Are There Chronic wOBA Over- and Under-Performers?

You know the third base pool is loaded at the top when there are three players at the position who returned more value than Alex Bregman did in 2019 5×5 Roto leagues. Yet, on draft day 2020, owners are likely to at least consider making the Astros’ 25-year-old the first third baseman taken, ahead of Rafael Devers, Anthony Rendon and Nolan Arenado. To this point, in the currently-under-way Pitcher List Experts Mocks, Bregman is the only third base-eligible player to be taken within the first 14 picks in all three drafts.

It’s not hard to see why. This season, he maintained his elite contact and plate discipline skills while tacking on 10 home runs, nine RBIs and 17 runs to his 2018 totals. In 2020, he would appear to be primed for another batting average around .290, and with a spot in the heart of the Astros’ order, he could clear the hurdles of 110-plus RBIs and runs yet again.

There is just one thing that has me putting the brakes on my own case of Bregmania. His Statcast power indicators are far from elite, and they don’t mesh with my idea of a 40-homer hitter’s profile. In 2019, his average exit velocity on flyballs and line drives was 92.3 mph, which ranked 114th out of 175 hitters who had at least 300 batted balls. His 37.5 percent hard-hit rate and 3.8 percent barrel rate (per plate appearance) were both outside of the top 100 as well.

Some of Bregman’s outsized power production can be chalked up to launch angle, as his 31.5 percent ground ball rate was the eighth-lowest among qualified hitters. Still, it seemed like there must be another explanation for his 41 homers and .296 ISO, and it turns out that there is. Bregman’s .296 Avg exceeded his xBA by 24 points, and his .592 SLG was 121 points higher than his xSLG. Yet I am not quite ready to peg Bregman as a regression candidate. In 2018, there was a similar, though less pronounced, pattern. He outperformed his xBA by 15 points and his xSLG by 64 points.

This, in combination with a similar discovery about Eduardo Escobar, made me wonder if certain players are somehow prone to overachieving their expected stats. I have taken a first pass at addressing this question by regressing the difference between a hitter’s actual wOBA and expected wOBA against the same difference from the previous season. I included all hitters going back to 2015 who had seen at least 2000 pitches in consecutive seasons and wound up with a pool of 322 hitters.

During this period, there has been a statistically significant relationship (R-squared = 0.17, p < .0001) between a hitter’s wOBA – xwOBA from a given season and the same differential from the previous season. The correlation coefficient for the previous season is 0.43, so on one hand, we can expect a hitter’s wOBA – xwOBA differential to regress towards zero the following season. However, a large differential like Bregman’s .049 from this season would typically only shrink by a little more than half. (The constant, incidentally, is less than .0002.)

Bregman is not alone in having consecutive seasons with a wOBA that was at least 20 points higher than his xwOBA. Of the 322 hitter seasons included here, 24 of them involved overperforming xwOBA by that magnitude for a second season in a row. However, 14 of those 24 cases represented just five players who had overperformed in at least three consecutive seasons. Charlie Blackmon and Arenado have overperformed their xwOBA by at least 20 points for five seasons running. Trevor Story has been in the same club for the last three seasons. (Noticing a trend yet?) Xander Bogaerts overperformed his xwOBA by at least 20 points in 2015, 2016 and 2017, and Didi Gregorious did the same in 2016, 2017 and 2018.

The Rockies and Red Sox are the teams with the two highest home BABIPs since 2015, as their respective parks have been among the most advantageous for hits in general and for doubles in particular. This creates a common link to connect Blackmon, Arenado, Story and Bogaerts. This explanation leaves Gregorius out, and it also makes it harder to explain how Bregman and Escobar have been able to overachieve their xwOBAs at this level for both 2018 and 2019. Even without a clear explanation, given that Bregman’s wOBA was 49 points above his xwOBA this season, regression may still only take him back to being an overachiever in the 20-to-25 point range in 2020. He has had an xwOBA in the .370s in each of the last two seasons, so he can be drafted with the expectation that he could be close to the .400 mark in 2020. That still renders him worthy of a late first round or early second round pick.

Escobar did not overachieve as dramatically as Bregman in either 2018 or 2019, so while he appears to be a candidate to overachieve in 2020, it may be by roughly 10 points instead of by 24 points, which was the case this season. Jonathan Villar, who overachieved his xwOBA by 27 points this year and by 21 points in 2018, is in a similar position. Both he and Escobar appear to be due for some minor regression. This, of course, does not take into account any changes in productivity that might occur if MLB changes the ball again.

It stands to reason that if there are some repeat overperformers, that there are some repeat underperformers as well. In this pool of hitters, there have been 13 instances of someone underachieving their xwOBA by at least 20 points on multiple occasions. Joe Mauer (2015-18) and Alex Gordon (2016-18) were the only players to do so more than twice. The only hitter to underachieve in both 2018 and 2019 was Marcell Ozuna. His .327 wOBA in 2018 was 32 points short of his xwOBA, while this season’s .340 wOBA was 40 points below his xwOBA. Again, without an explanation, it is hard to know if Ozuna has been incredibly unlucky or if there is some reason why he has fallen well below his xwOBA in successive seasons. The examples of Mauer and Gordon (who underachieved again in 2019, but only by nine points) suggest that there are some players for whom xwOBA may not be a useful indicator of future performance. It is probably safe to say, though, that we should not expect Ozuna to be anywhere close to a .380 wOBA hitter in 2020, just because his profile suggested he should have been one in 2019.

The graph also shows that there have not been many players who substantially overachieved in one season and then substantially underachieved in the following season, and vice versa. Lorenzo Cain overperformed his xwOBA by 17 points in 2018, only to underperform it by 30 points in 2019. Maybe luck, as well as thumb and knee issues, played a role in his subpar performance. Carlos Santana overperformed his xwOBA by 19 points this year after having had underperformed by at least 13 points in each of the previous four seasons. Fantasy owners are probably looking for Santana to regress after he had a career year in his age-33 season, but now we have yet another reason to not expect a repeat of his 2019 season.

Addendum: I did not consider the role that pulled flyball rate could play in wOBA – xwOBA, but fortunately, Jordan Rosenblum has. Check out his excellent work here. His findings go a long way towards explaining the outlier status of Bregman, Escobar, Mauer and Gordon and likely many others.