The Real Reason I Haven't Updated In Awhile
by Kyle Boddy on Nov 7, 2009 4:20 AM PST
in News
26 comments
While I assume these types of posts get repetitive over time, I feel obligated to at least inform my audience why I haven't been posting much other than the stock "I'm busy" or "I am working with my coaching/training site." Frankly, neither of those excuses are good enough, so I wanted to clear the air for the real reason that posts having been flying up with mechanical jargon and animated GIFs that you guys love so much: I'm worried that what I say is detrimental to the study of pitching mechanics.
There, I said it: I'm afraid that every mechanical breakdown of a pitcher I make is really just undermining the work of people like Dr. Fleisig at ASMI and Eric Cressey at Cressey Performance. So, I've decided to do something about it. During the long layoff, I've been doing a lot of research and interviewing to learn more and attempt to validate my beliefs, even if only in theory. The clients that I've coached and trained have increased their release velocity and pitch frequently without pain or soreness, which is a good start, but it's ultimately just anecdotal evidence that doesn't prove much. One of my clients is playing winter ball and was hesitant to go to the pitching coach, telling me that he didn't want to listen to a traditional coach who can't justify his views. I told him that such blind loyalty is dangerous, and that he should absolutely keep an open mind. I convinced him to lend his ear to the coach and take his advice for a trial run. The coach told him basic stuff like "stride more closed," "develop a short arm action," and "load the scapulas." The results were predictable - just weeks into the program, he developed pain and soreness in his throwing arm's bicep and anterior shoulder with decreased velocity and control.
One person who was kind enough to meet with me is PhysioCare's Andy Lodato, who graciously took time out of his busy day to show me the finer points of measuring shoulder internal/external rotation using a goniometer, a process described by Eric Cressey in the excellent article "What I Learned in 2008" on T-Nation.
If anyone in the Seattle area is in need of physical therapy, I highly recommend that you check out Andy's clinic in Woodinville. They have a great fully-equipped facility with a kind staff who knows their stuff.
In doing all this research and study, I've built up the basic page of my coaching/training site and have begun to sketch an opening assessment that will be done on all incoming clients - one such test is the total motion about the shoulder joint as described above, using a goniometer. And by studying those fields, I've decided that establishing a baseline is very important to any researcher's corner. As such, I'll be investing my own hard-earned cash to conduct research studies using high-speed video and imaging software to build a report much like ASMI's computerized biomechanical breakdowns. I'll be renting batting cage time, investing in the necessary lights and backdrop, and offering free pitching lessons to youth pitchers in order to measure a handful of variables to see what correlates best with release velocity and to see if there's any patterns with self-reported pain/soreness and mechanical flaws. While I doubt my own research will turn up anything groundbreaking, I look forward to building a portfolio of youth pitchers that I can review and perform calculations on over time.
And lastly, you'll all be happy to know that through a lot of my research and study that I've become more rock solid in some of my beliefs and that I've learned a ton over the past few months. I hope to be able to share the results of my introductory research study over the next year and see if we can come up with some interesting conclusions. In the meantime, you can look forward to a few mechanics posts coming this week.
As always, thanks for your patience and your understanding. Remember - no one has all the answers. I'm not even sure that we even have all the questions. Keep an open mind!
0 recs |
26 comments
Email
Print
Comments
Eagerly looking forward to the results...
…even though the longitudinal stuff could take years. Either way, the research sounds like it will be “ground-breaking” to me!
by TJ Eckleburg on Nov 7, 2009 5:56 AM PST reply actions 0 recs
I expect my research...
…to more or less confirm what ASMI already has – that front-knee stabilization, % of MER, and late torso rotation all contribute significantly to release velocity. What I’m really interested in are things that ASMI doesn’t necessarily measure right now – X-acceleration graph of the baseball, tracking the center of mass, and a few other variables that I’m researching right now.
Webmaster of Driveline Mechanics
http://www.drivelinemechanics.com - An Unconventional Look at Scouting
by Kyle Boddy on Nov 9, 2009 1:41 AM PST up reply actions 0 recs
Presumably your x-axis is a horizontal axis from the mound to home plate. If that’s the case, there are couple issues with calculating the x-acceleration.
1) Linear velocity OF THE PITCHER toward home plate is important, but it’s not the only variable to consider. Pitching a baseball at high speeds takes advantage (primarily) of rotational mechanisms.
Consider the following situation. You have a pitcher who (with his throwing arm abducted to 90 degrees and his elbow extended) can spin about a vertical axis passing through his c.m. If the pitcher received an angular impulse that accelerated him through, say, 360 degrees (way too large, and it never happens in real life, but I’ll use it to illustrate a point), then there will be times that the ball (as seen from the side) will be moving in the negative x direction. Despite the fact that the ball is being angularly accelerated so that it’s tangential velocity is always increasing, you’re going to find a negative x acceleration.
Now, of course, pitchers do not rotate 360 degrees before they release the ball, but they do rotate. As seen from the side, you might misinterpret a small linear acceleration in the direction of the pitch as insignificant and a non-factor in generating ball velocity. But that might not true, because the ball is probably being angularly accelerated. Below is a picture that will help illustrate that point.
2) I think all you’re going to see with the acceleration of the baseball is that for most of the pitch the ball does not have any pronounced acceleration and then for the last 0.15 seconds or its acceleration increases. What is this going to tell you? It does not say anything about how the rest of the body is moving. At best you can calculate the distal force made on the ball by the wrist, but that’s it.
I’m not sure what you hope to find by measuring that variable.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 9, 2009 8:46 AM PST up reply actions 0 recs
Also, any time you are dealing with acceleration data, it’s tricky to believe what you get. Acceleration is a second-order derivative, and as a result, produces excessive noise. Smoothing can work to some extent, but there are pitfalls in that as well.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 9, 2009 11:50 AM PST up reply actions 0 recs
The X-acceleration is important because of the vector required to deliver a pitch to the plate.
The pitch itself will fairly precisely travel along the X-axis toward the target. With rotation involved in the acceleration of the baseball, the peak velocity in the positive X direction will be shortly after the arm reaches a position that is perpendicular to this axis.
Because the ball is being thrown to a target roughly on the X-axis, it makes plenty of sense to measure its acceleration along that axis. The ball’s X velocity peaks before pitch release, there’s an improvement that can be made.
by NoNameOnCard on Nov 9, 2009 11:28 PM PST up reply actions 0 recs
Correction to last sentence....
If the ball’s X-velocity peaks before pitch release, then there’s an improvement that can be made.
by NoNameOnCard on Nov 9, 2009 11:29 PM PST up reply actions 0 recs
My point was that from 2D coordinate data (which I believe is what Kyle is using at the moment) you can only measure the x-position of the ball. If you attempt to calculate linear velocity or acceleration from just the these coordinates, you will arrive at false conclusions. Namely that at some time, the ball got a larger linear acceleration in the x-direction than it actually did.
Why does this happen? Because at different times, the ball receives linear accelerations in the x AND the y directions, increasing the magnitude of its velocity. As the pitcher’s arm rotates and gets closer and closer to release, the projection of the ball’s velocity vector onto the x-axis gets larger. The arm can continue to be accelerated in the x-direction, but it’s linear velocity will not have increased as much as someone analyzing 2D data might think.
I assume you mean the path of the ball after it leaves the pitcher’s hand; but if not, I have a big problem with this statement:
he pitch itself will fairly precisely travel along the X-axis toward the target.
Yeah, you want to throw a pitch along the x-axis, but until just before release, the ball does not move that way.
That’s a whole lot of x-axis deviation.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 10, 2009 5:08 AM PST up reply actions 0 recs
I read your post again, and I think it’s clear that you were referring to ball trajectory once it leaves the hand. So my bad on that.
But I think it’s clear that prior to release, the ball’s path is not confined to the x-direction. Significant amounts of the ball+wrist system’s velocity are generated in the y-direction.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 10, 2009 7:21 AM PST up reply actions 0 recs
But really, the biggest issue is the level of believability in acceleration data. Below is a series of unsmoothed raw coordinate data.
It isn’t beautiful, but this is the nature of all raw data. Whether you collect it from a marker system or you collect it by manually digitizing body landmarks, you will always have some noise.
When you calculate velocity from the raw data you get something a little messier.
This velocity is calculated from the same location coordinate data. This is only the first derivative and we have peaks ranging from ~ -15 m/s all the way to ~30 m/s. But you can still see a slight upward trend. In fact, a linear regression quickly reveals a slope of about m=29.
When you calculate acceleration, you’re in for some serious trouble.
Now we have peaks anywhere from -11000 m/s^2 to 10000 m/s^2.
This is unavoidable.
Smoothing takes some of this out. But at the same time, you lose a lot of the detail of your data. Additionally, during smoothing, a single bad point can drastically alter whatever polynomial you fit to your data.
Below is actual force data (directly proportional to acceleration) from a pitching study (Feltner & Dapena, 1986) that is already smoothed. I don’t know what the cutoff frequency was for this particular set of data, but these are the types of data that get reported by ASMI, your neighborhood biomechanist, or whoever. At the moment there is no real way to get around it. Thus is the nature of second-order derivatives.
Its difficult to see the scale, but there are force values ranging from -3000 to 3000 N. This still needs some smoothing, again, you lose a lot of the features and detail of a data set by smoothing.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 10, 2009 6:48 AM PST up reply actions 0 recs
To summarize my problem with all of this:
Measuring the apparent acceleration of the ball in an XZ plane has many problems. The biggest is that if the ball has a velocity in the y-direction and this velocity is redirected centripetally through some arc, the projection of the velocity on an XZ plane will become larger up to a point. This increased projection might be misinterpreted as a changing speed though in reality it is only the direction of the vector that is changing, not the magnitude.
To complicate this, acceleration is a second order derivative and data obtained in this manner is inherently noisy.
The solution would be to reconstruct the ball’s path in three-dimensions. The components of the ball’s acceleration in the X, Y, and Z directions and the resultant vector’s direction and magnitude can be calculated.
Filtering the noise can be done using low-pass bandwidth filters to smooth the 3D location data. Fourth degree polynomials can be fit to these coordinate data and differentiated to generate instantaneous velocity and acceleration of the ball. The data should then be resampled at approximately the same frequency it was recorded to produce discrete data points.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 10, 2009 7:36 AM PST up reply actions 0 recs
My explanation of why measuring x accel is a bad idea was kind of poor. This graphic should clarify things a little bit.
Basically you end up measuring a linear acceleration that isn’t really there. The only acceleration acting is centripetal, which acts to increase or decrease the projection of velocity onto the x axis.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 10, 2009 8:36 AM PST up reply actions 0 recs
You're thinking too mathmatically about this.
The end goal is velocity along the X-axis. Put simply, a faster pitch is going to have better X-axis acceleration than a slower pitch. This is indisuputable, agreed?
Measurements based on the x-coordinate of the ball will be imprecise to say the least, and yes, this will make it very hard to determine whether or not the ball reaches peak velocity prior to release rather than at release.
To sum up: the measurements/technology makes this nearly impossible to figure out with any kind of precision, but the logic is perfectly sound.
by NoNameOnCard on Nov 11, 2009 12:10 PM PST up reply actions 0 recs
The end goal is velocity along the X-axis. Put simply, a faster pitch is going to have better X-axis acceleration than a slower pitch. This is indisuputable, agreed?
Not quite. Think of it this way… Imagine a pitcher could accelerate a baseball from rest to 100 mph while it were still in his hand. Rather than accelerate it straight toward home plate, he can only do this in a direction perpendicular to home. If he could then redirect the velocity of the ball to home (this is the centripetal acceleration) then he could have a 100 mph pitch in the direction of home plate. Right? The magnitude of the ball’s velocity never changes after it was accelerated to 100 mph in the y direction. All that’s changed is the direction of the velocity vector.
This is a drastic oversimplification because no pitcher accelerates a ball strictly in the y-direction. But the ball path from separation to release is anything but linear, particularly the acceleration phase. The ball will acquire its speed by accelerations in the x, y, and z directions. This way, a centripetal acceleration will have constrained the ball to move in a circular arc, redirecting all (most) of the speed into a velocity vector pointing in the x direction at release.
The problem is that unless the ball is moving strictly in the xz plane, Kyle is not going to have a clue as to the actual speed of the ball, because he can not measure motion in the y direction. As the y velocity gets redirected toward home plate, it will appear as if the ball suddenly acquired a bunch of speed.
I know what Kyle is trying to do, but he isn’t measuring what he thinks he’s measuring. He’s going to end up with the x acceleration AND a sinusoidal term as the angle between the ball’s velocity and the xz plane changes.
My suggestion would be to film with two cameras and then use DLT to reconstruct the ball’s path in 3D. Then you’re only dealing with the noise from digitizing.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 11, 2009 12:26 PM PST up reply actions 0 recs
I have an idea of how to rectify this problem. If you could reconstruct the ball’s path in 3D and fit polynomials to the smoothed data so you have continuous functions, then you’d have no problem finding the instantaneous velocity and acceleration vectors and from that you could find the force acting on the ball. If you were to project the instantaneous force vector onto the velocity vector you would find the magnitude of the force that serves to increase the ball’s speed. The other orthogonal components are centripetal and act to keep the ball in a curved path. This would actually be interesting.
The end goal is velocity along the X-axis. Put simply, a faster pitch is going to have better X-axis acceleration than a slower pitch. This is indisuputable, agreed?
And to elaborate on what I said before, you don’t necessarily need a big acceleration at release, you want a big linear velocity. If I five miles to accelerate my car to 125 mph I would still be traveling as fast as the high-performance car that can accelerate to the same speed in only one mile.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 11, 2009 2:52 PM PST up reply actions 0 recs
I absolutely understand everything you're saying, but I think you aren't quite understanding me.
The origin of the initial velocity isn’t important, only that velocity toward the plate is NOT DECREASING before the pitch is released.
I’m not sure if this is exactly what Kyle is after, but it’s the impression I get from reading this (or gave myself when I decided to jump in) .
by NoNameOnCard on Nov 11, 2009 7:07 PM PST up reply actions 0 recs
The origin of the initial velocity isn’t important, only that velocity toward the plate is NOT DECREASING before the pitch is released.
I can agree with this statement for the most part. But just to be clear, it would be fine if the acceleration were to decrease (and in fact it likely will; as the arm moves faster and the muscles go into faster concentric conditions and make smaller forces). What is important is that the acceleration does not go below zero and the arm slows down. So yeah, I think I can get behind this statement.
I just hate the idea that anybody is working hard researching something, only to grossly misinterpret their results. If all he wants is to verify that the arm isn’t slowing down before release, then he’s pretty much ok because at release the arm is basically moving in the xz plane (although that’s an oversimplification). But if he’s trying to measure when the ball acquires most of its speed, then he’ll need a new method.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 11, 2009 7:42 PM PST up reply actions 0 recs
What is important is that the acceleration does not go below zero and the arm slows down. So yeah, I think I can get behind this statement.
This is exactly what I am looking into.
Sorry if I was unclear and non-responsive. I was enjoying reading this exchange and didn’t want to interject.
Webmaster of Driveline Mechanics
http://www.drivelinemechanics.com - An Unconventional Look at Scouting
by Kyle Boddy on Nov 13, 2009 3:07 PM PST up reply actions 0 recs
Agreed.
This is what I was getting at. Sorry if I was unclear.
by NoNameOnCard on Nov 15, 2009 9:51 AM PST up reply actions 0 recs
I’m usually critical of pitching mechanics blogs because of presentation, typically not content. I have a problem when authors present theory as if it were fact. I think this leads to a recycling of ideas, where readers are led to believe certain “truths” that may or may be correct.
The fact that you openly acknowledge this uncertainty is, to me, something that should not be understated. I would encourage you to continue writing, continue analyzing, and begin pursuing whatever personal research you might be interested in.
ASMI does not do that great of a job when it comes to researching pitching mechanics; they simply have more means than the average researcher. So critics be damned. As long as you’re not presenting your analyses as fact, but rather an educated opinion, I say keep going. How on earth could it actually be detrimental?
And as a selfish plug for my own research, I’m working through my thesis and finding some interesting stuff about the role of the stride leg, glove arm, and a dragging of pivot foot and how they actually contribute to pitching velocity.
My uncle says you've got a screw loose.
Your uncle molests collies.
by gorilla_baller on Nov 7, 2009 7:09 AM PST reply actions 0 recs
Thanks
As long as you’re not presenting your analyses as fact, but rather an educated opinion, I say keep going. How on earth could it actually be detrimental?
I just fear that my articles and opinions will be taken out of context. I know I can’t control this, but it bothers me all the same.
Webmaster of Driveline Mechanics
http://www.drivelinemechanics.com - An Unconventional Look at Scouting
by Kyle Boddy on Nov 9, 2009 1:41 AM PST up reply actions 0 recs
any patterns with self-reported pain/soreness
How would you measure this? Seems that every pitchers experience with pain is different than others. One pitcher might be used to the same amount of pain/soreness as another pitcher, and will rank the pain/soreness lower than a pitcher that is not used to pain/soreness even though they have the same amount.
Giant Dirtbags: John Bowker, Steve Hammond. MIA List: Todd Jennings, Brian Anderson
Jeremy Affeldt induces DP's
by Giant among Angels on Nov 7, 2009 9:09 AM PST reply actions 0 recs
If I may...
If you’ve ever been in a hospital, you have the classic 1 through 10 pain scale. It’s subjective, but it’s perfect because pain is relative.
The raw numbers themselves aren’t as important as the changes in the numbers are. If someone goes from 3 to 4, you should take notice. If someone goes from 6 to 2, that’s probably a good sign.
It’s really just another piece of the pain puzzle. The more information you can gather, the more you have to work with.
by NoNameOnCard on Nov 7, 2009 9:49 AM PST up reply actions 0 recs
Well said.
Trip basically outlined what I wanted to say – that pain is relative. All that matters is that we have a consistent measurement and enough data points to give us something to work off of.
Webmaster of Driveline Mechanics
http://www.drivelinemechanics.com - An Unconventional Look at Scouting
by Kyle Boddy on Nov 9, 2009 1:42 AM PST up reply actions 0 recs
Not to nitpick
Pain scales go from 0 to 10. You always want to give the patient or client the chance to have no pain (0 score). Often overlooked with pain is the type of pain: soreness, burning, sharp, achy, etc.
by seattle_pt1 on Nov 20, 2009 8:06 PM PST up reply actions 0 recs
You could arbitrarily set 1 to mean "no pain."
but I understand what you’re saying.
by NoNameOnCard on Nov 22, 2009 1:57 PM PST up reply actions 0 recs
Comments For This Post Are Closed
Secondary Sidebar
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2010 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
