What is Clutch? Definition, Formula, and Example
Clutch is a FanGraphs metric that measures how much better or worse a player performed in high-leverage situations compared to his own context-neutral production.
What is Clutch?
Clutch is a sabermetric statistic published by FanGraphs that measures the gap between a player's high-leverage performance and his own context-neutral baseline. It does not ask "was this player good in big spots?" — it asks "was he *better in big spots than he was overall*?" A 1.0 Clutch score means a player produced one win more than his own normal hitting line would have generated if those same plate appearances had been distributed across average leverage situations. Clutch can be positive or negative for any player and is calculated separately for hitters and pitchers.
How Clutch is calculated
The formula is:
Clutch = WPA / pLI − WPA/LI
Where:
- WPA (Win Probability Added) is the total change in team win probability a player produced
- pLI is the player's average Leverage Index across his plate appearances
- WPA/LI is context-neutral WPA — each event is divided by its own LI before summing, removing the leverage weighting
The first term, WPA/pLI, scales actual win-probability production by how high-leverage the spots were on average. Subtracting WPA/LI strips out the player's underlying offensive contribution. What's left is purely the over- or under-performance in leverage. FanGraphs' interpretation buckets are: 2.0+ excellent, 1.0 great, 0.5 above average, 0.0 average, −0.5 below average, −1.0 poor, −2.0 awful.
Worked example
David Ortiz finished his career with a +6.05 Clutch — among the highest figures ever recorded — meaning his postseason and high-leverage regular-season performance outstripped his (already excellent) context-neutral baseline by roughly six wins. On the other end, Alex Rodriguez posted a career −7.84 Clutch despite being a top-five hitter of his era; his high-leverage production trailed his overall line. In a single-season example, Freddie Freeman posted a +2.1 Clutch in 2023 — when leverage rose, his already-great numbers rose with it. Ronald Acuña Jr., the 2023 NL MVP, posted a −1.4 Clutch that same season — he was world-class in low- and medium-leverage spots but slightly less productive when the game was on the line.
Why Clutch matters
Clutch is one of the most cited and most argued stats in baseball. It surfaces narratives — "Big Papi was clutch, A-Rod wasn't" — with a number behind them. Front offices and analysts use it descriptively (what happened) but rarely predictively, because Clutch has very low year-to-year correlation. For DFS and prop bettors, a player's recent Clutch can flag whether late-inning matchups are likely to swing a slate. Hall of Fame voters frequently invoke clutch reputations, and the FanGraphs Clutch column is the closest thing to a defensible numerical proxy.
Limitations and common misconceptions
The single most important fact about Clutch is that it is largely not a skill. Year-to-year correlation for hitter Clutch is roughly 0.1 — barely above noise. A player who posts +2.0 Clutch one season is just as likely to post a negative figure the next. This means past Clutch should not be used to predict future Clutch. It also means the stat measures what happened, not who the player is. A second misconception: Clutch is not "performance in big spots." A player can hit .400 in high-leverage at-bats and still post a negative Clutch if his overall line was .450 — he was great, just *less* great when it mattered. Finally, Clutch is denominated in wins, so small samples (a single playoff series) can swing it more than a full regular season.
Related terms
In Legends Deck
Clutch is a hidden trait on hitter and reliever cards — it does not boost a player's base ratings, but it modifies outcome distributions in late-and-close at-bats inside the simulator. A high-Clutch career hitter like Ortiz has a tighter outcome variance with the tying run on base; a negative-Clutch superstar plays at his normal rating in those spots, not better. Because the underlying stat is noisy, we cap the in-game effect modestly — Clutch tilts the dice, it doesn't roll them.