Words like clutch and garbage time do a lot of work in sports writing and almost never get defined. Here is exactly what each one means when I use it: the filter that runs, the thresholds I picked, and the reasons you might pick different ones.
Every number on this site cites the definition behind it. Versions are permanent, so a piece written today still reproduces after I sharpen a definition next year.
A hitter's year-over-year change in OPS, ranked against every other hitter who was the same age entering the season. It answers the question a raw decline cannot: was this normal for a player his age, or unusual? A 25th percentile result means three quarters of his peers did better.
The four defensive penalties that hand an offense a new set of downs it did not earn: pass interference, roughing the passer, defensive holding, and illegal contact. Counted per 100 dropbacks so teams that throw more are not flattered.
Five or more pass rushers. Four or fewer is a standard rush. The count comes from FTN charting, which is why this split only exists from 2022 on.
Fourth quarter or overtime, inside five minutes, with the score within eight points, close enough that one score changes the outcome. This is the window where reputations get made, so it is the window we grade.
The clutch window, but only when the quarterback's team was behind. This is the honest version of "can he bring them back", it removes the easier job of protecting a lead.
Expected points added allowed per dropback on plays where the defense got no sack and no hit. With the rush removed from the play, what is left is mostly coverage.
How far a player sits from the middle of his own sport's distribution, so that distance can be compared with a player from another sport. Two numbers, because they answer slightly different questions: a z-score is how many standard deviations above his peer group's average he finished, and a percentile is the share of that peer group he beat.
Any play where the quarterback intended to pass, including sacks and scrambles, excluding spikes and aborted plays. Per-attempt stats hide sacks; per-dropback stats do not.
The gap between what a pitcher's earned run average actually was and what Statcast expected it to be from the quality of contact he allowed, his strikeouts and his walks. A positive gap means he pitched better than his ERA says. A negative gap means his ERA flattered him.
The distance between what a hitter actually produced and what Statcast expected from his contact quality. It separates a real decline from a run of bad luck: if the expected number falls with the actual one, the hitting got worse. If only the actual one falls, the outcomes did.
Plays run when the game was effectively decided, one team's win probability was above 95% (or below 5%). Defenses loosen, starters sit, and stats inflate. We strip these out so "production" means production that mattered.
Fourth quarter or overtime, a one-score game, and both teams still holding between a 10% and 90% chance to win. The game is live and the snap matters. Wider than the clutch window on purpose: this keeps the whole quarter rather than the last five minutes.
Saves and holds, set against the reliever's own ERA and WHIP. It asks whether the highest-leverage outs went to the pitchers who were actually getting outs.
A game where a team led by one to eight points with eight minutes left, and whether it went on to win. This is the outcome the rest of the analysis is trying to explain.
Fourth quarter or overtime with the offense tied or trailing by one score, and neither team above a 95% chance to win. The defense is trying to close the game out and the result is still genuinely in doubt. No time cutoff inside the quarter, because a defense that breaks with nine minutes left broke just the same.
For each hitter, the number of different batting-order slots he was used in across the season, and his average slot. A player used in two or three slots has a defined job. A player used in all nine does not.
Plays from the first three quarters with win probability between 20% and 80%, the game still genuinely in doubt, before desperation or coasting distorts play-calling. This is the fairest baseline for comparing players.
Every plate appearance taken by a player whose listed position is left, center or right field, added up for the whole team, then expressed as on-base plus slugging. It asks one question: when this team's outfielders hit, what happened?
Statcast's defensive measure: how many outs a fielder made compared with what an average fielder would have made on the same chances, accounting for how hard each ball was hit, how far he had to go, and how much time he had. Zero is average.
A sack or a quarterback hit. Hurries that force an early throw are not in the public data, so this undercounts pressure for every team equally. It is the honest measure available, not the complete one.
Behind at any point in the final eight minutes of the fourth quarter, or at any point in overtime. It is the situation behind the win-loss records people quote about who can and cannot come back.