In this age of analytics, when we evaluate batters and try to project the future, we talk about exit velocity a lot. Exit velocity measures how hard the bat hits the ball, and the typical metric is average exit velocity. Average exit velocity is just what it says: compute the ball’s average exit velocity off the bat for all of a batter’s contact during a specific period (like a season).
The recent fangraphs article, “The Doomed Search for a Perfect Way To Interpret Exit Velocity” (Andrews), inspired me to re-examine the exit velocity for Astros’ batters in 2023. Average exit velocity is attractive as a measurement, because a relatively small sample can predict how hard the batter will hit the ball in the future. However, average EV is not ideal for this role. As the article points out, over a short period like a season, the measure can be “noisy,” meaning that irrelevant or less meaningful data may obscure how good the batter is at hitting the ball hard.
One cause of this noise is the inclusion of weakly hit contact in the average. Every batter has some weak contact, and it probably doesn’t provide a meaningful signal about the batter’s hitting ability. One easy remedy for this problem is to narrow the sample to balls which are hit more solidly. An available alternative is termed “Best Speed,” which is simply the average EV for the >50% of highest velocity balls hit by the batter. The Best Speed Average EV is not available on the Statcast dashboard, but it can be found on the “custom leaderboard” tab at Baseball Savant.
Statcast expert Tom Tango says he ignores Average Exit Velocity and prefers the Best Speed version. He prepared a graph which shows that the 50th percentile of a player’s exit velocity is a starting point for correlating average exit velocity with next year’s wOBAcon (wOBA on contact).
My Astros’ analysis is based on comparing the top hitters’ average exit velocity with the best speed version of the metric. The difference between the Average EV and the Best Speed EV indicates which batters improve the most by using the >50% average. On average, MLB batters have a Best Speed about 10.5 mph higher than Average EV. I have also included “expected” Isolated Power (x-ISO), which is a comparative statistic for extra base power. ISO is Slugging Minus Batting Average.
As part of my analysis, I downloaded the Custom Leaderboard data, and tried to glean some additional insight with the statistical coefficient of determination (R square). The Best Speed metric exhibits a better relationship to ISO than Average Exit Velocity. Average EV explains 57% of the variation in x-ISO, while Best Speed explains 65% of the variation. Best Speed has virtually no relationship to the Statcast metric “Sweet Spot,” but it explains 66% of the variation in “Barrels.”
Barrels measures contact which produces an ideal combination of launch angle and exit velocity. (Not surprisingly, barrels frequently become home runs.) We probably should prefer the metric Barrels to Sweet Spot.
Alvarez, Diaz, Abreu and Altuve show Best Speed measures which exceed the Avg. EV by a margin that is well above average. Yordan Alvarez who is among the top 10 in MLB Average EV moves up a few slots on the Best Speed leaderboard. Diaz moves to 2d, just behind Alveras and past Tucker, on the Best Speed Astros ranking.
Altuve is a high contact hitter who is not normally viewed as a high exit velocity hitter—as evidenced by his sub-par 86 mph Average EV. But he has a substantially higher than average increase to EV when it is measured with Best Speed. Altuve’s high contact ability often results in weak contact, but that is not representative of his ability to hit the ball hard.
In my mind, Abreu is the most interesting case. And if you are the hopeful type, the comparison may provide some degree of optimism about Jose Abreu’s ability to improve his performance next season.
One of the warning signs which alarmed fans early in the season was Abreu’s abrupt decline in exit velocity. In mid-April, Abreu had a career worst 87 mph Average EV, which was 5.2 mph less than his previous year’s 92.2 mph. By June 1, his 87.6 mph Average EV was still far below his norm. After he was treated for a back injury around mid-season, Abreu’s power resumed. He ended up hitting 18 home runs on the season—more than 2022. His first half ISO of .108 approximately doubled to .209 in the second half.
Abreu’s 2023 season Avg. EV ends up at 98 mph (league average), but his late season improvement was close to his 2022 EV performance. In Sept./Oct., Abreu’s Average EV was 90.7 mph, second on the team to Yordan Alvarez at 94.6 mph. His ISO was .299 and his OPS was .835 for the same period.
Abreu’s second half wRC+ was 109, which happens to be the same as the ZIPS projection for full season 2024. So what does “Best Speed” have to do with this?
Abreu’s Best Speed EV for 2023 was an above average 100.82 mph. As noted above, the difference between his Best Speed and Avg. EV was the highest on the team, and well above average. My surmise is that the “bottom 50%” of Abreu’s EV contact is largely comprised of weak hitting in the first half of the season when his Avg. EV was alarmingly low.
We don’t know if Abreu’s early season lack of power was caused by his desire to play through a back injury. But it’s not unreasonable to project a significant rebound in Abreu’s power if the Astros’ medical staff has controlled his back issue.
And that’s what I mean by “the optimistic case” for an improvement in Abreu’s batting performance next season.