Ryan Gunther
@ryangunther1.bsky.social
- Writing a master's in statistics thesis that uses machine learning on baseball biomechanical data
- Occasional posts about progress on computer vision at the indy ball level
- Brock University MSc 2025
Github here: https://github.com/RyanGunther
- Occasional posts about progress on computer vision at the indy ball level
- Brock University MSc 2025
Github here: https://github.com/RyanGunther
98th percentile @fangraphs.com user, up from 93rd last year
December 5, 2025 at 5:12 PM
98th percentile @fangraphs.com user, up from 93rd last year
The most surprising part of this is that 17 people viewed Otto Lopez’s @fangraphs.com page more than me!
December 5, 2025 at 5:06 PM
The most surprising part of this is that 17 people viewed Otto Lopez’s @fangraphs.com page more than me!
Now we see why I maxed out at the collegiate level 🤣 just 11 big league homers were hit at or below this bat speed in 2025
(Me, today, measured by blast motion)
(Me, today, measured by blast motion)
December 4, 2025 at 11:54 PM
Now we see why I maxed out at the collegiate level 🤣 just 11 big league homers were hit at or below this bat speed in 2025
(Me, today, measured by blast motion)
(Me, today, measured by blast motion)
Hire me!
- Master's degree in statistics (thesis on hitting biomechanics)
- Experienced in computer vision, R, SQL, Python
- interned with 2 Indy ball teams, in pitch data (2024) & automation (2025)
- former collegiate player
- worked 3.5 years in finance/contract structuring
- Bachelors in finance
- Master's degree in statistics (thesis on hitting biomechanics)
- Experienced in computer vision, R, SQL, Python
- interned with 2 Indy ball teams, in pitch data (2024) & automation (2025)
- former collegiate player
- worked 3.5 years in finance/contract structuring
- Bachelors in finance
December 2, 2025 at 6:00 AM
Hire me!
- Master's degree in statistics (thesis on hitting biomechanics)
- Experienced in computer vision, R, SQL, Python
- interned with 2 Indy ball teams, in pitch data (2024) & automation (2025)
- former collegiate player
- worked 3.5 years in finance/contract structuring
- Bachelors in finance
- Master's degree in statistics (thesis on hitting biomechanics)
- Experienced in computer vision, R, SQL, Python
- interned with 2 Indy ball teams, in pitch data (2024) & automation (2025)
- former collegiate player
- worked 3.5 years in finance/contract structuring
- Bachelors in finance
In today's episode of unintuitive things I'm excited about - Isomap finds Ernie Clement and Ernie Clement's mirrored swing almost identical after dimension reduction!
This means that if lefties continue to have higher transformation scores than righties, I've found something significant!
This means that if lefties continue to have higher transformation scores than righties, I've found something significant!
December 1, 2025 at 11:20 PM
In today's episode of unintuitive things I'm excited about - Isomap finds Ernie Clement and Ernie Clement's mirrored swing almost identical after dimension reduction!
This means that if lefties continue to have higher transformation scores than righties, I've found something significant!
This means that if lefties continue to have higher transformation scores than righties, I've found something significant!
Here are two swings, same hitter, overlaid on each other, one makes contact with a ball in Zone 10, the other, Zone 11 (up+out vs up+in)
Barrels (red/purple) tunnel each other like 2 pitches tunnelling each other. But from overhead view you can see that they break off pretty early in the motion
Barrels (red/purple) tunnel each other like 2 pitches tunnelling each other. But from overhead view you can see that they break off pretty early in the motion
November 15, 2025 at 5:02 PM
Here are two swings, same hitter, overlaid on each other, one makes contact with a ball in Zone 10, the other, Zone 11 (up+out vs up+in)
Barrels (red/purple) tunnel each other like 2 pitches tunnelling each other. But from overhead view you can see that they break off pretty early in the motion
Barrels (red/purple) tunnel each other like 2 pitches tunnelling each other. But from overhead view you can see that they break off pretty early in the motion
Exciting new updates in baseball computer vision - I generated this (and many other clips like it) completely automatically through a pipeline I've built
Here's Adrian Del Castillo turning 96 above the zone into a gap double. Auto color-coded too, so the denser centers hold more weight in my model
Here's Adrian Del Castillo turning 96 above the zone into a gap double. Auto color-coded too, so the denser centers hold more weight in my model
November 14, 2025 at 6:53 AM
Exciting new updates in baseball computer vision - I generated this (and many other clips like it) completely automatically through a pipeline I've built
Here's Adrian Del Castillo turning 96 above the zone into a gap double. Auto color-coded too, so the denser centers hold more weight in my model
Here's Adrian Del Castillo turning 96 above the zone into a gap double. Auto color-coded too, so the denser centers hold more weight in my model
Putting the pieces together in my computer vision project - I've now taken Bogaerts swing footage, programmed the ability to automatically drill out "donut holes" i.e. the section between XB's arms at contact, then colour code so the centers of the data hold more weight
October 27, 2025 at 11:39 PM
Putting the pieces together in my computer vision project - I've now taken Bogaerts swing footage, programmed the ability to automatically drill out "donut holes" i.e. the section between XB's arms at contact, then colour code so the centers of the data hold more weight
No matter how I tune this, tree-based models only care to split for runners advancing from 2B on possible sac flies. From 3B it's pretty simple
Hangtime doesn't matter. Neither does OF initial positioning, launch angle, OF arms, runner speed, or where the ball's hit. If it's >260ft, you're scoring
Hangtime doesn't matter. Neither does OF initial positioning, launch angle, OF arms, runner speed, or where the ball's hit. If it's >260ft, you're scoring
October 24, 2025 at 3:57 AM
No matter how I tune this, tree-based models only care to split for runners advancing from 2B on possible sac flies. From 3B it's pretty simple
Hangtime doesn't matter. Neither does OF initial positioning, launch angle, OF arms, runner speed, or where the ball's hit. If it's >260ft, you're scoring
Hangtime doesn't matter. Neither does OF initial positioning, launch angle, OF arms, runner speed, or where the ball's hit. If it's >260ft, you're scoring
Finally figured this out on Monday. The below video isn’t perfect (helmet..?), but as long as it works conceptually, I can add more training data + annotations and build it into a smooth pipeline within a week or two
October 22, 2025 at 5:35 PM
Finally figured this out on Monday. The below video isn’t perfect (helmet..?), but as long as it works conceptually, I can add more training data + annotations and build it into a smooth pipeline within a week or two
I’ve struggled recently with trying to get my computer vision biomechanics model to properly segment the entire hitter’s body, but NOT the “donut holes” i.e. background that lands between the arms and bat, circled below
October 22, 2025 at 5:30 PM
I’ve struggled recently with trying to get my computer vision biomechanics model to properly segment the entire hitter’s body, but NOT the “donut holes” i.e. background that lands between the arms and bat, circled below
I’ve also never seen a baserunner headfirst slide back to a base that isn’t being thrown to. Watch top left of the screen 🤣
October 5, 2025 at 9:24 PM
I’ve also never seen a baserunner headfirst slide back to a base that isn’t being thrown to. Watch top left of the screen 🤣
Not sure I’ve ever seen a pitcher shake his head as he comes set. Either you shake off a pitch or you’re pissed about a bad call. But a borderline call went his way the pitch before this
Last pitch he threw today
Is this a “I hate the pitch call but I’ll go along with it anyway?”
Last pitch he threw today
Is this a “I hate the pitch call but I’ll go along with it anyway?”
October 5, 2025 at 9:23 PM
Not sure I’ve ever seen a pitcher shake his head as he comes set. Either you shake off a pitch or you’re pissed about a bad call. But a borderline call went his way the pitch before this
Last pitch he threw today
Is this a “I hate the pitch call but I’ll go along with it anyway?”
Last pitch he threw today
Is this a “I hate the pitch call but I’ll go along with it anyway?”
made some 3d visualizations of a motion capture session. batter approaching, making, and then following through his contact point
September 29, 2025 at 4:24 AM
made some 3d visualizations of a motion capture session. batter approaching, making, and then following through his contact point
A common issue with using computer vision for biomechanics is that baggy jerseys/pants, shin/elbow guards and long hair become part of the data. To combat this, I've used color gradients on my data - the edges now hold less importance than the denser centers
September 26, 2025 at 5:07 AM
A common issue with using computer vision for biomechanics is that baggy jerseys/pants, shin/elbow guards and long hair become part of the data. To combat this, I've used color gradients on my data - the edges now hold less importance than the denser centers
Here's my FIRST test of my self-trained computer vision segmentation model on completely unseen footage. Overall, SUPER happy with the results! A liiiittle bit of flickering and we lost a lower leg in the follow through. Nothing a bit more training data can't fix. A breakthrough for sure
September 12, 2025 at 4:44 AM
Here's my FIRST test of my self-trained computer vision segmentation model on completely unseen footage. Overall, SUPER happy with the results! A liiiittle bit of flickering and we lost a lower leg in the follow through. Nothing a bit more training data can't fix. A breakthrough for sure
I'm about 300 frames of manual annotation in to my segmentation project. Should be enough to begin training and automate this (incredibly lengthy) process
September 9, 2025 at 3:59 AM
I'm about 300 frames of manual annotation in to my segmentation project. Should be enough to begin training and automate this (incredibly lengthy) process
August 30, 2025 at 2:33 AM
I really like this paragraph I wrote tonight, summarizing a chapter of my work. There are bad / inefficient ways to move, but once you get past that, there are many different ways to get to successful swing outcomes in baseball
August 26, 2025 at 4:32 AM
I really like this paragraph I wrote tonight, summarizing a chapter of my work. There are bad / inefficient ways to move, but once you get past that, there are many different ways to get to successful swing outcomes in baseball
Made with inspiration from ViTPose in Google Colab. Will put the code on my Github this week
August 23, 2025 at 2:56 AM
Made with inspiration from ViTPose in Google Colab. Will put the code on my Github this week
I wonder if Andy Pages knows how much effort I've put into tracing his body to build a training set
August 20, 2025 at 5:14 AM
I wonder if Andy Pages knows how much effort I've put into tracing his body to build a training set
Hi Tom!! I'm curious how accurate / reliable the body movements (white dots below) in savant's swing visualizer can be trusted to be? I do notice variation in movement from batter to batter, so clearly there's some specificity to this, but would this be considered accurate for an avg swing?
July 9, 2025 at 4:50 PM
Hi Tom!! I'm curious how accurate / reliable the body movements (white dots below) in savant's swing visualizer can be trusted to be? I do notice variation in movement from batter to batter, so clearly there's some specificity to this, but would this be considered accurate for an avg swing?
I can’t sleep so I’m working on computer vision instead
July 8, 2025 at 5:39 AM
I can’t sleep so I’m working on computer vision instead
green dots are original marker placement, purple dots are the model's "best guess" at where those body parts should be based on the position of the red dots at that frame. This is a spectacular result
July 6, 2025 at 12:33 AM
green dots are original marker placement, purple dots are the model's "best guess" at where those body parts should be based on the position of the red dots at that frame. This is a spectacular result
(1/2) How I broke swing biomechanics into entirely data-driven segments, dynamic to each hitter and stance:
- Determine principal features of the swing - the markers that explain the most variance (a lot of biomech data is redundant)
- group the non PFs as isotropic (same in all directions) noise
- Determine principal features of the swing - the markers that explain the most variance (a lot of biomech data is redundant)
- group the non PFs as isotropic (same in all directions) noise
June 15, 2025 at 12:27 AM
(1/2) How I broke swing biomechanics into entirely data-driven segments, dynamic to each hitter and stance:
- Determine principal features of the swing - the markers that explain the most variance (a lot of biomech data is redundant)
- group the non PFs as isotropic (same in all directions) noise
- Determine principal features of the swing - the markers that explain the most variance (a lot of biomech data is redundant)
- group the non PFs as isotropic (same in all directions) noise