how did the whole mechanics—>injury predictive snowball ever start rolling in the first place? How did something so completely speculative end up being such a universal point of interest that there’s at least one person on every baseball website who fancies himself an expert?

If the whole “diagnosing injuries based on mechanics” thing is bunk, what do you think about the people who attempt to determine future injuries based on past injuries? Is there any reason to think that a past injury makes a future one more likely?

That I read about a few years ago from (I think) BP? I assume that’s still considered a decent guideline, since it was more about age and chance of injury than about any particular individual or throwing motion.

any opinion on the significance of the work being done at ASMI?

http://www.asmi.org/asmiweb/research.htm

With the explosion in popularity of this sub-genre of baseball commentary, I think more and more people are learning that there are 3,4, 5 mutually exclusive schools of thought on pitcher mechanics, and that each group thinks the others are basically grinding up pitchers’ tendons in sadistic ways.

In part, this is due to the fact that mechanics ‘experts’ are trying to solve two different problems. First, the problem of velocity – how can we get a given pitcher to throw as fast as possible on a consistent basis? Second, the problem of attrition – how can we keep a given pitcher healthy? The two may or may not have the same answer, and it’s sort of odd to hear basically everyone agree that they DO have the same answer (even if they all think the other answers are crazy talk).

To bring this back to this post though, the mechanics experts would argue that you’re missing the point – that getting a mathematical equation of injury risk is impossible (for the reasons you laid out), and not at all what they’re trying to do. They’d say that one motion produces less soreness after X pitches, or another is comprised of disjointed motions such that the arm/elbow has to do more ‘work’ in producing velocity – and those factors, while not quantifiable in any meaningful way, tell pitchers important things. Given the heat in discussion about what motion produces or reduces these ‘risk’ factors, I’m with you – most people talking about this have absolutely no idea what they’re talking about, or they’re confusing genetic variation in mechanical terms: Nolan Ryan did this and he lasted a long time, so do this (the ultimate in results-based analysis). Still, I’m just not convinced that the very real and intractable data problems in constructing a meaningful injury prediction system/rubric/test should lead us to throw up our hands and say to hell with evaluating mechanics. Think of the same argument in hitting: a mathematical equation for the likelihood of making contact with a slider is probably impossible too, but that doesn’t mean a hitting coach doesn’t know how to teach it. Or maybe it does, I don’t know – I’m kind of down on hitting coaches right now (“uh, open up a bit more. Yeah, no, that’s it, you’re cured.”)

I’m not sure it should lead us the other way either – people are remarkably quick to ‘guarantee’ injuries in dozens of pitchers they’ve seen twice on TV, or once in grainy camcorder footage in a pre-draft video. I don’t know who’s right, and so your injunction to be wary of internet blowhards is absolutely spot-on. But what I’d like to see more than perfect models of material properties or joint anatomy is more basic descriptions of where/how stress is added. Given the complexity of how stress is ‘pathed’ through the joints, even that may be asking too much, but I keep thinking there must be a way to get biomechanics into internet biomechanics discussions.
.

but I never really understood what a bioengineer (or biomechanics engineer) does once they get out of academia. It seems like a lot do prosthetic limb stuff. Everything else I can think of seems more like biochemistry, genetics, etc.

because I could never write that post myself. But towards the end you wrote:

Mark Prior had perfect mechanics, remember?

There have been some articles over the past year or two that suggest differently.

My pet peeve is PAP and a slavish obsession with arbitrary pitch count numbers. I await with anticipation the day that PAP dies a well deserved death

I want more from Graham. :-)

Biomechanics aside though, I think we are finally starting understand what kind of mechanics are harmful in a pitcher’s motion. Everybody and their monkey has an opinion on mehcanics based on theory, but its only until recently that a lot of people have been paying attention to “results-based” mechanical analysis.

Its fascinating how similar the durable greats are, and how similar a lot of the guys are that breakdown. Greg Maddux pitches unusually long and unusually healthy. Figure out what he does. Mark Prior has unusually short career. Figure out what HE does. Other analysts like O’Leary (really its only an extension of Marshall’s work though) have done this this enough times, that they started recognizing patterns in the durable and fragile pitchers. That is definetly where the analysis needs to START.

You WANT to be doing what most of the durable guys are doing, you DON’T want to be doing what most of the fragile guys are doing. Its common sense IMO. Simple, but I think pretty effective, and is FINALLY getting the ball rolling in actual injury-risk analysis.

because this should really be one of the SBN-wide linked articles.

How do you get an article up featured throughout the whole network, anyways?

Or if they get seriously hurt were they always destined to get hurt (due to either genetics or mechanics)? They throw enough pitches over a career that it seems like luck wouldn’t play much of a role compared to other things.

It seems like a simple question but I don’t know the answer. If luck doesn’t play a role then figuring out what leads to injury should be easier (from a statistics prospective).

...any time I hear anyone say something about a pitcher being more or less at risk of injury because of their mechanics, they’re making it up?

Good to know.

It’s one I’ve thought about before, too: attempting to find a correlation between the types of pitches a pitcher throws - velocity, pitch type, pitch movement using pitch f/x data - and the frequency/severity of pitcher injuries. I think we generally agree that the twin aims of analyzing pitcher injuries are 1) to prevent future injuries and 2) to predict the likelihood of future injuries. This type of analysis would probably not be too useful in solving the first aim, because we would never be able to imply that an injury was caused by the pitcher throwing a certain type of pitch. In other words, suppose we found a correlation between throwing ~85mph sliders with ~6” of horizonal break and rotator cuff injuries. We still would not be able to conclusively say “Aha! Pitcher A hurt himself by throwing too many sliders”, and similarly it would be foolish to run around advising pitchers to stop throwing this pitch lest they damage their valuable shoulder.

I believe there is, however, a chance that this line of research could lead to some predictive power for future injuries. The advantage of comparing pitch f/x data to pitching injuries instead of comparing mechanics to injuries, is that pitch types are much more easily measured and quantified. One of the main points that Graham brought up is that a pitchers mechanics are so hard to qualify, let alone quantify. When you say a pitcher has “good” or “bad” mechanics or that this pithcer’s mechanics are “different” or “similar” to that pithcer’s mechanics, what exactly do you mean by that? How good? How different?

Unfortunately, when comparing pitch type to pitcher injury, you will run into a similar problem when attempting to measure and quantify injuries. Do you seperate injuries by type, count the number of injuries, duration on the DL, length of time between successive injuries? Going down this road could lead to some ugly arbitrary point scale a la Pitcher Abuse Points or the BCS. Perhaps if you stuck to basic, well-defined injury qualifications you could at the very least end up with a conclusion along the lines of - for example - pitchers who throw 82-88mph curveballs between 15 and 30% of the time incurr DL stints due to elbow injuries more frequently than pitchers who throw 82-88mph curveballs between 5 and 15% of the time. It is easy to see how this type of information could potentially be useful for baseball clubs as they build their teams.

I say all of this with the following qualifications:
1) Yes, this method is “working backwards”, and therefore 2) No, it will not get to the root of what causes pitching injuries. 3) This method could lead to a near infinite number of correlations studies for each different scenario, which would then lead to sample size issues, especially given the recency of pitch f/x data. Also, 4) I doubt you would find many real strong correlations using this method. Meaning, 5) This line of questioning, if pursued, might prove to be futile.

Out of personal curiosity: Graham, where have you studied biomechanics?

I’m studying biomedical engineering, but I dont know much about biomechanics (or biology for that matter). I focus more on electronic devices.

Basically I think Graham is the man, and is probably going to be the guru internet guy on mechanics in a few years.

Still, lets not get this twisted. These aren’t my ideas or anybody others. One line of thought that is out there today is basically an extension of Mike Marshall’s. This is the same guy that IS an expert on biomechanics (like Graham is going to be) amongsts many other things, was arguably the most durable reliever in the history of baseball (no injuries, 14 year career, record 13 consecutive games pitched, pitched over 200 innings as reliever in ‘74, etc.) is a coach, and has dedicated his life to learning the craft of perfect pitching mechanics.

http://drmikemarshall.com/

Here are some of his credentials:
http://drmikemarshall.com/AcademicCredentials.html

O’Leary, Ellis, and others (it seems like many other pop up here and there) are some of the guys that do results-based analysis and conform to one theory of healthy mechanics, but really the vast majority of it is just a big confirmation of what Marshall already figured out. Marshall already KNEW the effective and healthy methods to pitching mechanics because he had already been there done that, and dedicated his life post-baseball to intense study on the topic. The guy is WAY ahead of the curve in pitching mechanics, so it was no suprise to see a lot of people discover that most of the durable pitchers in baseball history did many of the things he reccomended.

The reason I don’t think hes as well known is beacuse hes such a pain in the ass to deal with that hes been kind of brushed away in the baseball world—but the guy knows his shit. In 10-20 years we’ll probably start recognizing what he discovered.

I do think there are advances being made, and that they are worth paying attention to. I also think that any mechanical analysis that doesn’t jibe with the succeses and failures of real MLBers isn’t worth paying attention to (theres a lot of info out there that is downright false in my opinion).