In baseball statistics, LI stands for Leverage Index. This statistic is a quantification of how important a specific play is within the context of the game as a whole.
While every baseball fan has an intuitive sense of this concept, LI adds the step of assigning specific values to a play so we can confidently state which moments are most important in any one game, series, or season.
Working from this concept, there are a number of ways in which LI can be used to gain a better understanding of the game and its players.
The Concept of LI
It’s long been understood that some moments are more important than others in a game, but quantifying that importance was an impossible task. These days, thanks to the use of Win Expectancy, it’s possible to assign leverage values and determine which plays within a game actually are the most impactful.
In order to understand LI, you need to understand win expectancy, as the two are closely related. Win expectancy, as you might guess, is the odds of each team winning the game at any point in time.
For instance, the home team might have a .55 chance to win the game before it starts, while the road team has a .45 chance. The total of the two win expectancies will always equal 1. As the game progresses, these numbers will move up and down based on the score, the inning, and the base-out state within that inning.
Leverage index comes into play as the opportunity to swing the win expectancy in one direction or another. A high leverage at-bat is one where the win expectancy has the potential to move dramatically based on the outcome. The statistic is scaled to set 1 as an average-leverage situation.
When the bases are loaded late in a close game, you can be sure the LI will be high. On the other hand, a leadoff batter stepping to the plate to start a game is going to have a low leverage at-bat, since nothing that can happen in that plate appearance will dramatically swing the win expectancy toward either team.
Why is LI Important?
One of leverage index’s most useful applications is to determine which players are being used by their manager in the most important situations. This relates specifically to two categories of players– pinch hitters and relief pitchers. As managers can directly control when these players take the field, it says a lot about which players are trusted the most when you look at who is used in the highest leverage situations.
Whether it is a key relief pitcher brought in to get out of a jam, or a pinch hitter used when runners are on base late in the game, the team’s best subs tend to appear when leverage is the highest.
How is LI Calculated?
Leverage index is not a statistic that you would want to calculate on your own. Since the game is constantly moving and win expectancy is constantly changing, it wouldn’t be practical to do these calculations live. As long as you understand the underlying components that contribute to leverage, such as base-out states and overall win expectancy, you should have everything you need to understand this stat.
To see how leverage works and to see how win expectancies evolve over the course of a game, the scoreboard at FanGraphs is a great place to go. You can see the leverage index for any past game that you’d like to review, and you can even watch it unfold live if you follow along while a game is in progress.
If you are watching a game on TV, try monitoring the leverage index using this scoreboard to gain a better understanding of which situations mean the most in each game.
What is a Good LI?
With LI, it isn’t really about ‘good or bad’, but rather about separating high leverage situations from low leverage situations.
The high-leverage situations are those where the outcome of the game truly hangs in the balance. On the other end of the scale, low-leverage situations are the least important moments of a game, when the win expectancy is unlikely to swing dramatically in one direction or another.
As you look at LI, remember that 1.0 is average leverage. From there, the moments get more important as the numbers get higher, with anything above 2.0 being considered a high leverage situation.
Once you become more familiar with this stat, you will realize that high leverage situations are far less common than low leverage situations. In fact, most of the at-bats that take place during a game count as low leverage situations, with only a small percentage moving up above 2.0.
What are the Problems with LI?
One of the problems with Leverage Index is the temptation to use it to determine which players are ‘clutch’.
Generally speaking, most baseball statisticians agree that clutch performance is not a true skill which will be replicated over and over again. More likely, a player who is performing particularly well in clutch situations is benefitting from a small sample and those performances will average out over time. In the end, with a large enough sample, most players tend to perform near their career average levels regardless of leverage.