WAR for College Softball

Nobody publishes WAR for college softball, so we built it from the ground up instead of borrowing baseball's constants. The batting side starts from wOBA using run values fit on this season's own play-by-play, applies an opponent adjustment shrunk to game-level calibration, and expresses the result as wRC+, where 100 is league average and every point is one percent better.

The pitching side rests on a softball-native FIP: a log-linear model fit per inning on 1,051 official pitcher seasons, weighing home runs at +1.58, walks and hit batters at +0.73, and strikeouts at -0.69 per inning. Fitting on softball data matters because the sport's run environment is not baseball's: top strikeout rates reach 49 percent and the league allows 5.84 runs per seven innings.

Runs convert to wins at 10.48 runs per win, derived from Pythagenpat rather than a cross-team regression, which schedule strength badly confounds. Replacement level is a .250 team. Two honest gaps: WAR here has no fielding or baserunning component, because the public play-by-play names the position that fielded each ball but never the player. Treat it as offense plus pitching, which is most of the game but not all of it.

This dataset is free to use under CC BY 4.0: use it anywhere, including commercially, with attribution to Pioneer Pennant (pioneerpennant.com).