Statcast xStats: Using Sabermetrics for 2023 Fantasy Baseball

Statcast is a valuable tool for fantasy baseball analysis, and it can be easy to look at a stat called "Expected Batting Average" and blindly use it as your projection moving forward. Of course, proper use of these metrics is a little bit more nuanced than that.

First, a disclaimer: This article is about the "Expected Stats" found on Baseball Savant. It is not about the various "xStats" developed by fantasy analysts such as Mike Podhorzer or used in projection systems such as Ariel Cohen's ATC. Those tools have value, but any attempt at an in-depth analysis of them would involve far more math than this column is intended to get into.

Expected Statistics is yet another stat accessible via the Leaderboards tab on Baseball Savant. You can sort by players (including hitters and pitchers) and by team. With that out of the way, let's begin by identifying what the Expected Metrics are and how they work.

• Fantasy baseball injury reports
• Fantasy baseball trade analyzer
• Daily MLB starting lineups for fantasy baseball
• Fantasy baseball BvP matchups data (Batter vs. Pitcher)
• Fantasy baseball PvB matchups data (Pitcher vs. Batter)
• Who should I start? Fantasy baseball player comparisons
• Fantasy baseball closer depth charts, bullpens, saves
• Fantasy Baseball live scoreboard, daily leaderboards

How To Use Statcast's Expected Metrics In Fantasy

The first is xBA, or Expected Batting Average. This statistic is calculated using Hit Probability, itself a stat measuring how often a batted ball with a particular exit velocity and launch angle has fallen in for a hit since Statcast was introduced in 2015.

For example, a line drive to the outfield that has historically fallen in for a hit 80 percent of the time counts as 80% of a hit by Hit Probability. xBA is simply a batting average produced using Hit Probability, actual K%, and official ABs. If you play in a traditional 5X5 roto league, this is the xStat you'll probably use the most.

As of January 2019, the Hit Probability formula was modified to include the batter's Statcast Sprint Speed, more accurately representing his ability to beat out a ground ball. That said, the adjustment feels like it may be too light in certain circumstances, so you may still want to make a slight adjustment upward for true jackrabbits.

Next up is the Expected Slugging Percentage, or xSLG. It is calculated in the same manner as xBA, except that each batted ball is weighted according to its probability of being a single, double, triple, or home run instead of just a hit. If your league counts slugging percentage, you might get good use out of this stat.

Finally, we have Expected Weighted On Base Average, or xwOBA. It is calculated the same way xSLG is, except real-world walks and HBP are added to the equation. Each result is also assigned a linear weight with more math than the simple multiplication used to calculate the slugging percentage. It offers the most real-world value but doesn't translate that well to fantasy unless you play in a realistic points format.

The principal value of all three metrics is to take both luck and defense (and therefore actual results) out of the picture, allowing a player to be judged solely on his contact quality.

Putting xBA to Work

We'll assume that you play 5x5 roto and stick with the simpler xBA from here on out. Generally speaking, a player who posts a higher xBA than the batting average would be expected to improve his average moving forward, while the opposite is true if a player's average is higher than his xBA.

Baseball Savant's Leaderboards allow you to sort players by the difference between their BA and xBA, so finding some samples is easy. The three biggest xBA overachievers are all fantasy relevant: Paul Goldschmidt (.317 vs. .261 xBA, a 56-point differential), Xander Bogaerts (.307 vs. .259, 48 points), and Jeff McNeil (.326 vs. 280, 46 points). Goldschmidt figures to put up solid numbers based on his power and role in the middle of a strong St. Louis lineup but expecting a repeat of last year's average would be unwise.

Goldschmidt's Statcast Sprint Speed of 26.2 ft./sec is below average, too, so he's not beating out many grounders. Both Bogaerts (27.9 ft./sec) and McNeil (27.4 ft./sec) have above-average wheels, per Statcast, but aren't fast enough to overcome the differentials above. Since both are generally rostered for their batting average in fantasy, it may pay to fade them this year.

Bogaerts' power production is best compared to a random number generator, and he won't pile up runs scored if he isn't getting on base. McNeil has never contributed much power and may find it difficult to accumulate counting stats as the fifth hitter in the Mets lineup.

Going the other way, Carlos Santana posted the best positive differential with a .253 xBA against an actual batting average of .202. These advanced stats don't understand that certain players are more susceptible than others to the shift, so you should check those numbers before you blindly project improvement. Santana was shifted nearly every time up and hit just .200 against it last year. He's also slow with a Statcast Sprint of 25.7 ft./sec, so he isn't legging out many singles.

Pitchers illustrate another problem with xBA. Tony Gonsolin was the "luckiest" pitcher according to the metric in 2022, posting an xBA of .220 despite a batting average against of .172. The metric doesn't consider a defense behind a pitcher, however, so outstanding glovework could help sustain such a gap moving forward. The Dodgers were a solid defensive club overall last season with eight OAA, ranking ninth in the league. However, Gonsolin was the beneficiary of all of them, something that probably won't repeat in 2023.

League-wide, major leaguers posted a .243 batting average and .240 xBA in 2022. It was the first time the differential increased year over year, with both 2021 and 2019 coming in at two points. While exciting, this suggests the technology isn't foolproof. It is always best to utilize Statcast Expected Stats as part of a broader analysis, rather than your sole data point.

Conclusion

In summation, Expected Stats allow you to evaluate a player's performance based on his exit velocity and launch angle, taking variables such as the opposing defense out of the calculus. This can give you a better sense of a player's true talent level, but there are limitations on what you can do with it.