Going Deep: Interpreting BABIP and Players with Negative Regression on the Horizon

I’m back with Part Two of my articles of interpreting BABIP, this time we are focusing on averages that are much higher than a player’s normal range. This process is easier than explaining a player with a very low BABIP as most of the time it’s simply luck or being skewed by a hot streak.

First off, a BABIP over .400 is almost assuredly going to regress as there hasn’t been a season-long BABIP over .400 since 2002. Heck every year there is only two to three BABIPs that even end up over .370. Cody Bellinger is an excellent hitter with the potential to be one of the league’s best, but his .397 BABIP is much much higher than his career norm of right around .300.

This is the important part of looking at BABIP when the numbers are super high, it’s less about asking IF they will regress (they will), it’s more about how far will they regress.

The process is incredibly similar to last week with a few added wrinkles. Unfortunately, I’m going to make you do some math but don’t worry it’s like middle school level math so it is anything too crazy.

• First thing I do is determine the player’s normal BABIP range. If they are a young player or have a wide range of BABIPs I’ll usually just go with what Fangraph’s Depth Charts has listed.
• Once I have that baseline BABIP I regress the player’s current output to their normal BABIP using a simple proportion. This will pump out the new regressed stat. Please note that this is not meant to be ANYTHING set in stone. This is simply a quick and dirty way to get a rough idea of what their numbers might regress to.
• If those rough adjusted numbers look radically different than the player’s normal or projected output for the season I for any articles that might imply a swing change. You’re looking for things like actual mechanical changes (holding their hands lower, removed/added a leg kick, that sort of thing).
• Then I move on to the batted-ball data to see if there seems to be a change in approach. Mainly you’re looking to see if the player is hitting more or fewer line drives and fly balls. Line drives will boost BABIP while a high BABIP with a ton of fly balls makes me suspicious. It’s also worth noting the player’s HR/FB% and how often that player is hitting doubles. If a player’s HR/FB% is down but his doubles are up, then that can inflate BABIP as a home run is not considered a ball in play.
• Finally, I check the player’s xStats at Baseball Savant. Just like the previous section, I want to look at xBA, xSLG, and xwOBA to see if the batted-ball luck tells us anything. Then I also check out if there are any notable changes in exit velocity, launch angle, BBL%, and hard-hit rate as increases in these stats can be indicative of a swing/approach change or a player putting it all together.

The key to all of this process is that you are looking for any evidence to support those new higher stats. For instance, Bellinger is currently hitting for a 34.9 LD%. That represents a 15% increase from 2018 and would have led the league last year by 2%. If Bellinger’s LD% held up all season long it would be third-highest single-season LD% of ALL TIME. To be skeptical of that number isn’t a diss on Bellinger, I’d be skeptical if I saw that for any player. Since we’re using Bellinger as the example let’s start our player analysis by running him through the process.

 

Cody Bellinger, 1B/OF, Dodgers

 

Season	BABIP	AVG	HRs	OBP	OPS
2019 Current	.397	.433	13	.509	1.42
2019 Regressed	.310	.338	10	.397	1.10

1. As you can see after running those base stats through the proportions that even if he was having a more normal BABIP season he would still be mashing the ball. Again I cannot emphasize this enough that you do not take these regressed numbers as gospel, it’s just to normalize his numbers to a reasonable range.
2. It turns out Bellinger has made a lot of changes to his swing especially when it comes to hitting against lefties. A large portion of his success has been fueled by hitting over .300 against lefties in 2019 with a mere .250 BABIP. According to this article by Zach Kram over at The Ringer, the Dodgers’ batting coaches have got him standing more loosely in the box as opposed to his previous stance which was very straight and rigid. This is the kind of thing we’re looking for. If this means Bellinger is more comfortable at the plate this could give us a reason to suspect some of that new production is sustainable.
3. We talked a bit about his LD% already, I don’t expect that to hold up. In the past, he’s been a below average LD hitter even during his blazing hot rookie year. Yet, given his new stance and a young stud player’s natural skills progression, I don’t think it is going to fall all the way back down to last year’s 19.0% either. Somewhere in the upper 20s around 27.0% or 28.0% is reasonable. That’s still an elite LD% and portends really great things to come.
4. Along with that LD% he has a 34.9 FB% which doesn’t really mesh with how many HRs he’s hit so far this season. His current 43.3 HR/FB% would be the second-highest of all time if it continued throughout the season. So at the moment, his power looks great despite fewer flyballs because of a historic HR/FB%. You can’t bank on that continuing. My guess is that over the course of the season, his LD% and HR/FB% come back down while his FB% increases.
5. Let’s talk about his spray chart for a second. As mentioned in The Ringer article, Bellinger is pulling the ball more than ever to the tune of near 10.0% increase overall. This is fantastic because when he pulls the ball in the air, it goes. This might be what has me the most excited about Bellinger, it reminds me a lot of the approach changes Francisco Lindor and Jose Ramirez made during their breakout years.
6. Here’s the nutsy-cuckoo part. All the xStats almost completely mirror his current production.
   

   Season	AVG	SLG	wOBA
   2019 Actual	.433	.918	.560
   2019 xStats	.417	.819	.540

   That makes it a bit harder to be skeptical of Bellinger’s season so far but of course, I still expect those numbers to come back down to earth. The xStats confirm the validity of Bellinger’s new approach, negating some poetential regression.  He’s also barreling the ball at nearly double the rate he did last year and while his launch angle is down (hence more line drives, fewer fly balls) his exit velocity is up to 93.6 which puts him in the top 5% of the league.

It’s entirely plausible that Bellinger’s AVG ends up somewhere between .280 and .300 (especially if he continues to swing at just 21.1% of pitches outside the zone, limiting his strikeouts) which would represent a career high. Given how hard he is hitting the ball, combined with how often he is pulling it in the air, I really do think 40+ HRs is well within range. His numbers are going to come back to down to actual human levels at some point but I do think we are seeing a new improved baseline.

 

Paul DeJong, SS, Cardinals

 

Season	BABIP	AVG	HRs	OBP	OPS
2019 Current	.385	.340	5	.393	1.005
2019 Regressed	.312	.275	4	.318	.814

DeJong has looked like the player we’ve been waiting to break out for years now. Unfortunately, on the surface, it appears like solid incremental skills progression with a healthy chunk of luck.

1. Proportional regression pegs DeJong as closer to a .275 hitter with 4 HRs and a much more appropriate OBP for a hitter with a 6.3 BB%. Considering he was a .243 hitter last year that’s a great improvement. Is the regression fair or do the underlying numbers know something the proportions don’t? Are his numbers somewhere in between the BABIP fueled numbers and his regressed numbers? I think so.
2. Nick Gerli wrote about DeJong and his 2019 success earlier this week and he does a fantastic job explaining how his mechanics are tighter and more consistent this year. I cannot recommend the piece enough as Nick does a great job really illustrating much of DeJong’s adjustments.
3. We start with DeJong’s greatly improved K% (17.9% which is down from 25.0% last year), a big part of what held him back in years past. Some of that K% gain is due to some regression as he made solid improvements in his O-Swing% (down 2%) and Contact% (up 2%) but that’s likely not enough to explain a 7.1% drop in his K%. That will probably settle back into the low 20s as he has shown improvement, just not that much improvement.
4. So far he’s made a 2.0% improvement in his LD% while reducing his GB% by nearly the same amount which certainly could help explain the bump in AVG to .275. At 26.5% DeJong is an above average LD hitter and so I wouldn’t be shocked if in reality, his AVG ends up closer to .280 or .285. Nick also talks in his piece about how DeJong is pulling the ball in the air which could boost his power output, evidenced by the 5.6% drop in his OPPO%. I see a bunch of incremental improvements as opposed to the signs of a full-on breakout.
5. Check out what xStats has to say about his batted ball luck so far this season.
   

   Season	AVG	SLG	wOBA
   2019 Actual	.340	.612	.425
   2019 xStats	.307	.570	.390

   The xStats definitely say that DeJong has gotten lucky so far but they still peg him as .390 wOBA hitter, which is elite. Exit Velocity and Launch Angle are consistent with last year but he is barreling the ball nearly 3.0% more up to 12.0% from 9.0% last year. This reinforces that he is definitely improving, just not by leaps and bounds.

Overall I’m actually really high on DeJong this year. If these baseline numbers continue all throughout the year, this sets the table for the real DeJong breakout in 2020. I see something along the lines of a .280 AVG with 25 to 30 HRs while hitting in a pretty good lineup the rest of 2018.

Andrew Benintendi, OF, Red Sox

 

Season	BABIP	AVG	HRs	OBP	OPS
2019 Current	.369	.306	2	.384	.866
2019 Regressed	.321	.268	1	.334	.753

1. Benintendi is an interesting case. I don’t buy the regression at all but they still give me pause. As you will soon see very little about his profile has changed and what has changed I find troubling. There are also a lot of signs that his numbers are sustainable so let’s take and see what we think.

2. I’m not seeing anything out there about any swing changes for Benintendi.

3. This is where I see some things I don’t really like. Benintendi’s K% is up 4.0% all the way up to 20.0% which is a fine number but you never like seeing a player’s K% jump that much. It supported by an 8.0% jump in his O-Swing% to 36.4% and 1.9% jump in his SwStr%. This makes me tend to think that Benintendi’s AVG is a bit of a luck-fueled mirage so far this season.  Add in the fact that his FB% is up 6.4% to 41.8% and I’ll admit I’m worried. Normally adding more fly balls is a good thing (groundballs suck!) but only if it brings with it a corresponding increase in power and so far that has not been the case.

His HR/FB% is down to 7.1% and his ISO is essentially the same as last year. You want that FB% increase to come at the expense of a player’s GB%, but a lot of those flyballs came at the expense of his LD%.

4. Then just when you’ve got it all figured out the xStats tell a totally different story.

Season	AVG	SLG	wOBA
2019 Actual	.306	.482	.371
2019 xStats	.298	.522	.385

In addition, his Exit Velocity is consistent with last year at 88.0 MPH and his Launch Angle is up to 14.7 degrees. That feels like someone who should be hitting for more power especially when you factor in his BBL% is up to 10.3%.  It’s probably somewhere in between. If the new fly ball tendencies hold up and he continues to hit the ball hard, Benintendi should produce something like last year with more HRs, so say a .280 to .290 AVG with 20 to 25 HRs. With more data we’ll have a better idea of what kind of hitter he’ll be moving forward.

If you would like some specific interpretation on a player, please feel free to name one in the comments or reach out to me on Twitter @DanielJPort and I will try to get my take on as many I can!

Photo by Brian Rothmuller/Icon Sportswire