New Research Shows Two Twins Are Biggest Injury Risks in MLB

[Matthew Trueblood]

According to a new model for injury prediction from Sports Info Solutions, Andrelton Simmons is the position player most likely to miss time due to injury in 2021. Next on the list? Byron Buxton. We talked to a co-creator of the model to find out why.Obviously, it will come as no great shock to Twins fans that Buxton is considered one of the league's biggest injury risks. He's battled concussion issues, shoulder injuries, and wrist trouble. He's suffered from back spasms, had toe trouble, and strained his groin. He's even had migraines, perhaps as a secondary symptom of repeated concussions.

 

Simmons is also known for his fragility, but since he's new in town, Twins fans might be surprised to learn that he is not only more likely to get hurt than Buxton, but more likely than any other big-league regular. Nonetheless, the highly sophisticated system co-created by Sports Info Solutions (SIS) colleagues John Shirley and Matt Manocherian says precisely that. To find out more, I talked to Shirley about how the analysis was performed, and on what it's based.

 

Video scouts employed by SIS watch every big-league game, every season. As far back as 2015, they have logged not only the movements and actions on every play of each game, but also all injury-related events. According to Shirley, this ranges from the obvious (and easily found elsewhere) like players being helped off the field with major leg injuries to the almost unnoticeable (and rarely logged), like a player rolling their ankle on a base but staying in the game, or being hit by a pitch in a certain body part, or grabbing their arm between pitches. All of those injury events are catalogued and cross-referenced with other sources, such as injured list transactions. Where needed, the company's injury coordinator, John Verros, follows up on the report.

 

"For instance," Shirley said, "John might check whether a player rolled an ankle inward or outward, and adjust the prognosis on that basis. It's that level of detail."

 

With that extremely detailed database in place, the co-creators separated hitters from pitchers, to build distinct models for the two player types. For each, they then used machine learning techniques to train the database to predict injuries over various periods—one month, two months, a full season.

 

Shirley noted that the number of layers and data points involved make it difficult to boil down any of the model's findings to a simple explanation. That's reasonable; that web of interdependencies is one reason why it's so hard to predict injuries at all. However, with Simmons, there are certain markers that pushed him toward the top of the list.

 

"I think just that he's such a high-usage player," Shirley said, noting that Simmons "makes a lot of diving plays, covers a lot of ground, is kind of willing to throw his body around a bit."

 

Much of that can also describe Buxton, of course, which has fueled many of his own injury issues over the years. Shirley also noted that, since Simmons has had ankle injuries in consecutive seasons, the model takes not of the risk that such issues can become chronic, or that they can cause injuries at different points in a player's kinetic chain.

 

Both Simmons and Buxton are elite defenders. That, as it turns out, can be a leading indicator of injury problems. Shirley's use of the term "high-usage player" is telling. In other sports, we have become accustomed to the idea that some players bear much heavier workloads than others, even on a per-play basis. Those players are more exposed to injury risk. The same is true in baseball, in ways we might not have fully appreciated until now. Because Simmons and Buxton can get to so many balls, they push their bodies to the limit in efforts to do so. That comes at a higher cost than anyone fully understood, prior to the collection of this much detailed information and the presentation of the findings it fueled.

 

One of the surprising things they found, Shirley said, was that the model often slightly decreased injury risk as players aged. Traditionally, we imagine that players grow more vulnerable to injuries as they age; our bodies break more easily and recover more slowly after age 25.

 

However, players also change their behavior as they age. No longer being able to make certain plays can encourage a guy not to overextend themselves, and they can become more durable even as they become (slightly) less valuable on a per-play basis.

 

For now, Simmons and Buxton are as likely as any other players in baseball to get hurt and miss time. That should sour any solace a Twins fan was finding in the White Sox's misfortune this week. The good news, though, is that the Twins are aware of this, too, and have built their club accordingly. Tomorrow, we'll discuss the implications of these findings when mapping the roster, for Opening Day and beyond.

 

MORE FROM TWINS DAILY

— Latest Twins coverage from our writers

— Recent Twins discussion in our forums

— Follow Twins Daily via Twitter, Facebook or email

 

Click here to view the article

[ToddlerHarmon]

The good news is that Nelson Cruz will play DH forever.

 

I am curious what the injury exposure for just batting and baserunning. Speed probably amplifies the risk here, too

[Channing1964]

Wow...this information was always available? Our trend setting FO must have missed something with their massive information data collection support group.

[rdehring]

Interesting. Got me thinking about the Twins record when Buxton is playing. What is it, something like winning 2 of 3 and a losing record when he isn't. Will it be something similar when Simmons plays? If so, let's pray that they aren't injured at the same time. And will it be winning 3 of 4 when both are in the lineup?

[lukeduke1980]

After reading the headline I assumed Buxton and Donaldson

[MN_ExPat]

Normally I bit my tongue on this kind of thing.... but this is just silly to me.      This seems like the usual human attempt to quantify everything.        It's a highly competitive sport played at the highest levels, players just get hurt and some more than others.

 

It's just a fact of the game.

[Doctor Gast]

We're blessed to have such gifted players to playing at 2 of the most crucial positions. The problem is that there's a big drop off of talent when they aren't playing. Our success to arrive and advance in the PS largely is based on their health. Donaldson is also crucial and his health is very important. We need them all healthy for the PS.

[vavo]

Interesting analysis. I tend to be kinda skeptical of the approach of dumping a bunch of data in a bucket and flipping the "machine learning" switch. Sometimes it gives extremely valuable results, sometimes it produces garbage. As you said, it's also difficult to determine what it actually learned.

 

That being said, the high-usage concept is interesting, but I thought about it a little differently. Buxton and Simmons play the two positions that see a lot of action. Is there evidence that SS and CF see more injuries than say, 3B/LF?

[chpettit19]

 

Wow...this information was always available? Our trend setting FO must have missed something with their massive information data collection support group.

No, this data was not always available. This is an incredibly new set of information. And the idea that our trend setting FO would have this data from their support group is ridiculous. No major league team has the resources to track this stuff on their own. Its why companies like SIS exist. And once a team goes to them with a request for new information it takes a year or 2 to get the data together and finalize the algorithms and data. And, while those guys at SIS are crazy smart, this is one of the more questionable data sets out there.

[dbminn]

Nice work, Matt. Glad you followed up with the source.

[Dodecahedron]

I have an uncontrollable urge to type something like, "Buxton? Injury risk? Fake news! He has never ONCE been injured!"

 

However, I will resist that temptation.

 

I think the question with Buxton is not "Will he be injured" but "How many times will he be injured" and "How many days will he miss due to his injuries?"

 

He is amazing when he is on the field, and the Twins are rather mediocre when he's not out there. When 85 games of Buxton gets you in the playoffs with 100 wins, I'm wondering what 160 games of Buxton would produce. The team was 62-25 in games he played in 2019, a pace of 121 wins.

 

But this doesn't fix the usual problem: 121 wins does not equate to a single win the playoffs.

[The Wise One]

That players who hit their bodies up against solid objects have more injuries than those who do not I thought would be common sense. Who knew that it had to be quantified to be valid? 

[jkcarew]

 

The good news is that Nelson Cruz will play DH forever.

I am curious what the injury exposure for just batting and baserunning. Speed probably amplifies the risk here, too

Luckily, Simmons and Buxton have been able to mitigate this throughout their careers by employing low and really low respective OBP’s

[jkcarew]

It seems a bit questionable to lump Simmons and Buxton together as ‘high usage’. Any SS is in an entire different category of usage than any CF, including Buxton. That’s still true, even in today’s game. IMO, Buxton is just Buxton. Center Field didn’t do much to create material injury issues for Puckett or Hunter who both played the position aggressively (and in the case of Hunter, covered a ton of ground).

 

Call me skeptical regarding how much this is moving the needle regarding the predictability of injuries. Worth a try, I guess...if someone’s going to pay for it.

[prouster]

Someone had grant money to burn.

[Fatbat]

Not saying the model is bad or inept but Eloy Jimenez should be #1 on the list.