Advanced Stats in MLB Game Summaries: WPA & Leverage

A walk-off grand slam and a routine seventh-inning groundout can both reduce to a single tally in a hit or out column, yet one of them may swing a team’s chance of winning by more than 90 percentage points while the other barely moves the needle. That gap is exactly what advanced situational metrics were built to measure. Win Probability Added (WPA), Use Index, and Run Expectancy each translate a raw event into the value it carried in the moment it happened, and they have become standard furniture in modern baseball recaps. This article explains what each metric means, where it came from, and how to read all three inside a daily MLB game summary.

Why Advanced Stats Belong in Modern Game Summaries

Traditional box score lines tell you what happened. They rarely tell you how much each event mattered. A two-out single with the bases empty in a blowout and a two-out single that drives in the tying run in the ninth inning look identical in a hit column, but their effect on the result could not be more different. Situational metrics close that gap by weighing every play against the exact game state in which it occurred, so the story of a close game reads differently from the story of a rout. Readers who already know how to read an MLB box score get the most out of these numbers, because the advanced figures sit on top of the same events.

Three metrics anchor most current recaps, and each answers a separate question. Run Expectancy asks how many runs a base-out situation is worth on average. Use Index asks how much the situation could swing the final outcome. WPA asks how much a specific play actually changed a team’s probability of winning. Used together, they convert a flat list of events into a record of where a game was decided.

None of these figures replace the narrative of a recap. They sharpen it.

A Short History of Win Probability and Run Expectancy

The idea of valuing baseball situations by their run or win consequences is older than the recent analytics boom. George Lindsey published An Investigation of Strategies in Baseball in the journal Operations Research in 1963, building one of the first run expectancy tables from thousands of recorded plays (INFORMS, 1963). His work showed that the average number of runs scored from a given base-out state could be estimated and used to judge strategy.

Brothers Eldon and Harlan Mills carried the concept into the win column. Their 1970 system, Player Win Averages, assigned credit for every plate appearance according to how much it shifted each team’s chance of winning, a direct ancestor of today’s WPA (SABR). The method was ahead of its computing era and faded from public view for decades.

Modern situational analysis took its current shape in the 2000s. Tom Tango, working with Mitchel Lichtman and Andrew Dolphin, formalized Use Index and popularized win-probability reasoning in their 2007 book The Book: Playing the Percentages in Baseball. Public sites then made the numbers free and accessible at the play level, which is why a casual reader can now look up the WPA of any plate appearance from a regular-season game.

Statcast, deployed across all 30 ballparks by Major League Baseball in 2015, layered tracking data on top of this situational foundation (MLB.com). The result is that a single play can now be described by both what the ball did and what the play was worth.

Win Probability Added (WPA): Crediting the Swings

Win Probability Added measures the change in a team’s win expectancy from the moment before a play to the moment after it, then credits that change to the players involved (Wikipedia). Win expectancy itself comes from historical data: given the inning, the score, the number of outs, and which bases are occupied, what share of teams in that exact spot have gone on to win? A hitter who improves his team’s win expectancy from 40 percent to 55 percent earns +0.15 WPA on that play, and the opposing pitcher is charged the matching –0.15.

One elegant property makes WPA easy to audit. Because win probability always begins at 50 percent at first pitch and ends at 100 percent for the winner, the winning team’s combined player WPA in any game sums to exactly +0.5, and the losing team’s sums to –0.5 (FanGraphs library). That balance lets a recap point to the handful of plays that account for most of that half-win.

The size of a single play’s WPA depends almost entirely on timing. A solo home run that breaks a scoreless tie in the first inning might be worth around +0.10, while the same home run as a walk-off swing can be worth several times more. A walk-off grand slam by a team that trailed by three runs with two outs in the bottom of the ninth can approach +0.95 WPA, near the theoretical ceiling for one play, because it carries the team from near-certain defeat to a finished win. The table below shows representative values to illustrate the range.

A few related variants appear in deeper recaps. Championship WPA (cWPA) rescales the idea to a team’s odds of winning the World Series rather than a single game, so a September plate appearance for a contender can carry visible cWPA while an identical play for an eliminated team carries almost none.

Use Index: Measuring the Pressure

Use Index, introduced by Tom Tango, scores how much is at stake in a given moment by comparing the potential swing in win probability for that situation against the swing in an average situation (Wikipedia). The scale is anchored so that an average plate appearance has a Use Index of 1.0 (FanGraphs library). A spot twice as pivotal as average reads 2.0, and a low-stakes moment in a blowout might read near 0.1.

Analysts often treat any moment above roughly 1.5 as high use, with the most extreme late-game, close-score situations climbing well past 5.0 (FanGraphs library). The value rises with later innings, tighter scores, more runners on base, and fewer outs to work with, because each of those conditions widens the range of outcomes a single play can produce.

Use is most useful for evaluating relief pitching, because relievers do not all face the same pressure. Variants such as gmLI capture the average use a reliever inherits when he enters the game, which separates a closer who routinely protects one-run leads from a long reliever who mops up lopsided games. When a recap describes a reliever as having worked the highest-use outs of the night, this is the number behind the claim. For more on how recap writers fold these moments into prose, see our guide on how to write an MLB game recap.

Run Expectancy and RE24: The Value of Every Base-Out State

Run Expectancy answers a simpler question than win probability: from a given base-out state, how many runs does the average team score through the end of that inning? There are 24 such states, formed by eight possible base configurations multiplied by three out counts (Wikipedia). The full grid is called the run expectancy matrix, and its companion stat, RE24, credits a player with the change in run expectancy his plays produce.

The numbers reward what intuition already suggests. With the bases empty and nobody out, the average team scores roughly 0.48 runs over the rest of the inning, while with the bases loaded and nobody out that figure climbs to about 2.29 runs (run expectancy data compiled by Tom Tango, published via Baseball Reference). Every out erases a large share of that potential, which is why a strikeout with the bases loaded shows up as a big negative for the hitter and a big positive for the pitcher.

RE24 reads the matrix in motion. A double that moves a runner from first to third and puts the batter on second adds the difference between the new state’s expectancy and the old one, plus any runs that scored. Because it counts every base advanced and not just runs batted in, RE24 often rewards productive outs and well-placed singles that a basic line ignores. If you want to see how the same play looks across formats, our breakdown of the game summary, box score, and play-by-play shows where each number lives.

How These Metrics Work Together in a Game Summary

The three numbers describe the same play from different distances. Run Expectancy treats the play in isolation, asking only about runs in the current inning. Use Index sets the stakes, telling you whether the moment was loud or quiet. WPA delivers the verdict, recording how much the play moved the needle toward a win. A clean way to remember the relationship is that WPA is roughly the run impact of a play scaled by how much that inning’s runs matter to the final result.

Consider a tie game in the eighth inning with a runner on second and one out. Run Expectancy says the situation is worth about 0.66 runs (Tom Tango run expectancy data via Baseball Reference). Use Index flags it as high, perhaps near 2.0 (FanGraphs library). A go-ahead double in that spot therefore produces a modest RE24 gain but a large WPA, because the run it drives in arrives when the win hangs in the balance. The same double in the first inning of a scoreless game would carry nearly the same RE24 and a fraction of the WPA.

This layering is why a strong recap can rank the night’s biggest plays without resorting to opinion. The plays with the largest WPA are, by definition, the ones that decided the game.

Reading Advanced Stats Without Getting Lost

Start with the win probability graph if the summary includes one. The line runs from 50 percent at first pitch to 100 percent for the winner, and the steepest climbs and drops mark the turning points of the game. Hovering over those steep segments usually reveals the play and its WPA, which gives you the night’s decisive moments in seconds.

Next, scan for any play with an absolute WPA above about 0.20, since those are the swings large enough to reshape a result. Then check the Use Index attached to key relief appearances to judge whether a pitcher earned his outs under real pressure or in calm water. Reading the numbers in that order, from the graph to the big plays to the use, keeps the focus on what changed the game rather than on raw totals.

• Use the win probability graph to find turning points at a glance.
• Treat WPA above 0.20 in absolute terms as a game-shifting play.
• Read Use Index to separate high-pressure outs from low-stakes ones.
• Lean on RE24 to credit productive plays that RBI totals miss.

Limitations and Common Misreadings

These metrics describe what happened, not what a player is likely to repeat. WPA in particular is heavily context-dependent, since a hitter only accumulates large totals if his teammates create high-use chances for him. A great hitter buried in a bad lineup can post a modest seasonal WPA despite excellent underlying numbers, which is why analysts pair WPA with context-neutral measures rather than reading it alone.

Sample size is the other trap. Over a single game or a short stretch, use and timing dominate, so a fluky walk-off can hand a journeyman a bigger WPA night than a star earns. Run expectancy values also drift from season to season because they depend on the league’s overall run environment, so the exact figures in any matrix are a snapshot rather than a constant. Treat the numbers as a precise account of one game and a noisy guide to talent, and they will rarely mislead. If you are new to the broader vocabulary, our explainer on what an MLB game summary is covers the basics these stats build on.

Frequently Asked Questions

What does WPA mean in a baseball game summary?

WPA stands for Win Probability Added, and it measures how much a single play changed a team’s chance of winning. Each play has a win expectancy before and after it, drawn from historical data on the inning, score, outs, and runners, and the difference between those two figures is the play’s WPA. A hitter who lifts his team from 40 percent to 55 percent earns +0.15, while the opposing pitcher is charged –0.15. Across a full game, the winning team’s player WPA sums to +0.5, which makes the metric a tidy way to identify the plays that actually decided the result.

What is a high Use Index?

Use Index is scaled so that an average plate appearance equals 1.0, meaning anything above that line is more pivotal than a typical moment. Analysts commonly treat values above roughly 1.5 as high use, and the most extreme late-inning, close-score, bases-loaded situations can climb past 5.0 or even 10.0. The figure rises with later innings, tighter scores, more baserunners, and fewer outs, because each of those conditions widens the range of outcomes a single play can create. In recaps, Use Index is used most often to judge whether a reliever recorded his outs under genuine pressure or in a low-stakes situation.

How is Run Expectancy calculated?

Run Expectancy is built from large samples of historical play data. For each of the 24 base-out states, formed by eight base configurations and three out counts, analysts average how many runs teams actually scored from that point through the end of the inning. With the bases empty and no outs, that average sits near 0.48 runs, while with the bases loaded and no outs it rises to about 2.29 runs, according to run expectancy data compiled by Tom Tango. The companion stat RE24 then credits players with the change in run expectancy their plays produce, including every base advanced and not only runs batted in.

Where can I find these advanced stats for a game?

Public reference sites publish WPA, Use Index, and run expectancy figures at the play level for regular-season and postseason games, usually inside a game’s play log or win probability chart. Many summaries display a win probability graph that runs from 50 percent at first pitch to 100 percent for the winner, with each steep climb or drop tied to a specific play and its WPA. Official league data through Statcast, deployed across all 30 ballparks in 2015, supplies the tracking layer that sits alongside these situational numbers, so a modern summary can show both what the ball did and what the play was worth.

Is WPA a good measure of player skill?

WPA is an excellent record of what a player contributed in specific games, but it is a weak measure of repeatable skill on its own. The metric depends heavily on context, because a hitter can only build a large total when teammates produce high-use chances for him to capitalize on. Two players with identical underlying performance can finish a season with very different WPA simply because of when their hits arrived. For that reason analysts pair WPA with context-neutral measures and read it as a story of outcomes rather than a forecast. Over a single game it reflects timing more than talent.

How do WPA, Use Index, and Run Expectancy relate?

They describe the same play at different distances. Run Expectancy values a situation only in runs for the current inning, ignoring the score and the clock. Use Index sets the stakes by measuring how much the moment could swing the final result. WPA delivers the verdict by recording the actual change in win probability. A useful shorthand is that WPA is roughly the run impact of a play scaled by how much that inning’s runs matter to the outcome, which is why an identical hit can carry a small WPA in the first inning and a huge WPA in a tie game in the ninth.

Related Reading

• MLB Game Summaries: Daily Baseball Reports & Box Scores (main pillar)
• Anatomy of an MLB Game: The Inning-by-Inning Timeline
• Best Practices for Accurate, Trustworthy MLB Recaps
• Game Summary vs. Box Score vs. Play-by-Play Explained
• How to Read an MLB Box Score: Every Stat Explained
• How to Write an MLB Game Recap That Readers Finish
• What Is an MLB Game Summary? A Guide to Baseball Recaps
• What MLB Recap and Live Score Services Cost in 2026

Sources

• Win probability added – https://en.wikipedia.org/wiki/Win_probability_added
• Use index – https://en.wikipedia.org/wiki/Leverage_index
• Run expectancy – https://en.wikipedia.org/wiki/Run_expectancy
• FanGraphs library, WPA – https://library.fangraphs.com/offense/wpa/
• FanGraphs library, Use Index – https://library.fangraphs.com/misc/li/
• MLB.com Statcast glossary – https://www.mlb.com/glossary/statcast
• SABR, The Story of Win Probability – https://sabr.org/journal/article/the-story-of-win-probability/
• Lindsey, An Investigation of Strategies in Baseball, Operations Research (INFORMS), 1963 – https://pubsonline.informs.org/doi/10.1287/opre.11.4.477

Cricket Match Scorecards: Test, ODI & T20 Results Database

{“@context”:”https://schema.org”,”@type”:”Article”,”headline”:”Advanced Stats in MLB Game Summaries: WPA, Leverage, and Run Expectancy”,”author”:{“@type”:”Organization”,”name”:”Editorial”},”publisher”:{“@type”:”Organization”,”name”:”Daily Match”},”about”:[“Win Probability Added”,”Leverage Index”,”Run Expectancy”],”inLanguage”:”en”}