect good agreement. Conclusions: The estimator provides a method of quantifying exposure for pitchers to help plan safe participation and control for confounding factors when attempting to understand the risks of pitching.