Analytics Glossary

Total WAR accumulated over an entire career.

Percentage of batted balls hit to center field.

The difference between the odds when a bet was placed and the closing line, indicating sharp betting.

Proportion of variance explained by a model, ranging from 0 to 1.

Percentage of swings on which a batter makes contact, calculated as (Swings - Whiffs) / Swings.

Measure of linear relationship strength between two variables, ranging from -1 to 1.

Technique for assessing model performance by training and testing on different data subsets.

Current market value of one win above replacement, typically $8-10 million.

Runs saved or cost by turning (or not turning) double plays.

The magnitude of a difference or relationship, independent of sample size.

Combining multiple models to improve prediction accuracy (random forest, boosting, etc.).

Runs saved or cost by avoiding or committing errors.

Percentage of batted balls classified as fly balls, calculated as FB / Balls in Play.

Measure of how much each input variable contributes to model predictions.

A double play in which the batter hits a ground ball resulting in two outs.

Betting half the Kelly-optimal amount for risk reduction.

Percentage of fly balls that are infield pop-ups, calculated as IFFB / FB.

The probability implied by betting odds, calculated as 1 / decimal odds.

Optimal bet sizing formula maximizing long-term growth: (bp - q) / b, where b = odds, p = probability, q = 1-p.

Measure of prediction accuracy for probability estimates, penalizing confident wrong predictions heavily.

Relief appearance that decreases team's win probability by at least 6%.

Park factor calculated over multiple seasons (typically 3-5) to increase stability and reduce single-season variance.

True implied probability after removing bookmaker's margin.

Percentage of batted balls hit to the opposite field.

When a model learns noise in training data, performing poorly on new data.

The highest WAR total achieved over a player's best consecutive seasons (typically 7 years).

Composite metric combining chase rate, zone swing rate, and contact rates to measure a hitter's approach quality.

Updated belief about a parameter after incorporating data.

Whether an effect is large enough to matter in real-world terms.

Initial belief about a parameter before seeing data in Bayesian analysis.

Bets from recreational bettors, often favoring favorites and overs.

Percentage of batted balls hit to the pull field (left field for right-handed batters, right field for left-handed batters).

Statistical measure indicating the proportion of variance in the dependent variable explained by the independent variable(s), ranging from 0 to 1.

The number of runs a player contributes above a replacement-level player, the foundation of WAR calculations.

Statistical technique for modeling the relationship between variables, or the tendency for extreme performances to move toward the mean over time.

Net profit divided by total amount wagered, expressed as percentage.

RE24 converted to wins using runs per win.

Standard deviation of prediction errors, measuring model accuracy.

The change in run expectancy caused by a specific event, used in linear weights calculations.

Number of observations needed for statistical stability, varying by metric.

Bets from professional or highly-informed bettors that move lines.

Estimate that pulls extreme observations toward the population mean.

Relief appearance that increases team's win probability by at least 6%.

The horizontal angle of a batted ball relative to center field, measured in degrees.

Percentage of batted balls where the ball is hit near the sweet spot of the bat, resulting in optimal energy transfer.

The sample size at which a metric is approximately 50% skill and 50% noise.

Measure of the precision of an estimate, decreasing with sample size.

Fantasy metric measuring value of statistics toward category standings.

The probability that an observed result occurred by chance, typically p < 0.05.

Projection system emphasizing recent performance and regression.

The difference between a player's projected production value and their salary cost.

The distance the bat travels during a swing from start to contact, measured in feet. Shorter swings typically correlate with better contact rates.

The bookmaker's commission built into betting odds, typically 4-5% for MLB.

A player's production above replacement level in fantasy context.

Sum of a player's seven highest single-season WAR totals.

Run cost of grounding into double plays relative to opportunity.

Percentage of swings that result in misses, calculated as whiffs / swings.

The change in win probability caused by a specific event, summed to show a player's contribution to team wins.

How well a metric predicts itself from one season to the next.