Baseball analytics glossary

Every stat and term used in TwentySeven, defined in plain English. No formulas unless the formula is the point.

Win probability

Win probability (WE) also: win expectancy: The estimated chance a team has of winning the game at a specific moment, shown as a percentage. It comes from looking at the current score, inning, outs, and baserunners, then asking how often teams in that exact situation won historically. The two teams’ figures always sum to 100%. At the first pitch of an even game the home team sits slightly above 50% because it bats last. Full guide →
Win probability curve also: win probability chart, WP graph: A line plotting one team’s win probability across the whole game, from first pitch to final out. It rises when that team helps itself and falls when it doesn’t, and it always ends pinned at 100% or 0%. A jagged, back-and-forth line means a close game; a line drifting smoothly to one side means a comfortable one. Near-vertical jumps mark the biggest plays. Full guide →
Win probability added (WPA): The change in a team’s win probability caused by a single play, credited to the players involved. A hit that lifts the team from 40% to 55% is worth +0.15 WPA to the batter and −0.15 to the pitcher. Summed across a game or season, WPA measures a player’s impact on outcomes rather than raw production — it knows the difference between a home run in a blowout and one in a tie game. Full guide →
Championship win probability added (cWPA): The same idea as WPA, but measuring the change in a team’s chance of winning the World Series rather than the current game. Regular-season plays typically move it by hundredths of a percent; a World Series Game 7 at-bat can move it twenty points or more. Full guide →
Leverage index (LI): A measure of how much the current moment could swing the game, scaled so that 1.0 is an average situation. Above roughly 1.5 is high leverage; below about 0.85 is low leverage. Leverage describes the situation, not the outcome — it tells you how much is at stake before the pitch is thrown. Full guide →
High-leverage situation also: high leverage spot: A moment where the next play could move win probability substantially — typically a close score, late innings, and runners on base. Conventionally any situation with a leverage index above about 1.5. These are the spots managers save their best relievers for, and they are a small fraction of the pitches thrown in a game. Full guide →

Run scoring

Run expectancy (RE): The average number of runs a team scores from a given base-out state through the end of the inning, measured across thousands of historical innings. Bases empty with nobody out is worth about 0.48 runs; bases loaded with nobody out is worth about 2.29. Full guide →
Base-out state: The combination of which bases are occupied and how many outs there are. Eight base configurations times three out counts gives exactly 24 possible states, and every moment of every inning is one of them. Full guide →
RE24 also: run expectancy 24: A stat crediting a player with the change in run expectancy his play caused: run expectancy after the play, plus runs scored, minus run expectancy before. It measures runs created in context, where WPA measures wins. Full guide →
Run differential: Runs scored minus runs allowed, over any span from one game to a full season. Roughly ten runs of season differential is worth about one win. It predicts future record better than current record does, because a record treats a one-run win and a ten-run win identically while differential keeps the margin. Full guide →
Run environment: How easy it is to score in a given era, league, or ballpark. A high run environment makes each individual run less valuable, which is why many statistics are park- and league-adjusted before being compared across seasons.

Team and season

Pythagorean expectation also: Pythagorean record, expected win-loss: An estimate of the record a team should have based only on runs scored and allowed. The original formula is runs scored squared divided by runs scored squared plus runs allowed squared; modern versions use an exponent near 1.83. A team well above its Pythagorean record has usually been winning close games and losing blowouts, a pattern that rarely repeats. Full guide →
Pythagenpat: A refinement of Pythagorean expectation that sets the exponent from the game’s run environment instead of fixing it at 2 or 1.83. It is more accurate at scoring extremes, though the difference is usually under one win per season. Full guide →
Playoff odds: A team’s estimated chance of reaching the postseason, produced by simulating the rest of the season many thousands of times and counting how often the team qualifies. Full guide →
Monte Carlo simulation: A technique that answers a question too tangled for a formula by simulating it many times and counting outcomes. In baseball it is how playoff odds and championship odds are generated. Full guide →
Strength of schedule (SOS): A measure of how strong a team’s opponents are, usually the weighted average of their winning percentages. Past SOS tells you how much to trust a record; remaining SOS tells you what to expect. Full guide →
Magic number: The combined number of wins by the leading team and losses by its closest rival required to clinch a division or playoff spot. When it reaches zero the spot is secured.
Elimination number also: tragic number: The mirror image of the magic number: the combined number of losses by a trailing team and wins by the leader that would eliminate the trailing team from contention.

Player value

Wins above replacement (WAR): An estimate of how many more wins a player is worth than a freely available replacement-level substitute, combining batting, baserunning, fielding, and positional difficulty. Roughly 2 is a solid regular, 5 is All-Star level, and 8 or more is an MVP season. Full guide →
Replacement level: The production a team could get for nothing — a minor-league call-up or waiver claim. A team composed entirely of replacement-level players would win roughly 48 games in a 162-game season. Full guide →
fWAR and bWAR: The FanGraphs and Baseball-Reference versions of WAR. They differ mainly on pitching: FanGraphs builds pitcher value from fielding-independent outcomes, while Baseball-Reference uses runs actually allowed adjusted for team defense. The two should never be compared directly against each other. Full guide →
wOBA also: weighted on-base average: A rate stat that weights each way of reaching base by how much it actually contributes to scoring, rather than treating a walk and a home run as equally valuable. It is scaled to look like on-base percentage, so roughly .320 is average.
wRC+ also: weighted runs created plus: A park- and league-adjusted measure of total offensive value where 100 is exactly league average. A wRC+ of 140 means a hitter created 40% more runs than an average hitter in the same conditions.
OPS and OPS+: OPS is on-base percentage plus slugging percentage, a quick summary of offensive production. OPS+ adjusts that figure for park and league and sets 100 as average, so 130 means 30% better than average.
FIP also: fielding independent pitching: An ERA-scaled estimate of a pitcher’s performance built only from strikeouts, walks, hit batters, and home runs — the outcomes that do not depend on the fielders behind him. A large gap between ERA and FIP often signals unusual defense or luck.
BABIP also: batting average on balls in play: The batting average on balls hit into the field of play, excluding home runs and strikeouts. League average sits near .300, and unusual values often regress, which makes it a common luck indicator.
ERA+: Earned run average adjusted for ballpark and league, scaled so 100 is average and higher is better. An ERA+ of 150 means a pitcher was 50% better than league average at preventing earned runs.
Exit velocity: The speed of the ball off the bat, measured in miles per hour. Harder contact produces better outcomes on average, which makes exit velocity one of the core Statcast batted-ball measurements.
Barrel: A Statcast classification for a batted ball whose combination of exit velocity and launch angle has historically produced at least a .500 batting average and 1.500 slugging percentage. Barrel rate is a strong indicator of power quality.

Situations

Late and close: A defined split requiring two conditions at once: the seventh inning or later, and the batting team leading by one, tied, or with the potential tying run at least on deck. Sample sizes are small — around 100 to 150 plate appearances a season — so the split describes what happened but predicts poorly. Full guide →
Clutch: Performing better in high-pressure situations than in ordinary ones. Research has repeatedly found that clutch hitting is not a stable, repeatable skill: players who excel one season mostly do not the next. Full guide →
Walk-off: A play by the home team that ends the game immediately, scoring the winning run in the bottom of the final inning. It carries the largest possible single-play win probability swing, since it takes the team straight to 100%.
Save situation: A defined condition for a relief pitcher who finishes a game his team wins: entering with a lead of three runs or fewer, or with the tying run on base, at bat, or on deck, or pitching at least three effective innings. Not the same as late and close, which includes tie games and applies to any plate appearance.
Hold: Credited to a relief pitcher who enters in a save situation, records at least one out, and leaves with the lead intact without finishing the game. It exists to give setup relievers a countable statistic.
Quality start: A start of at least six innings with three or fewer earned runs allowed. It is a deliberately low bar — six innings and three earned runs is a 4.50 ERA — and is best read as a durability measure rather than an excellence one.