Velocity’s Relationship with Pitcher Arm Injuries

Cliff Lee doesn't throw hard, but he hasn't been completely healthy in a couple years. (via Keith Allison)

Cliff Lee doesn’t throw hard, but he hasn’t been completely healthy in a couple years. (via Keith Allison)

Pitcher injuries have already taken a bite out of the 2015 season already. Yu Darvish, Cliff Lee and Marcus Stroman will each miss part or all of the 2015 season. Teams have always had to deal with injuries, but recently the injuries seem to happen more often and with more severity. One pitching-related item that also has recently increased is fastball velocity.

Since PITCHf/x began tracking pitch velocity, the average fastball speed has increased each year. Can this velocity increase help explain the increased injury rate? Can a jump in velocity lead to an injury? Finally, does an injury always mean velocity loss?

Before we dig in on these fascinating questions, let’s lay out a simple ground rule: I am going to look only at arm injuries. Pitchers can hurt themselves in a variety of ways, including some not related to baseball at all. But I want to see if throwing harder puts undue stress on the arm.

While people consider throwing a baseball an unnatural motion, very few position players tear out their shoulders or elbows playing catch. The issue is the velocity and number of instances throwing the ball in a short time. As Dr. Glenn Flesig told The New York Times in 2013:

Not because throwing isn’t natural,” said Glenn Fleisig, research director of the American Sports Medicine Institute in Birmingham, Ala., and a specialist in pitching mechanics.

“What’s not natural is throwing a hundred pitches from a mound every fifth day,” he said. “That amount of throwing at that intensity is not natural.”

Earlier this year, Fleisig acknowledged to Grantland’s Jonah Keri that velocity could be an issue:

Velocity is a factor. All things being equal, throwing 95 miles per hour is more stressful than throwing 90. But throwing 95 miles per hour with good mechanics is less stressful than throwing 90 miles per hour with bad mechanics. Throwing 95 miles per hour with proper rest is less dangerous than throwing 90 miles per hour without rest.

With that in mind, let’s begin to break this down. Here are the average fastball velocities since PITCHf/x was installed in ballparks in 2007. From left to right, the first column includes cutters, sinkers, split fingers, and two-seam and four-seam fastballs. The middle column is just the harder fastballs — two- and four-seamers (there are three classifications listed because four-seamers in PITCHf/x are listed as FA and FF. Finally, I included the velocity values from Baseball Info Solutions (BIS) going back to 2002.

Average Fastball Velocity, 2002-2014
Season All (PITCHf/x) Just FT, FF, FA (PITCHf/x) Fastballs (BIS)
2002 89.6
2003 89.4
2004 89.6
2005 90.1
2006 90.3
2007 90.9 91.1 90.2
2008 90.7 90.9 90.5
2009 91.0 91.2 90.5
2010 90.8 91.3 91.0
2011 91.0 91.5 90.8
2012 91.0 91.7 91.2
2013 91.2 91.8 91.5
2014 91.2 91.9 91.6

Additionally, here are the BIS data grouped by different fastball speeds.

Image1

We can see some pretty big shifts here. For instance, in 2002, less than half of major league pitchers had an average fastball over 90 mph. Now the number is over 75 percent. Those who average over 93 mph have gone from 12 percent to 32 percent.

Overall, there has been an increase of two mph over the past dozen or so seasons. Everyone reads about the velocity increase and then sees this popular image of Tommy John surgeries.

With Tommy John surgeries and velocity up, velocity must be the problem, right? Not so fast. Here are pitcher arm injuries broken down by year, which originally appeared in an article I wrote here last December:

shoulderVSelbow2

Since 2002, the number of shoulder injuries has gone from roughly 6,500 days lost per season to fewer than 3,000 last year. During the same time frame, the number of days lost to elbow injuries went from about 4,750 to more than 8,000. So while Tommy John surgeries are definitely up, it seems the overall effect is muted.

Just to be sure, let’s group together the total number of days lost per season to pitcher arm injuries:

Days Lost Per Season to Pitcher Arm Injuries, 2002-2014
Season Days Lost
2002 11,724
2003 9,645
2004 11,077
2005 9,434
2006 10,826
2007 13,984
2008 13,597
2009 12,829
2010 10,256
2011 10,342
2012 12,535
2013 14,133
2014 11,835

So, velocity is increasing, pitchers are losing the same amount of days as they did more than a decade ago.

Now, this doesn’t mean velocity isn’t a factor, but it shouldn’t be the only reason people look to when it comes to pitcher arm injuries. If I used the same logic, I could say that the increase in fastball velocity has led to a significant drop in shoulder injuries.

I do believe arm injuries are just being transferred from the shoulder to the elbow because of shoulder strengthening.  From an article I cited last summer:

One huge positive advancement recently has been the ability of pitchers to develop their shoulder muscles–key word: muscles. Muscles can be strengthened. One example is the Tampa Bay Rays pitching staff, as noted here:

[James Shields] devoted himself to the Rays’ shoulder-strengthening program, a 30-minute workout using bands, dumbbells and weighted balls twice a week for 30 minutes.

“No matter where I pitch,” explains David Price, [then] the Rays current pitching ace, “I’m taking this program with me. It’s the best. I tell everybody that comes here, ‘You probably won’t be very good at these [exercises] for a year. It’s tough on your arm at first. It makes you pretty sore. But once you get acclimated to it, it’s great.’ If I didn’t do it now? I would feel it big time.”

A stronger shoulder means that the pitcher can throw harder with less stress to the area. The problem is that the ligaments in the elbow now become the weakest point. If we combine the number of elbow- and shoulder-related disabled list stints longer than 90 days from 2002 to 2011, we find that 51 percent were related to shoulder injuries, with the lowest value being 39 percent in 2004. In 2012 and 2013, however, that total dropped to just 33 percent.

By looking at pitcher groups by average velocity, I can find how often flame-throwers were injured compared to pitchers at other velocity ranges. I will use three velocity ranges from our earlier graph of yearly average pitch speeds.

I’m not just measuring every pitcher, of course, so let’s set up a framwork. I am going get the injury rate of the season in question and then the set speed in the next season. Also, I wanted to look for somewhat capable major league pitchers, so will look for a minimum of 20 innings pitched. Here are the results:

Disabled List Trips by Velocity
MPH Count DL Trips for season Avg days DL Trip chance for next season Avg days
> 96 101 3.0% 64 27.7% 73
93 to 96 1,031 13.7% 45 20.6% 70
> 93 1,132 12.7% 45 21.2% 70
90 to 93 2,308 13.5% 47 15.2% 70
87 to 90 1,655 12.8% 46 11.2% 60
< 87 511 10.8% 51 11.9% 80

The results are not surprising. Pitchers who threw harder were more likely to go on the disabled list than those who didn’t. The difference in injury likelihood is not really in play for the season the pitcher threw hard. The real difference is in the next season. A pitcher who throws a fastball harder than 93 mph is almost twice as likely to end up on the DL, for an average of 60 days, than one who throws less than 90 mph. I included the days lost, but there isn’t much of a difference between categories or any noticeable trend.

If two pitchers are going get the same production, but one throws significantly slower, a team may want take the soft tosser for health reasons. For example, Matt Shoemaker and Yordano Ventura both factored into the AL Rookie of the Year race last season, but with markedly different average fastball velocities. Ventura set the world on fire with his 97 mph fastball while Shoemaker lulled batters to sleep with a 90.5 mph fastball. They had similar results, but this season Ventura is more likely to be DL-bound.

With general fastball velocity shown to cause more injuries, do velocity spikes also cause injuries? This study has been three years in the making since I read the following quote from Keith Law:

The team’s most promising major league starter at this point is Danny Duffy, who has shown increased velocity so far this year and missed his last start with a sore elbow. After a promising side session on Sunday, he’s expected to make his next start, but a velocity spike plus immediate arm soreness is a dangerous sign. Speaking anecdotally, it seems we often hear about a pitcher throwing harder than ever right before something goes “snap!”

Law is not the only person to make this observation. Every year some pitcher seems to experience a decent velocity spike and then ends up hurt. Last season, the pitcher was Tyler Skaggs. Skaggs’ fastball was at 89.2 mph in 2013 for the Diamondbacks. In 2014, it jumped to 92.1 mph and a couple months later he was on the disabled list.

Well, the axiom couldn’t be more wrong. I looked at pitchers who were mainly starters in season one and two (GS/G >= .5), pitched a minimum of 20 innings in season one, were 27 years old or  younger in season one and saw their velocity increase 1.5 mph in season two. In total, 49 pitchers met this requirement. Only nine of them went on the disabled list with an arm-related injury (18.4 percent). If I look at the entire population, (removing the velocity change requirement) 42 percent of pitchers went on the DL with an arm-related injury.

One big reason I could see is the pitcher was dealing with an injury during season one and is finally feeling better in season two. Or, the pitcher could have built up his strength over the offseason and could be healthy and can throw harder. While a velocity jump may be considered a warning that injury is coming, it really isn’t. A starter seeing a velocity jump historically has been healthier than his peers.

Finally a look at how much velocity loss we can expect from a pitcher who experiences an injury. I am going to look at four groups of pitchers for this. First, starters (GS/G >= .5) and relievers (GS/G < .5) in the season of the arm injury. Then, I am going to group these pitchers in major injuries (>= 45 days on the DL) and minor (< 45 days on the DL).

Starting Pitchers, < 45 Days on Disabled List
Injury Yr -1 Injury Yr Yr +1 Yr -1 to Injury Yr Injury Yr to Yr +1 Avg Days Missed Samples
Other Arm 90.9 90.4 90.5 -0.5 0.1 25 43
Elbow 90.9 90.3 90.5 -0.6 0.2 28 65
Shoulder 90.6 90.2 90.2 -0.4 0.0 23 116
Starting Pitchers, >= 45 Days on Disabled List
Injury Yr -1 Injury Yr Yr +1 Yr -1 to Injury Yr Injury Yr to Yr +1 Avg Days Missed Samples
Other Arm 91.2 90.9 91 -0.3 0.1 74 26
Elbow 90.9 90.6 90.7 -0.3 0.2 100 79
Shoulder 90.8 90.1 90.2 -0.6 0.0 92 78
Relief Pitchers, < 45 Days on Disabled List
Injury Yr -1 Injury Yr Yr +1 Yr -1 to Injury Yr Injury Yr to Yr +1 Avg Days Missed Samples
Other Arm 92.7 92.6 92.2 -0.1 -0.4 26 43
Elbow 91.7 91.5 91.0 -0.2 -0.5 27 93
Shoulder 92.0 91.5 90.2 -0.4 -1.3 25 107
Relief Pitchers, >= 45 Days on Disabled List
Injury Yr -1 Injury Yr Yr +1 Yr -1 to Injury Yr Injury Yr to Yr +1 Avg Days Missed Samples
Other Arm 92.1 91.8 91.4 -0.4 -0.4 89 25
Elbow 91.9 91.5 91.2 -0.4 -0.3 107 98
Shoulder 90.7 91.2 91.2 0.5 0.1 93 87
All Pitchers, Disabled List
Injury Yr -1 Injury Yr Yr +1 Yr -1 to Injury Yr Injury Yr to Yr +1 Avg Days Missed Samples
Other Arm 91.8 91.4 91.3 -0.3 -0.2 46 137
Elbow 91.4 91.0 90.9 -0.4 -0.1 67 335
Shoulder 91.0 90.8 90.4 -0.3 -0.3 53 388

No real surprise here, with pitchers throwing with less speed in the season they got hurt. Additionally, they didn’t see any bounce back in the next season, which is expected. Starters will see about a 0.5 mph drop with an injury and get a small bit back the next season. The relievers will see less drop (or an increase with a rested shoulder), but the decline will be greater the next season. The overall numbers show a one-third mph drop in the season of the injury, and then a small decline the next season.

Conclusion

High velocity pitchers are more likely to go on the disabled list for an arm-related injury than those pitchers whose fastballs don’t register as much heat. However, it looks like a velocity jump is not a predictor of a pitcher getting injured. Finally, an arm injury is generally a sign of velocity loss, with there being little chance of the pitcher regaining the velocity. High velocity readings are not going away; we now know more and more their effects on pitchers.

Next Steps

The key for future studies on this subject is accruing more data. We have injury transaction and velocity information going back only to 2002. The more data that comes out each season, the more we can be sure of the effects. I don’t see the number of hard throwers declining in the near future. With more and more hard throwers making it to the majors, the sample to test will become larger.


Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
14 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
royalguy
8 years ago

velocity is becoming the premium medium for going to the majors. Blazing fastballs get outs.
There was an extraordinary article that came through the community research portion of the site that did a study on velocity or movement. They found velocity is more valuable than movement. Velocity is what’s drafted for too.
Every 10 year old pitcher knows how hard he throws and what that means he can do to hitters. It’s the premium in pitching! You don’t brag about the swing and miss rate of your curveball. You brag about your velocity. It’s not going to change because that’s how outs are gotten at every level of baseball from little league to the MLB!

andrew
8 years ago

Could be that improved shoulder strength/health is enabling higher velo and exposing the elbow?

Overall DL days are fairly flat, but looks like shoulder and elbow are on offsetting trends.

Guy
8 years ago

Nice analysis. In terms of the link between velocity and injury risk in year two, is your sample large enough to add age to the analysis? It seems to me that young flamethrowers (age <25?) are especially likely to get hurt. And is there any notable difference between starters and relievers here?

Ruff
8 years ago
Reply to  Guy

I agree. I feel as if the lowest velocity group would be significantly older than the other groups, especially the 96+. I would love to see if taking a 20 year old at 98 mph is actually riskier than a 27 year old throwing 98.

Jon Roegele
8 years ago

Nice work Jeff. Regarding the velocity spike and arm injuries, one thing I’ve always been interested in is even short term spikes just before injury – over a short enough period that you might not pick up the spike if looking at the seasonal average. I suppose the Josh Kalk (and Kyle and Noah) Injury Zone work might pick up on this. I’m thinking about for example Drew Hutchison; if you check his final two starts I think before his Tommy John in 2012, his FB velocity was noticeably higher.

Also related to this topic was when I looked at velocity of pitchers (controlling for role and age) and Tommy John surgeries specifically.

Velocity and Tommy John Surgeries

Julien Assouline
8 years ago

Hey Jeff I really enjoyed the article. Especially the part about velocity spike and drop. Something I’ve been interested in for a while, was really cool to see the results.

I was also curious as to some of your methodology. For example, with velocity, you broke them down in buckets of three. Was there a specific reason why you chose three as oppose to let’s say five or four? Also how did you define a relief pitcher? Did you set an innings limit? These are elements I’ve been struggling when it comes to my own research.

rustydude
8 years ago

If there are Pulitzers given out for well-researched/well-written sports articles, this piece should qualify for a nomination. Thanks, Jeff!

evo34
8 years ago

Nice work. One issue I see with reading anything into the final few tables, however, is that there is a huge selection bias, right? I mean, I’m sure lots of injured pitchers didn’t make it back at all or only for a few innings of terrible performance. Or am I misreading something?

evo34
8 years ago

Also, I’m not following the claim that “an arm injury is generally a sign of velocity loss, with there being little chance of the pitcher regaining the velocity.”

In your previous article, you found that healthy pitchers lose 0.37 mph/year on average:

Average Change for Healthy Pitchers
K/9: -0.26
BB/9: + 0.03
ERA: + 0.22
FBv: -0.37

So one would expect your injured subset in this article to show more than 0.74 mph of velo loss from Yr-1 to Yr+1. But they actually show -0.2 and -0.7 for SPs and RPs, respectively, on the DL > 45 days with an elbow injury. That suggests that for those seriously-injured-elbow starting pitchers who make it back to the majors, they may experience an unexplained velocity increase (compared to their non-DL counterparts) of ~ +0.5 mph. Would be curious to see TJS vs. non-TJS elbow injury pitcher comparison.

Dan
8 years ago

MLB has become velocity blind. Velocity is great but in this day and era of MLB the next Greg Maddox would be overlooked and might not even get drafted. There are throwers and there are pitchers. The art of pitching is becoming a lost art. I once won a bet by pitching to the 2,3, n 4 hitters on a high school team. That I would strike them out or if they hit it it wouldn’t leave the infield. I also had to call what infielder a fair batted ball would be hit to. I also was not allowed to throw a pitch above 60 MPH, walk or hit a hitter, if I did, I loose the bet. I won, two strike outs on 7 total pitches and a called ground ball to the 3rd .baseman on 3 pitches.