The percentage of a pitcher's swings against that result in a swing-and-miss.
Strikeouts are downstream of swings and misses. A pitcher who gets hitters to swing through his pitches will, in any plate appearance long enough to reach two strikes, generate strikeouts. Whiff% measures the underlying skill directly: of all swings against this pitcher, how many came up empty?
The metric is computed per pitch type and overall. A pitcher's slider whiff rate is meaningful — it tells you how often that specific pitch beats the bat. Aggregating across all pitches gives the overall whiff rate, which correlates strongly with strikeout rate.
Whiff% is more stable in small samples than strikeout rate, because it removes the at-bat-level variance. A pitcher who gets 35% whiffs on a thousand swings is a high-strikeout pitcher; one who gets 22% is not. The metric stabilizes in roughly 200-300 swings — fast enough that early-season trends are informative by mid-May.
The metric also slices by location. Whiffs in the zone are different from whiffs out of the zone — chase rate matters separately. The two together explain most of the variance in strikeout rate. Modern pitcher development is built explicitly on whiff% optimization: pitch design, sequencing, and pitch mix are all evaluated against their effect on the whiff rate against major league hitters.
Our pitcher strikeout-prop model is built on per-pitch whiff rates against handedness, weighted by projected pitch mix. The model outperforms naïve K/9 projections because it conditions on the lineup the pitcher will actually face.