Finding Possible Hitter Injuries Using xwOBA

I’ve always been a week or two behind evaluating hitters so I continue trying to find ways to gain an edge. Today’s stab in the dark is trying to see if StatCast data can determine if a player is hurt. A recent study of mine found no correlation between playing through an injury and exceeding their projection in the next season. Instead of looking at preseason projections, I’m going to go a little more in-season today and determine how much of an impact an injury has on a hitter’s in-season production.

As Al showed, there are a ton of metric available StatCast batted ball metrics to use. I started down the path of using several of them but quickly found there is more to injuries than just power. A hitter’s plate discipline and speed results (e.g. turning singles into doubles) also matter. Instead of incorporation all of them in, I decided to use BaseballSavant’s xwOBA metric to measure a hitter’s production since it combines all of these factors.

The second major input I used was the time frame of 30 days before or after the reported injury. In my previous study on playing through an injury, I noticed some hitters struggled right after the injury but eventually go themselves righted. I just need some baseline and 30 days is it for now. This time frame could be completely wrong, but it’ll work to get the discussion going.

And the final key is that THIS IS JUST A STARTING POINT. Limited time and resources keep me from completely diving in and exploring all avenues. Right now, I am just exploring the surface with the knowledge, I’m going to diving in more later.

So, using the list of 2018 players who played through injury, I found those who had a set event (e.g. hit by pitch). Also, they need to have a “healthy” time frame before the injury. In several instances, the listed hitter started the season injured.

In all, 22 players were used and here are their before and after xwOBA ranked by difference using Savant’s search function.

The average drop was .025 xwOBA and the median at .020.

On average, playing through an injury will bring down a hitter’s production but not always. Eight of the 22 saw an improvement through random variance. Or the injury information was incorrect.

The key takeaway is when should I, as an owner, begin to worry a player has an ailment, investigate them further, and possibly release or bench them.

I’m going to go with an xwOBA drop of 70 points for now. I might even round the value up to 100 points for ease of analysis. If a hurt a player can see their production jump over 60 points, a healthy player could see the same amount of decline.

This initial investigation got me interested in two further areas. The first is to add more than the 30-day time frame. I’m thinking something like previous season and 15-day intervals up to 45 days to help determine the healing rate.

The other big step, for me personally, is to be able to quickly pull the information in-season. Being able to streamline the process is huge when doing it 180 times during the season.

With this small study, xwOBA may be able to help find hitters who are struggling with injuries. The next big step is to see if recovery time can be quickly estimated. I’m not sure when the next study will happen as I’m likely forced to cover moves once the Hot Stove gets going. It might be Monday. It might be in January.