Quantitative vs. Qualitative Mechanical Analysis in Pitching Biomechanics

Written by Chris Langin

Introduction

“Johnny! If you used those legs, you’d throw a good 5-7 miles per hour harder!”

If you’re reading this blog, you’re likely familiar with the mechanical jargon used throughout the world of baseball.

From youth fields across the globe to the 9-figure stadiums that play host to the world’s most talented hurlers, there’s undoubtedly a coach, parent, or spectator discussing their displeasure with a pitcher’s mechanics at this very second.

At some point during your baseball career, you’ve likely found yourself in Johnny’s shoes. You’ve also recently noticed that the umpire has yet to throw somebody out of a game for throwing the baseball too fast.

This realization seems pretty insightful, but given certainty isn’t your strength, you discuss the topic with your family during dinner. Your intuition turns out to be correct; the pitcher is, in fact, not restricted to any sort of speed limit.

A few months go by and you’re not quite certain how fast you’re throwing the baseball, but you know you want to throw it faster.

Then, one day during your bullpen session, you receive some critical feedback regarding your pitching technique: “If you used those legs, you’d throw a good 5-7 miles per hour harder!”

At first, you’re frustrated that you’ve done something incorrect. Usually, this type of thing only happens when you forget to do your homework. However, you grin inside when it sinks in that using your legs better could improve your velocity.

This is some of the best news you’ve heard. Heck, you couldn’t even stand on two feet and walk at one point in your life. Getting into those legs a bit more during your pitching motion seems elementary.

Three days later, your next bullpen opportunity arrives. “Just use your legs,” you tell yourself. Your confidence is extremely high while you’re playing catch. You lift your lead leg up and down for several reps, acknowledging that you indeed will use this leg today.

You alert your catcher that the ball is going to be coming in firmer today. He smells like chlorine and doesn’t seem to care much. After mimicking a few more leg lifts for good measure, you finally toe the rubber. You close your eyes and take a deep breath. Finally, you convince yourself you’re ready for this life-changing moment.

You wind-up and leverage your legs for the first time during an overhead toss in your life. The ball arrives at the plate and the dirt in front of the mound remains undamaged. Your catcher lifelessly tosses the ball back to you.

“How was that?” you ask.

“Little slower than last week,” he responds.

After three more tosses that end in a similar result, you scream this question at the top of your lungs:

“How the hell do I use my legs?”

Motivation

Helping Johnny out is the motivation of our quantitative/qualitative study.

All performance coaches have good intentions:

Leverage the resources at your disposal to identify your athletes’ limiting factors.

Relay that information in a digestible manner, with the goal of improving that athlete’s performance.

With our motion capture lab, we’re able to precisely quantify how our athletes are moving to identify their movement limitations and track changes. However, most athletes and trainers don’t have access to a lab, so have to make due with what they do have. While the lab is the optimal solution for assessing mechanics, we’d be remiss if we claimed players and coaches are helpless without it.

While a trainer’s misplaced overconfidence in their understanding of how an athlete’s mechanics work could get in the way of the athlete, the other extreme of providing no feedback limits how much you can help an athlete.

Many of our pitching trainers at Driveline came from low-level college ranks. If a fresh set of baseballs wasn’t in the cards, there certainly wouldn’t be any motion capture labs in the budget.

While you can track if your athletes improve, there are a variety of competing variables that could be responsible such as strength, bodyweight, sleep, and simple throwing fitness improvements. Thus, it can be tough to know what is a result of your mechanical feedback (qualitative in the absence of a lab) or not, to use that as a proxy for how good you are at assessing mechanics.

This hindsight inspired us to partake in a qualitative/quantitative pitching kinematics study. Trainers could uncover their “true talent” level in evaluating kinematic positions in slow-motion. This also allows us to start peeling away at Johnny’s question: “How do I use my legs?”.

Johnny’s original question was a bit too broad, though he’s on the right track to question what exactly that meant. We intended to quantify some of the following mechanical jargon commonly passed around in the gym: 

• “Is this athlete linear or rotational?” 
• “Does this athlete have a pushy or whippy arm action?” 
• “Is this athlete glute or quad dominant?” 
• “Is this athlete’s pelvis open or closed at foot plant?”

Methodology

We wanted to identify whether there was validity to these qualitative categories. If so, what biomechanics actually define these classifications? The open/closed pelvis question is more easily defined. This question identified our trainers “true talent” in analyzing biomechanics qualitatively.

For reference, pitching assessments at Driveline last five days. On day one, athletes throw in the motion capture lab. Athletes have their biomechanics meeting on day three. The R&D department processes that report over that time period, giving us about a 48-hour window to provide a qualitative assessment, without it potentially being biased by having seen the actual report. 

For an athlete to qualify for our study, trainers/athletes must meet the following criteria:

• Must be the athlete’s first motion capture assessment at Driveline 
• Slow-Motion Video must be taken from the athlete’s open side 
• Trainers must complete questionnaires prior to Day 3 of their assessment 
• 3 trainer responses required

All in-gym trainers at Driveline took part in the study. To account for trainer bandwidth concerns, we assigned three trainers to each athlete. We took the weight of those three trainer samples over the weight of each individual trainer to get a better feel for department opinion. While we may incorporate this in a future analysis, there was a fear of significant noise given we didn’t have a 100+ athlete sample.

We used iPhone side-view video with a standardized set up. If parallax error was present, we omitted the athlete from the sample. We also saw some user errors occur when taking videos. We omitted athletes that had their videos return to regular speed prior to ball release.

This study started right at the start of collegiate summer. Our department analyzed the first athlete on May 3rd, and the last athlete on June 20th, creating a sample of 56 athletes. If two or more trainers chose the same positive response for a category, that athlete was binned accordingly.

Example: Two trainers responded with Athlete A being rotational, one responded “undecided”. That athlete would be a rotational thrower.

However, if the response was even across the board (one rotational, one linear, one undecided) we omitted the athlete from our analysis.

Results

Let’s first begin by looking at our “qualitative categories”.

Rotational vs. Linear Thrower

Most compelling differences:

• “Rotational” pitchers were more counter-rotated with their torso
• “Rotational” pitchers had more horizontal abduction (scap retraction) at foot plant
• “Linear” pitchers had a longer stride length
• “Linear” pitchers had a more open pelvis and torso at foot plant

Whippy vs. Pushy Arm Action

Whippy pitchers…

• Had higher max external rotation and shoulder external rotation at foot plant
• Were more counter-rotated with their torso
• Threw three miles per hour harder on average

Glute vs. Quad dominant

• “Glute” dominant pitchers had more shoulder abduction at foot plant
• “Glute” dominant pitchers had a more open torso at foot plant
• “Quad” dominant pitchers threw two miles per hour harder on average
• “Quad” dominant pitchers showed a more notable peak downward center of gravity (z) velocity (body movement down towards the ground)

And finally, our lone qualitative kinematic question to judge our trainers “true talent” in assessing positions with slow-motion side view:

Open vs. Closed Pelvis at Foot Plant

Open pitchers had…

• A more open pelvis and torso at foot plant
• Higher max hip-shoulder separation and hip-shoulder separation at foot plant
• Higher peak COG (x) velocities (vector pointing from second base to home)

Application

Slow motion video is a resource that should continue to be leveraged by coaches, trainers, and athletes. The pitching department could discern significant differentiators in pelvis angle at foot plant.

These findings could apply to daily throwing evaluation between a trainer and athlete. This study however doesn’t suggest trainers have the talent level to quantify kinematic positions at foot plant with even 90% certainty.

However, binning athletes in ~15 degree buckets (Open/Closed) could have some credence. Adding in multiple opinions from other trainers/coaches likely ups the certainty level in the accuracy of these buckets as well.

Any type of remote training especially gains value from these insights. While our in-gym population still benefits from feedback between retests, they’re still provided a retest report every 6 weeks that is void of subjective evaluation. Online trainers may have an athlete for 4+ months without a retest motion capture during that time period. The ability to bucket these athletes via collaboration between their co-workers and managers likely adds further accuracy to their evaluation.

In addition, our lab could detect (based on our trainers’ judgement) the validity of several qualitative categories.

Limitations

The methods for this study were not very rigorous. Largely, this study’s intention was to detect signal for future work on the topic. With limited previous work, an extremely rigorous protocol didn’t seem efficient given the scarcity of trainer hours during the summer.

Selection bias is also unavoidable in this study. Evaluators in this study were all trainers from Driveline. Trainers evaluate several hundred motion capture reports a year. Identifying how accurately an average coach, or another portion of the population, could make kinematic evaluations would be a significant insight.

Future Work

Further evaluation of qualitative categories is on the agenda. Quantifying these categories with kinematic variables provides us with more closure on what they’re detailing.

Potential examples include:

“Good/Bad Downward Move?”

“Vertical vs. Non-Vertical Shin Angle?”

“Heel Connection vs. Early Extension?”

The same exists with qualitative kinematic evaluations. It’s likely that certain kinematic variables (torso rotational velo,  are easier to evaluate without a motion capture session than others.

A study showcasing the error range to expect from the “coaching eye” provides coaches with some clarity during their evaluations. Certain metrics may warrant confident feedback when using side view as a resource. For others, perhaps voiding any opinion is best.

Potential examples include:

“Elbow Flexion Inside 90 degrees vs. Outside 90 degrees.”

“Shoulder Horizontal Abduction: Good/Average/Bad?”

Finally, we’d love to introduce a larger population into our evaluation pool. Detailing what the average coach’s evaluation looks like is the ultimate insight we’re looking for.

Most coaches/trainers aren’t going to have the luxury of constant access to a biomechanics lab. Furthermore, their livelihood isn’t dependent on delivering these mocap findings to athletes 365 days a year. The practical implications of these studies for a broader public depends on getting a study that includes evaluations from a variety of coaches.

And Johnny, to answer your question… just throw in a biomechanics lab.