Discover more from The Dynasty Dugout
Statcast 101: Expected Stats
Chris Clegg breaks down expected stats, how to use them , and what they mean.
Statcast data has changed the game and the way we evaluate players, well at least some people. The baseball community is pretty split on what matters when it comes to evaluating players as you’ll often hear a wide variety of opinions from just watching a Major League broadcast on television.
Looking for an edge is something we all do, whether it be in life, fantasy baseball, or actual MLB organizations. Finding and creating new metrics is constantly happening in the game we love, so seeing what matters and what doesn’t, especially when it comes to advanced metrics is important.
When doing offseason research, Fangraphs and Baseball Savant should both be go-to resources for you. But what if you find a metric on there that you just are not sure what it means? This article series will help answer those questions for you as we dive into statcast and other advanced metrics that you can use to analyze a player for fantasy baseball.
Statcast Data and What It Means
We often hear someone refer to a player’s statcast page and say it is full of red or full of blue. Here is what they mean when they refer to the statcast page or bubbles.
In the picture above, you can see that Ronald Acuña Jr. ranks exceptionally well in every hitting category. When a player ranks well in a category, you will see color and a percentile. If a player ranks highly, that data set will be red. As a player fades toward the 50th percentile, that color will fade to neutral and eventually blue. A slider will be blue on the far extreme if a player ranks poorly in a category.
It is easy to look at a statcast profile like Acuña’s full of red and automatically assume that the player is good. Duh, Ronald Acuña Jr. is good, more like the best player in baseball good. The same can be said for a player with a lot of blue in his profile. While it is likely true, it is worth understanding what each stat means and how it correlates to actual stats that matter for Fantasy Baseball.
Here is a player profile that is not so red. This player has a 2024 ADP near 100 largely due to speed and stolen bases.
When you first get to a player’s profile on Baseball Savant, you will find a nice picture of the player with the player’s position, whether they bat/throw right or left, the player’s height, weight, and age. If you move to the middle of the page, you will see MLB percentile rankings. This list shows several different statcast numbers and their percentile rankings among hitters or pitchers. Finally, to the far right of the page, you’ll see a hitter’s spray chart on just batted balls that became hits.
If you scroll down on the player’s profile page, you will find a player’s stats for both the current/previous season and the player’s career. Then you see the statcast data. This is ranked on the sliders at the very top of the page. This view gives you a nice overlook at how a player may have improved or regressed each season.
Right below the statistics is a nice chart. The default view is the percentage of each pitch type a player saw in a season. The pitch types are broken down by fastball, breaking ball, and offspeed. The pitch tracking section below the chart shows how a player performed based on pitch type.
This chart is handy for a lot of things. If you can click into the drop-down box labeled “PITCH %,” you can change the input to a variety of statistics, including all statcast data. In most cases, I switch “PITCH GROUP” to “ALL PITCHES,” which provides an excellent overview of how a player has performed. You can look at a player’s performance over the season, month, or game. If you are a visual person like me, you will likely find this page section very useful.
This is all great, but how did we get here? How is this tracked? Well, I am glad you asked. This is a Hawkeye camera if you have ever seen one of these things at the ballpark. There are 12 of these around the ballpark to track everything that happens during a baseball game.
Starting in 2015, MLB began using statcast and tracking every little detail of the game. It began using TrackMan, but in 2020, they switched to Hawkeye which is shown above.
The possibilities here are endless, and every year, Hawkeye software gets a little better and is able to track more. The newest iteration in 2023 is bat speed for all players, which has a large correlation in power metrics.
Before we get too into the weeds with Hawkeye, let’s talk about a couple of metrics, what they mean, and how to use them.
Expected Stats: What Are They and Are They Useful?
You often hear it quoted “Player X is overperforming his xBA, therefore regression is coming.” I have been guilty of this myself. I have written articles in the past about over and underperformers in expected stats. I look back and realize I probably had it all wrong.
Expected stats are like the shiny toys that everyone loves. People love quoting these like they are gospel and can predict a player’s performance will pick up or slow down based on these. But what is the intended purpose of expected stats and what do they tell us?
Expected Batting Average
Expected Batting Average (xBA) is a statcast metric that measures the probability that a batted ball will become a hit. Every batted ball is assigned an xBA on comparable hit balls in terms of exit velocity, launch angle, and on topped or weakly hit balls, sprint speed.
In the game feed picture above, you can see this played out. Each batted ball is assigned that xBA based on the factors mentioned in the previous paragraph. Using the expected outcomes of each player’s batted ball helps form their season-long xBA. Strikeout totals are factored into the equation, resulting in a player’s overall Expected Batting Average.
A player’s xBA for the season is calculated by taking the sum of all xBA from individual batted ball events. The sum is then divided by all batted ball events. After that, strikeouts are factored in which results in a season-long Expected Batting Average.
Expected Batting Average can help give us a better picture of a player’s true skill level than batting average itself. xBA removes defense from the equation, meaning the results are based more on the hitter’s skill level. Because hitters can influence exit velocity and launch angle to an extent, xBA is useful, but there are many more things that matter than strictly exit velocity and launch angle. Once the ball leaves the bat, a hitter has no control over the outcome.
Before 2023, Corey Seager seemed like an obvious breakout hitter if he could stay healthy. The shift was going away and his batted ball data was extremely good. Seager posted just a .245 batting average in 2022, but his expected batting average sat at .283. It was a sign of bad luck, but in large part due to the shift.
You know the end result, Seager went on to hit .327 this season and mash 33 home runs on his way to becoming World Series MVP.
On the flip side you see some players that constantly over or underperform their Expected Batting Average, which means it is not extremely useful for every player.
With doing a full study on every player since statcast began in 2015, data shows little predictive value with an R2 of .114.
Expected Batting Average can be a useful statcast metric to use when evaluating what should have happened. It does not tell us what we can expect in the future. Using xBA to predict a change in a player’s performance may not always be the smartest thing to do.
If you want to look at Expected Stats with more incorporated than just exit velocity and launch angle, check out Crosby Spencer’s article and leaderboard on our site:
Much like Expected Batting Average, the Expected Weighted On-base Average is calculated similarly. Statcast data like exit velocity and launch angle are used. Still, xwOBA takes it a step further and assigns each batted ball a single, double, triple, or home run probability based on the results from comparable batted balls since statcast was implemented in 2015.
All hit types are valued similarly for xwOBA as they are for wOBA. The formula for wOBA is: (unintentional BB factor x unintentional BB + HBP factor x HBP + 1B factor x 1B + 2B factor x 2B + 3B factor x 3B + HR factor x HR)/(AB + unintentional BB + SF + HBP). “Factor” that you see listed in the formula indicates the adjusted run expectancy of a batted ball event in the context of the whole season.
A player’s season-long xwOBA is calculated based on the batted ball data mentioned and factors such as walks, strikeouts, and hit-by-pitch. xwOBA is based on the batter’s quality of contact rather than actual outcomes and can be more useful.
xwOBA is less useful for Fantasy Baseball purposes in general. It is, however more useful in OBP leagues, which are becoming increasingly popular in the Fantasy Baseball community. Did a hitter’s xwOBA suggest they under or overperformed? It is possible to see positive or negative regression the following season.
xwOBA has a slightly stronger predictive value of a player’s future wOBA, than Expected Batting Average does. The r2 value for previous year xwOBA to wOBA is .218 using nine years worth of data points. Previous seasons wOBA is much less predictive with an r2 of .191. Despite that, it seems that year-over-year wOBA is easier to predict than batting average.
Expected Slugging Percentage
Again, as the previous two stats mentioned, Expected Slugging Percentage is a statcast metric calculated using launch angle, exit velocity, and on certain batted ball types, sprint speed. All hit types are calculated and valued similarly for xSLG as they are for standard slugging percentage. Doubles are worth twice as much, triples three times as much, and home runs four times as much as singles.
The formula for slugging percentage is: (1B + 2B*2 + 3B*3 + HR*4)/AB). Knowing the expected outcomes of each individual batted ball from a player helps formulate the player’s xSLG. Like the other statcast metrics, Expected Slugging Percentage is based on the quality of contact rather than actual outcomes.
With nine years of statcast data to this point, it is easy to compare individual batted balls. If a single batted ball is close to other batted balls based on exit velocity and launch angles, it is easy to quantify. If you have a batter ball that has an exit velocity of 108 miles per hour and a launch angle of 21 degrees, but it ends in a flyout, that batted ball gets a zero in the slugging percentage department. But, if that similar batted ball type goes for a home run more often than not, the xSLG will help give you a better idea of what happens more often than not on those similar batted balls.
Like xBA and xwOBA, Expected Slugging Percentage can be useful because it can indicate a player’s true skill. Removing defense from the equation and using factors that a hitter can influence like exit velocity and launch angle can give a better idea of the hitter’s skill level. It does a great job of explaining what should have happened or usually happens when a player has a similar batted ball event. If a player has experienced bad luck, you can expect their xSLG to be much higher than their actual slugging percentage.
Expected Stats Summary
Expected averages are not always perfect, but are sometimes helpful when evaluating hitters for Fantasy purposes. They are not predictive and were never designed to be. Expected stats fulfill their goal of doing what they were created to do which is to paint a bigger picture of a player’s actual performance.
According to Tom Tango, MLB Senior Database Architect of Stats, expected stats were designed to only be descriptive. If the goal was to be predictive, they would have been designed differently.
Expected stats tell you what the likely outcome should have been. They are not predictive so use them accordingly.