GUIDES · BETSLIP DOCTOR
What does an expected Statcast batting statistic add to the observed result?
An expected Statcast batting statistic asks what comparable batted-ball inputs tend to produce; an observed statistic records what actually happened. MLB’s xBA estimates hit probability from contact characteristics, xSLG assigns probabilities to hit types and their total-base values, and xwOBA applies weighted on-base values while retaining real outcomes for events such as walks and strikeouts. These measures help describe contact quality alongside results, but none changes the official credited outcome or predicts what will happen next.
For Bettors and cappers learning to read baseball records · Updated
Expected is a model summary; observed is the recorded outcome
A hard-hit ball can become an out, and a softly hit ball can fall for a hit. The observed batting line records the official result. An expected measure estimates outcomes from comparable tracked contact, so it can show how the contact profile differs from the results that happened. A gap is context for review, not proof that the observed result was wrong or that future results will move toward the estimate.
The three expected statistics answer different questions
MLB’s glossary describes xBA as a hit-likelihood measure for batted balls, based mainly on exit velocity and launch angle, with Sprint Speed used for certain contact types. Seasonal xBA also keeps real strikeouts in the calculation. xSLG models probabilities of singles, doubles, triples, and home runs, then values those outcomes by total bases. xwOBA uses expected batted-ball outcomes with weighted on-base values; MLB’s definition retains actual data for non-contact events such as walks, strikeouts, and hit-by-pitches.
- xBA: modeled hit probability on contact, expressed on a batting-average scale.
- xSLG: modeled hit-type probabilities translated into total bases per at-bat.
- xwOBA: modeled contact outcomes in a weighted on-base framework, alongside actual non-contact events.
Compare the same player, period, and opportunity set
Before comparing expected and observed values, check whether the figures cover the same dates, competition, and sample. A short stretch can produce a large gap by chance. Leaderboard qualification is a separate question from a metric’s definition; use the current table’s stated minimum rather than assuming one threshold applies everywhere. BSD does not claim to retrieve, calculate, or display every Statcast metric discussed here.
Hypothetical worked example: two contact outcomes
Toy model only—not MLB data or an MLB model output. Consider two at-bats with no walks, strikeouts, or other plate-appearance events, so this example isolates contact. Ball A is hit at 91 mph and a 12° launch angle; the teaching model assigns probabilities of 0.14 single, 0.04 double, 0.01 triple, 0.01 home run, and 0.80 out. Ball B is 104 mph at 29°; the toy probabilities are 0.30 single, 0.20 double, 0.05 triple, 0.15 home run, and 0.30 out. The toy xBA is (0.20 + 0.70) / 2 = .450. Expected total bases are 0.29 and 1.45, so toy xSLG is 1.74 / 2 = .870. If A is actually an out and B an actual home run, observed AVG is .500 and observed SLG is 1.000. xwOBA is not calculated here: its season-specific weights and the actual non-contact outcomes would also be needed. These different numbers describe model inputs and credited results; none is a forecast.
Related questions
Does a higher expected average change the official batting average?
No. The observed batting average uses credited hits and at-bats. xBA is a separate modeled summary.
Do xBA, xSLG, and xwOBA mean the same thing?
No. They use different outcome scales and event treatment. Compare each with its own definition and sample.