How win probability works in baseball

Win probability — often called win expectancy, or WE — is the single number that answers “who’s more likely to win right now?” It’s the estimated chance a team has of winning the game at one specific moment, expressed as a percentage from 0 to 100.

Because one team’s gain is the other’s loss, the two figures always add up to 100%: if the home team sits at 68%, the visitors are at 32%. At the first pitch of an evenly matched game the home side starts a hair above 50% thanks to home-field advantage, and by the final out the winner has climbed to 100% while the loser has fallen to 0. Everything interesting happens in between.

Win probability isn’t a prediction of the final score — it’s a live read on how much danger or safety a team is in at that instant, given everything that has happened so far. Understanding that one idea unlocks a whole layer of the game.

The number doesn’t come from a hunch — it comes from history. To find a team’s win expectancy, the model looks at the exact game state and asks: across thousands of real games in the same situation, how often did the team go on to win? If teams trailing by one run in the bottom of the seventh with a runner on second and one out won 44% of the time historically, that’s the win probability.

The inputs that define the “situation” are the score difference, the inning, the number of outs, which bases are occupied, and whether the team is home or away — plus the overall run-scoring environment of the era. Feed those into a decade or so of comparable games and the winning percentage of all those teams becomes your estimate.

Modern versions layer in team and pitcher strength, but the core is beautifully simple: match the moment to the past, and count how it usually ends.

Not every input carries equal weight, and knowing which ones dominate helps you feel the swing before the graph even updates.

• Score, scaled by time. A one-run lead means little in the second inning and almost everything in the ninth. The same gap grows more decisive with every out that passes, because there’s less time left to erase it.
• Base-out state. Runners in scoring position with fewer outs pump up the offense’s chances; two outs and empty bases deflate them. The classic contrast: a one-run lead in the bottom of the ninth is far tenser than a one-run lead in the top of the third.
• Home-field. The home team bats last, so it always carries a small structural edge in tie or trailing spots late.

Put together, these are why a single pitch in the ninth can move win probability by 30 or 40 points while an identical pitch in the third barely nudges it.

Most win-probability tools plot the whole game as a single line, and once you know the shape you can reconstruct a game you never saw. The vertical axis is one team’s chance of winning, 0% to 100%; the horizontal axis is time, first inning to last.

Three things to look for:

• The line rises when the batting team helps itself and falls when it doesn’t. A near-vertical jump marks a big play — a home run, a bases-clearing double, a double play that snuffed out a rally.
• Long, flat stretches are low-drama innings where nothing much was at stake.
• The line always ends pinned at 100% or 0%, because by the final out the result is certain.

A game with a jagged, back-and-forth line was a nail-biter; a line that drifts smoothly to one side was a comfortable win. The chart is the story of the game in one glance.

A scoreboard tells you what happened; win probability tells you how much it mattered. Down 4–2 sounds the same in the third and the ninth, but those are wildly different situations — roughly a coin flip in one, near-hopeless in the other. The raw score can’t distinguish them. Win probability can.

That’s the whole appeal for fans who want more than the line score. It turns a two-run deficit into a real number you can feel, flags the moment a routine-looking game quietly tipped, and lets you glance at your phone and know instantly whether it’s time to pay attention.

It’s also the foundation for two deeper ideas: leverage index, which measures how much a moment could swing, and win probability added, which credits players for the swings they cause.

Once you start thinking in win probability, a plain scoreboard feels thin. That’s exactly why TwentySeven is built around it. Open any MLB game and you get the full win-probability curve — the same story-in-one-line chart described above — recalculated on every single pitch, alongside the run differential and the plays that actually decided the outcome.

You don’t even have to keep the app open. Pin a game to your Lock Screen with Live Activities and a live win-probability mini-graph updates on every pitch, right there in the Dynamic Island, so a glance tells you exactly where things stand.

From here, two ideas go deeper: read how leverage index measures the tension of a single moment, and how win probability added turns those swings into a player-by-player impact score.