The Odds of Matt Harvey Breaking Down

Yesterday, it was reported Matt Harvey may need Tommy John surgery because of a torn UCL in his right elbow. Some people may say they saw the injury coming and the Mets were crazy to let him throw over 175 innings this season, but the evidence doesn’t really support those ideas. After looking over the history of other 24-year-olds, it appears that the pitcher’s ability to throw hard and recent small velocity drop were the only identifiable injury indicators.

Myself and others have looked at many indications of a pitchers chances of getting hurt. High increase in innings for a young pitcher (Verducci Effect). Velocity and Zone% drop (PAIN Index). Inconsistency in release points and velocity late in a game. High breaking ball usage. Bad Mechanics. High fastball velocity.

With so many possible explanations, I figured I would find at least one major issue sticking out, but I didn’t. I will step through each of these possible theories and just show how it was all but impossible to know this injury was coming.

Verducci Effect

Tom Verducci of Sports Illustrated created the theory which a “pitcher who is under 25 years old and who had an increase in his workload of 30 innings or more in the previous season is at greater risk for injury or for a steep decline in performance.” Matt Harvey fits into this category for 2013 since his innings did increase 30 IP from 2011 to 2012 (135.2 to 169.1 increase), just exceeding the 30 IP limit.

To put this simply, the Veducci Effect has been proved worthless for several years now and should not be used to evaluate pitcher injuries. I am sure it will be mentioned on TV several times in the up coming days, but please know that the theory has been roundly debunked, and is not any type of valid analysis.

Declines in Velocity and Zone% (PAIN)

Since the beginning of the season, I have been working on a metric to help find pitchers who may be injured. Basically it looks for pitchers with a declining fastball velocity and have the inability to throw strikes. I named it PAIN. Just before Harvey’s injury was reported yesterday, I put him at a PAIN value of -86 which is outstanding (+100 are pitchers at risk). He has improved on his 2013 first half fastball velocity (95.2 to 95.7 mph) and Zone% (51% to 53%), so this marker did not pick up any red flags.

One possible issue was a small drop in his average fastball velocity over his last half dozen games.

Most of the decrease was giving up gains he saw right before the All-Star Game. After the game, his velocity was down about 1 mph. Some concern, yes. Worried after the early season gains, no, not really.

Late Game Inconsistencies

I continued the work of Josh Kalk and Kyle Boddy to find injured pitchers by looking at inconsistencies late in their starts. I created an application to find these instances.

Looking at Havery over the past 2 seasons, his recent late game consistency was impeccable (a value of 100 indicates a possible injury).

Basically, Harvey maintained his velocity late into the game while keeping a consistent release point.

Breaking Ball Usage

Matt Harvey is known for his beautiful slider. In case you have seen him throw in the last five minutes, here is another look:

The slider is Matt’s best pitch and he throws it regularly because it is so effective. Some people think throwing too many breaking balls may make a pitcher injury prone.

In a study I performed, I found pitchers with high slider usage are more injury prone. The problem is Harvey doesn’t throw his slider enough to make a difference. A pitcher needs to throw their slider over 30% of the time to have a higher injury risk. Harvey throws his less than 20% of the time, because the rest of his pitches are pretty good as well.

Pitching Mechanics

Throwing a baseball hard is hard on a body. Throwing the ball with bad mechanics is even harder on a body. I know little about pitching mechanics, except I have no idea how Chris Sale‘s arm doesn’t explode every time he throws a ball. Others though will give their expert opinions.

Jim Margalus of put together a great list of people praising Matt Harvey‘s mechanics. Harvey’s mechanics have been described as “good”, “clean”, “flawless” with “no violent movement”. Former major leaguer (3.2 IP in 1970), Dick Mills even created a video touting Harvey’s mechanics

Some flaws with Harvey’s mechanics probably could be noted by someone, as it could with any pitcher. The general consensus before the injury was of a pitcher with a sound delivery.

High End Velocity Pitchers

Matt Harvey throws the ball hard compared to other starters. Of the 2013 staters with over 120 IP, Stephen Strasburg is the only other pitcher to throw his average fastball over 95 mph.

I have always wondered if throwing continually at extreme velocities can injury a pitcher more often than slower tossers. With hard throwers such as Stephen Strasburg, Alexi Ogando and Michael Pineda dealing with major shoulder and elbow injuries, I finally dug into a few numbers. Since 2008 (when Pitchf/x data has been reliable), there have been 16 instances of a starter averaging over 95 mph.*

Assuming Harvey spends time on the DL in 2014 and ignoring Strasburg for now, 50% of these pitchers spent time on the DL in their next season. In the past, a starting pitcher who threw 120 innings in one season has a 39% chance of going on the DL the next year. An increase, but not without its flaws. I am not excited about just a sample size of 15, so I expanded the search to include 22 pitchers who averaged between 94 and 94.99 mph. Of these additional pitchers, 36% went on the DL the next year bringing the overall total to 42%. Not a much of change from the league average of 39%.

It seems that extreme hard throwers have some additional injury risk, if a person believes in a sample of 15 pitcher seasons. By expanding lowering the velocity threshold down to 94 mph, not much additional risk exists.


Putting it all together, there were no glaring red flags pointing to an injury issue with Matt Harvey, just a couple of slightly pink ones. He was experiencing a small drop in velocity and may throw the ball too hard on a regular basis. Other pitchers do the same things daily and don’t tear their UCL. Additionally, he showed no signs of bad mechanics, inability to throw strikes, high work load, too many breaking balls or late game inconsistencies. His injury is just another example of a pitcher trying to do his best and his body not letting him.


* Ubaldo Jimenez (2008, 2009, 2010), Stephen Strasburg(2012, 2013), Michael Pineda (2011), Matt Harvey (2013), Justin Verlander (2009, 2010, 2011), Jeff Samardzija (2012), Felipe Paulino (2011), David Price (2012), David Price (2010), Alexi Ogando (2011)

Print This Post

Jeff writes for FanGraphs, The Hardball Times and Royals Review, as well as his own website, Baseball Heat Maps with his brother Darrell. In tandem with Bill Petti, he won the 2013 SABR Analytics Research Award for Contemporary Analysis. Follow him on Twitter @jeffwzimmerman.

49 Responses to “The Odds of Matt Harvey Breaking Down”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. anonynous says:

    What a shame. This all could have been avoided if the Mets had gone with the Harvey Rules like the Yankees used with Joba.

    +9 Vote -1 Vote +1

  2. TKDC says:

    Great article. Does the history of unpredictability in pitchers’ injuries make it fairly silly to try to peg guys as injury risks based on mechanics? It seems there are a ton of guys people like to say they “knew” would get hurt based on mechanics, but so many with supposed flawless mechanics get hurt, too. Is there really a good reason to think that one way of doing something that is extremely unnatural and violent is so much better than a different way?

    Vote -1 Vote +1

  3. Mike says:

    Or maybe that elbow was damaged at UNC, where he wasn’t exactly treated with caution.

    Vote -1 Vote +1

  4. olethros says:

    The Mark Prior Hypothesis: pitchers with a reputation for flawless mechanics are doomed to career destroying injury.

    +28 Vote -1 Vote +1

    • Jeff Zimmerman says:

      Good one, but Dick Mills would disagree.

      Vote -1 Vote +1

      • Paula says:

        I am of the opinion that very few people listen to Dick Mills anymore. There isn’t a single source that everybody agrees upon, but most everyone agrees he doesn’t know what he is talking about. He doesn’t use any scientific data to back up his claims; he only uses dubious anecdotal evidence.

        Vote -1 Vote +1

  5. hard sliders says:

    Nice analysis. One possible correlation to quantify:

    1. Very hard slider (top 5%-10% in MLB?)
    2. Lots of very hard sliders (over 15% of pitches as sliders?)

    Correlated with *significant* injury or actual surgery (decrease of at least 50% in innings from when hurt?). I think *any* DL time may obscure the data.

    Compared to league average of significant injury time.

    Opposite may be instructive too…how many pitchers require elbow or shoulder surgery that were within a standard deviation of avg fastball velocity, slider velocity/quantity?

    Also, not sure why Strasburg seems to be removed from the analysis above if I am reading it correctly?

    Vote -1 Vote +1

    • Jeff Zimmerman says:

      His 2013 season is removed, we don’t know if he will be hurt in 2014 yet.

      I like the ideas on the sliders, sounds like a nice off season project.

      Vote -1 Vote +1

      • Richie says:

        I’ve never seen anyone examine the Weaver Rule (young pitcher = long relief, period, end of discussion). How closely did Weaver follow his own rule, and what was the health record of his young pitchers?

        Vote -1 Vote +1

      • Dave Cornutt says:

        Interesting… Phil Hughes is throwing 23% sliders this year, a lot more than he has thrown in the past. Might be worth keeping an eye on.

        Vote -1 Vote +1

      • murphym45 says:

        We’ve already seen one pitcher scrap his hard slider this season in Andrew Cashner, and he’s been relatively healthy after being plagued by injuries earlier in his career. However, he also stopped striking guys out.

        Vote -1 Vote +1

    • Alan says:

      How many fastball/slider pitchers throw less than 15% sliders?

      Vote -1 Vote +1

  6. Richie says:

    He was throwing at a pace of 210+ innings for the whole year. Maybe high work load for a 24-year-old isn’t a red flag, but I’d say he did have one these past 5 months.

    Vote -1 Vote +1

  7. Joaquin Andujar says:


    Vote -1 Vote +1

  8. cass says:

    I think it’s the velocity, but we don’t have a big enough sample size to know for sure. How about a graph of next-year injury percentage by each half mph of average fastball velocity? 93.0-93.5, 93.5-94.0, 94.0-94.5, etc.

    +10 Vote -1 Vote +1

    • JayT says:

      Yeah, I’d be very curious to see this too. Do guys that throw it 89-90 have a much lower injury rate than guys that throw hard?

      Vote -1 Vote +1

      • what...? says:

        An issue i could see with that though is survivor bias. Not many young guys pitching at 89-90. Not many old guys pitching 93+. Don’t see how you can adjust for age without making the sample sizes ridiculously tiny.

        Vote -1 Vote +1

  9. Gaupo says:

    Check out Gerrit Cole’s PitchFX and it becomes obvious what the Pirates think. Cole has thrown over 85% fastballs (four seam, two seam and cut) and less than 10% sliders even though his slider is nasty and they’re in a pennant race.

    Vote -1 Vote +1

  10. FeslenR says:

    He’s a Met, that alone should explain why it happened…”snakebitten, baby!”

    sliders shmiders, injuries happen….

    Vote -1 Vote +1

  11. Cuck city says:

    Why can’t arm injuries just be random?

    Vote -1 Vote +1

    • Richie says:

      Well, literally, because nothing is. A coin will land heads or tails as a result of how high you flip it, at what angle, with what velocity, how hard the landing surface is. Just a matter of figuring out the factors.

      Which just seems like we ought to be able to figure for arm injuries, dadgummit. And given how significant they are to baseball, we’re going to keep trying.

      Vote -1 Vote +1

      • Cuck city says:

        Well I guess its the improbability of all the injury factors lining up in the (in)correct circumstances on a particular pitch. Clumpiness of dirt on the mound, Arm angle, ‘effort’, etc.. and not usage over time.

        Vote -1 Vote +1

      • Boom Shakalaka says:

        Well, you can try to figure out all the factors if you want, but you’re also going to have to figure out contributing factors that may exist outside of the game of baseball.

        The fact is that starting pitchers are only on the field once every five days and who knows what they are doing off the field. Maybe Matt Harvey plays golf on his off days and has a swing worse than Charles Barkley that torques his elbow. Maybe he was trying some wierd sex positions with Anne V. Maybe he hit his funny bone one day like any normal person would randomly and his UCL tore because it was already under stress from pitching.

        The point is you may be able to quantify as many factors as you can think of, but you will never be able to quantify randomness.

        Vote -1 Vote +1

    • what...? says:

      Because the entire point is to figure out if it’s random or not. Whoever figures out the pattern (if it exists) is gonna make some SERIOUS money.

      Vote -1 Vote +1

    • SmileBunch says:

      I think we assume from the get go that there is a random component, but with analyses like this we’re attempting to identify patterns that are non-random within the data.

      It would be tough to measure, but it seems like there are probably factors other than workload, mechanics, and work type that would play a role such as nutrition, body composition/build/musculoskeletal frame, weight gain/loss, genetic factors that influence tissue durability, etc.

      Vote -1 Vote +1

  12. Matt says:

    Can we get some analysis on the Mets medical and conditioning staff? It may be confirmation bias, but it seems the Mets have a history of not treating minor injuries like the forearm tightness Harvey admitted to dealing with. Their mismanagement of injuries has long been documented anecdotally, but I’d love to see some data to back it up, even if it’s from a small sample size. Harvey would be very wise to follow fellow Boras client Carlos Beltran and get independent doctors to handle his entire surgery and recovery process.

    Vote -1 Vote +1

  13. Jason says:

    I know (meaning, i think i know) the Rays agree with the “too many sliders is a bad thing” mantra. Im pretty sure ive read somewhere that the Rays discourage the pitch. David Price had a filthy slider. They “made” him stop using it. Jake McGee also had a sick slider..he throws it about once a month now. Chris Archer is the only Rays pitcher that seems to throw it very often.

    Vote -1 Vote +1

  14. Adam G says:

    I think the idea of “clean” or “smooth” mechanics should be shelved for a while. When throwing a ball overhand with that much torque, there is nothing natural or healthy about it. It’s like saying a running back takes hits “cleanly” or something like that (not a great comparison, I know). It’s a physical motion that is mostly unparalleled in other sports, and frequently causes significant damage to the joints involved.

    Obviously, there are mechanical features to the overhanded throwing motion that are more problematic than others, but in general it’s not good for you. Pitchers like Chris Sale and Tim Lincecum stick out for their “weird” or “bad” mechanics, but what about Randy Johnson? Madison Bumgarner? Walter Johnson? These guys also have weird motions and have been relatively healthy. And if you look closely, Nolan Ryan had a slightly odd arm angle.

    This rant doesn’t really add much to the discussion, but I guess what I’m trying to say is that after all these years, this talk of “good” vs “bad” mechanics is still usually vague and of little value. Maybe we’ve all been looking at this problem from the wrong angle. Instead of seeing injured pitchers as an outlier, accept the fact that the UNinjured pitcher is the outlier. Instead of saying, “why did pitched X have this injury?” and instead start saying, “why have these pitchers NOT been getting injured?”

    Vote -1 Vote +1

    • Ruki Motomiya says:

      Except more pitchers are going to be uninjured at any one time then injured, so the injured pitcher is ALWAYS the outlier. If your logic is “Well, they all get injured eventually!”, then healthy position players are also outliers.

      You can’t be an outlier (an observation that is numerically distant from the rest of the data) if you are less numerically distant than the thing you are trying to replace as an outlier, unless you want to say multiple outliers, in which case this accomplishes nothing.

      Vote -1 Vote +1

  15. Phantom Stranger says:

    I’d be less concerned about high-velocity fastballs and more about the ability to throw high-velocity off-speed pitches. I cringe every time I see Jose Fernandez throw his devastating slider. That much break on your off-speed stuff has to be generated by placing tremendous biomechanical stress on your joints and ligaments.

    It’s very possible that some arms are simply not built to handle the repeated stress of throwing the modern slider at typical velocities these days.

    Vote -1 Vote +1

  16. vincent says:

    Why, in today’s game, when pitchers are better conditioned, have better travel accommodations, and are generally facing watered down competition due to the number of teams, have low innings and injuries that accompany them become almost normal? I understand data highlighting favorable match ups has changed the way pitchers are used, but still. Strasburgh and Harvey would not have had an inning limit if they pitched in fifty or sixty years ago. Would they have blown out there arms if they came up throwing heavy innings? Just wondering if there was a prevailing SABR thought on this.

    Vote -1 Vote +1

    • Kogoruhn says:

      Pitchers fifty or sixty years ago got injured just like today except we just tend to forget the ones who had devastating injuries. We remember the ones who managed to stay healthy while we forget the ones like Gary Nolan who had their careers ruined by an injury.

      Vote -1 Vote +1

  17. John Sterling says:

    I keep telling you nerds, but you won’t listen.

    Vote -1 Vote +1

  18. Roman says:

    dropped in keeper league, have a very bad gut feeling about this and my gut is always 100% huge.

    Vote -1 Vote +1

  19. DC says:

    I think you are reading way too much into this injury by looking for the symptoms of a symptom, which I feel is a trap that sabermetics tends to set people up with. What causes the UCL to tear? Not decreased velocity, or erratic throwing but almost all non-contact UCL injuries are a result of micro-tears brought upon by repetitive stress turning into a large tear. The idea that Matt Harvey has flawless mechanics is because there is no such thing. Throwing a baseball is never going to be something that isn’t stressful on the elbow joint. Do “good/bad” mechanics change the stress loads? Sure but I would argue as to what extent that actually matters when your arm is violently whipping one way or another. So the only other options we have to consider is histology. To some extent tissues strength is limited by our genetics. Yes we can strengthen a tendons and ligaments by proper aligning their fibers but there comes a point where enough is enough and that point is not the same for me as it is for Matt Harvey. I would estimate that 60% of the problem is unavoidable because the motion and 15% is unavoidable because of the tissue matrix. That leaves 25% though that I believe is in our control and this has nothing to do with innings or pitch count (unless they are exceedingly high i.e. 300 plus innings and maybe 170 pitches a game). But rather with muscle strength affecting the elbows stability and overall range of motion. During a pitch, almost all the time the arm during the deceleration phase is supinating. This causes the stress felt in the medial elbow. In order to counteract this motion the arm must have sufficient strength in the muscles that produce pronation (i.e. the aptly names pronator, brachioradialis, etc.) If they cannot offset the force being produced during supination then the arm will fall farther into supination increasing the stress on the medial elbow and thus the UCL. But how can we quantify this? Perhaps the best way to do this is through ROM. I plan on accomplishing an in-depth study on this subject within the next 3 years, but the basic premise is if there is a pronation/supination ROM that may be part of a danger zone? By establishing healthy baselines, a pitcher can be evaluated quickly and hopefully reliably before and after each time they take the mound (practice or game).

    +6 Vote -1 Vote +1

    • Ruki Motomiya says:

      I like this post.

      Vote -1 Vote +1

    • Jeremiah says:

      What do you mean by ROM? Also, to what extent are modeling techniques like Finite Element Analysis used or applicable to this issue?

      Vote -1 Vote +1

      • Shawneecubs says:

        ROM= range of motion. This study has done before. It found that fastballs had the highest torque on the shoulder and elbow. Curveballs and change ups were 2nd/ 3rd for highest torques. It’s one of the reasons pitch counts are such a big deal now.

        In my work with young pitchers, I would guess 50% at least have some ucl laxity and therefore micro tears. It’s whether they are symptomatic or not that’s most important.

        Finally, I have heard one of the reasons pitchers break down is the use of weight training. The tendons in the shoulder specfically are tiny and can only hold up to a certain amount of force.

        Hope it helps.

        Vote -1 Vote +1

  20. Jon L. says:

    I knew it! I knew Harvey would get hurt and it would be totally unforeseeable. Just wishing I’d said something…

    Vote -1 Vote +1

  21. Michael says:

    To me, pitch counts, more so than innings, are more important. Higher pitch innings, or higher stress innings, are probably even more important than overall pitch counts. What’s amazing to me is that everyone seems to forget that it takes throws to actually warm up the arm. How many throws does it take for Harvey to warm up? How many throws does he make in the bullpen before his start, or on his side days? How many throws does he make in between innings? In a 7 inning start, where he throws 105 pitches, he may make an additional 56 in between innings, another 40 in the bullpen before the game, and 30 or 40 to actually warm up before he starts his bullpen. That’s 230-240 throws. And while we’re at it, what is his off-season throwing program? How quickly does he try to get up to velocity? Does he ease his way into it, or does he ramp himself up too quickly? How soon does he start throwing breaking stuff. It matters to build your arm strength up. I’d argue that the studies aren’t precise enough. While you might not be able to follow Matt Harvey (or whomever else) around everyday, that’s the type of data that I would be interested in.

    Vote -1 Vote +1

  22. Jonathan Sher says:

    Excellent analysis, Jeff.

    One tiny suggestion: Your small sample size of 15 hard-throwers might be reduced to 14 as Pineda had shoulder woes in the minors before he became such a hard thrower. At the time he was topping out in the low-90′s, I recall.

    Vote -1 Vote +1

  23. No Need to Argue says:

    Harvey did something physically stressful. That stress damaged his ulnar collateral ligament.

    Vote -1 Vote +1

  24. Joel says:

    Excuse me doctor, I think I know a little something about medicine.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>