Statcast Projected Isolated Power Performance in Predicting 2017 Performance

Last February, I took Statcast batted ball data to project Isolated Power performance.  The idea was that big in game power comes from hitting the ball hard and barreling up on it.  I lacked launch angle data, but considered batted ball types.  However, incorporating ground balls, fly balls, and line drives did not improve the model.  Regardless, a model knowing only average distance and barrel per at bat, you could fairly accurately predict Isolated Power performance.  Therefore, if a player underperformed according to the projection then you might expect a bounce back the following season.  However, I wondered whether there were reasons why a player would typically under or over perform the projection.

Looking at the 2016 season, it found the following players whose actual 2016 ISO was the most underperforming in comparison to the model projections:

	2016 ISO	2016 xISO	Diff
Miguel Cabrera	.247	.295	-.048
Josh Harrison	.105	.147	-.042
Brandon Belt	.199	.239	-.040
Howie Kendrick	.111	.149	-.038
Kendrys Morales	.204	.242	-.038
Buster Posey	.147	.184	-.037
Albert Pujols	.189	.226	-.037
Alex Gordon	.160	.197	-.037
Adeiny Hechavarria	.075	.109	-.034
Yonder Alonso	.114	.147	-.033
Troy Tulowitzki	.189	.222	-.033
Nick Markakis	.129	.161	-.032
Mitch Moreland	.189	.220	-.031
Yadier Molina	.120	.151	-.031
Adam Jones	.171	.201	-.030

Looking at this data, Kendrys Morales jumped out to me.  He was quickly scooped up by the Blue Jays in what seemed like a fairly curious move.  He had been a decent hitter for the Royals, but was not doing anything incredibly productive.  His lack of position also hurt roster flexibility.  A three year deal for a player like that seems a bit like folly.  At the time of looking at this model, I thought well maybe the Blue Jays think his ISO underperformed because of Kaufman Stadium.

Mid-season, you heard similar things about Adeiny Hechavarria when the Rays traded for him.  He was hitting the ball hard and barreling it, so perhaps the Rays thought they could better channel that into more productive hitting.  Anyway, how did these guys do in 2017 compared to 2016?

	'16 ISO	'17 ISO	Diff
Albert Pujols	.189	.145	-.044
Yadier Molina	.120	.166	.046
Miguel Cabrera	.247	.149	-.098
Kendrys Morales	.204	.196	-.008
Howie Kendrick	.111	.161	.050
Alex Gordon	.160	.107	-.053
Nick Markakis	.129	.110	-.019
Troy Tulowitzki	.189	.129	-.060
Mitch Moreland	.189	.197	.008
Adam Jones	.171	.181	.010
Buster Posey	.147	.142	-.005
Yonder Alonso	.114	.235	.121
Josh Harrison	.105	.160	.055
Brandon Belt	.199	.228	.029
Adeiny Hechavarria	.075	.145	.070
		32+	-.020
		31-	.047
		All	.010

One thing you will notice is that I reordered them by age.  Moreland and up are 2017 seasons played as 32 or older.  Jones and down are age 31 and younger.  What this paltry little sample seems to suggest is that underperforming your expected ISO is a major red flag for players in the mid to late 30s.  It is indicative of something else happening that is eroding performance.  However, for younger players, who are in less of a decline phase age-wise, show significant rebounding in performance.

What does this mean going forward?  Below are the players who most underperformed their expected Isolated Power performance:

Player	2017 ISO	2017 xISO	Diff
Miguel Cabrera	.149	.230	-.081
Mitch Moreland	.197	.248	-.051
Alex Gordon	.107	.157	-.051
Kyle Seager	.201	.248	-.047
Jose Peraza	.066	.113	-.047
Justin Turner	.208	.252	-.044
Matt Carpenter	.209	.250	-.041
Nicholas Castellanos	.218	.258	-.040
Jed Lowrie	.171	.209	-.038
Alcides Escobar	.107	.144	-.037
Albert Pujols	.145	.180	-.035
Chris Davis	.208	.241	-.033
Shin-Soo Choo	.162	.194	-.032
Dansby Swanson	.092	.124	-.032
Ian Kinsler	.176	.207	-.031
Jose Bautista	.164	.195	-.031
Hanley Ramirez	.188	.218	-.030
Nick Markakis	.110	.140	-.030
Yadier Molina	.166	.196	-.030
Joe Mauer	.112	.142	-.030

What we see above in this list is a lot of older players who failed to live up to the projections.  If the 2017 data is indicative of anything, this does not bode well for most of these guys.  Younger players on the list are a mix in availability like Nick Castellanos are supposedly available in trade or like Dansby Swanson are not available.  It is these players who we might expect as they age they refine their skills and are able to turn their barreling and distance into something more useful.

Chris Davis, for the Orioles interested readership, will be entering his age 32 season.  Above, that barely places him into the upper range that saw a major erosion in performance.  Molina and Kendrick were really the only two players who bounced back.  The others treaded water or further collapsed.  For those hoping for Davis to reclaim his past greatness, it is a weak indicator that perhaps that simply is unlikely to be in the cards this upcoming season.

Using Statcast to Project Isolated Power

Through my youth, there were basically only a few statistics one needed to know.  Homeruns, batting average, runs batted in, stolen bases, and, only if you were a bit wonky and had time on your hands, doubles and walks.  My youth was largely dictated by what someone back in the late 1800s who knew more about cricket than baseball thought newspaper readers would want to know.  In the past 16 years, though, we have experienced a renaissance as fans and clubs have begun to use computers, statistics, and approaches developed in other field to know more about the game.

Some of these new approaches required new technology.  One such approach is Statcast, which uses multiple cameras to identify elements such as players, the baseball, and a bat.  This is combined with radar data and the reward is a ton of data.  While one can use this data to evaluate pitchers, baserunners, and fielders, we will be using it in this column to discern ability in hitters.  Specifically, whether exit velocity of a batted ball can be related to the power metric, isolated power.  And, then, if one or two seasons of exit velocity data can be used to accurately project future performance.

Now, the first step in figuring out how useful these measurements might be is to compare them in season.  I only looked at player who had 300 plate appearances in both 2015 and 2016.  What we find is that average hit distance (forgive the error in the graphic below, it is average hit distance not home run distance) and barrel rate correlate very strongly with isolated power in the same year (p < 0.01 for both variables).  This means that these two ways to measure velocity and contact quality are related in-season to isolated power.  The regression model connecting those measurements to the metric isolated power was also significant (<0 .01="" p="">
Below is a graph comparing expected ISO with ISO for 2015 with the accompanying R2 value:

So all of this informs us that hit quality is connected to isolated power.  That should be obvious, but it is helpful to be able to see that here.  However, what we are really interested in is whether these values are meaningful from one year to the next.  In other words, is this simply a descriptive correlation or is it a predictive correlation. 

We will do a very simple comparison.  We will simply compare R2 values for expected ISO using the 2015 developed model vs. 2016's ISO.  This simple comparison will help show whether the formula using Statcast measurements better correlates with next season's values than simply using the actual ISO from the year before.  The comparison between 2015 ISO and 2016 ISO is not shown, but the R2 was 0.5026

What we find is that the Statcast method improves the predictive capability by about 15%.  That is remarkable, but is not earth shattering.  If your decision making process was simply finding the players with strong ISO, then this technique would help but it might take a decade or so for that to be able to be seen through the noise.  In general, I do not find this to be much of a silver bullet.  That said, it may be a hesitant flag for some players and suggest some players that should be expected to regress downward or upward.

Here is a list of players who the model thinks most underperformed.  In other words, who does this model think should have had a bigger 2016 than they actually did.

	2016 ISO	2016 xISO	Diff
Miguel Cabrera	.247	.295	.048
Josh Harrison	.105	.147	.042
Brandon Belt	.199	.239	.040
Howie Kendrick	.111	.149	.038
Kendrys Morales	.204	.242	.038
Buster Posey	.147	.184	.037
Albert Pujols	.189	.226	.037
Alex Gordon	.160	.197	.037
Adeiny Hechavarria	.075	.109	.034
Yonder Alonso	.114	.147	.033
Troy Tulowitzki	.189	.222	.033
Nick Markakis	.129	.161	.032
Mitch Moreland	.189	.220	.031
Yadier Molina	.120	.151	.031
Adam Jones	.171	.201	.030

One name that jumped out to me was Kendrys Morales.  He had a solid year last year, but the model thinks it should have been considerably better.  If the model better accounts for his talent, then we might see something closer to that expected isolated power.  It may well be that playing in Kansas City depressed his value a bit and some of his hard hit balls should have fallen in.  A different point of view would be that perhaps his isolated power was depressed because he is below average in converting singles into doubles.  That might explain why Pujols is up here as well.

Here is a list of players who the model thinks most overperformed:

Player	2016 ISO	2016 xISO	Diff
Brian Dozier	.278	.201	-.077
Nolan Arenado	.275	.224	-.051
Mookie Betts	.216	.170	-.046
Robinson Cano	.235	.190	-.045
Ryan Braun	.233	.189	-.044
Curtis Granderson	.228	.187	-.041
Jay Bruce	.256	.216	-.040
Edwin Encarnacion	.266	.229	-.037
Zack Cozart	.172	.137	-.035
Anthony Rizzo	.252	.217	-.035
Ben Zobrist	.174	.140	-.034
Carlos Santana	.239	.208	-.031
Didi Gregorius	.171	.140	-.031
Jose Bautista	.217	.187	-.030
Gregory Polanco	.205	.175	-.030
Josh Donaldson	.265	.235	-.030
Rougned Odor	.231	.201	-.030

I would have thought that the model would list speedster after speedster, guys who stretch singles into doubles.  That does not appear to be the case here.  Many of these players are rather plodding.  The closest Oriole on this list is Jonathan Schoop who comes in at a -.024, which is not a good thing to hear given how uneven and somewhat underwhelming his season was last year.

This made me wonder though about how things change over time.  For instance, is the over or under production from batted ball performance to expected batted ball performance a skill.  Are over producers always over producers.  What was remarkable was that the average difference between 2016's difference and 2014's difference was .011.  The greatest difference was .046.  This suggests that there is some element that I am missing.  The ability to over or under produce appears to be repeatable, so therefore likely having to do with a skill.  The next step is finding that skill.