Stephen Sutton-Brown
@srbrown70.bsky.social
2.8K followers 120 following 1.7K posts
Braves, heat maps, cars, tractors, and music Creator of StuffPro and Arsenal Metrics at Baseball Prospectus He / him
Posts Media Videos Starter Packs
forks and slurves, using statcast tags
describe some of what the models are "seeing" beyond just raw stuff, and two need to not be pitcher-agnostic to add predictive value to these types of models. IOW: Is this pitch more or less deceptive than expected due to who threw it and what their arsenal looks like?
One thing that I can't remember if I've said explicitly: Since working on modeling pitch detection I've found that current Stuff and Pitching models do a good job at "understanding" tunneling and deception at the population level. Tunneling and deception metrics thus, one, help
Correlation of a pitcher's average bat speed vs expected to various BPro metrics along with some updated leaderboards adjusting for what pitch types a pitcher throws and the counts they reach. Correlations are spearman correlations. Everything is min 500 swings against.
just the model notation, it'll solve for a batter specific intercept to replace the 1
of course, the model is set up as bat_speed ~ 1 + A*y_intercept + B*y_intercept^2 and I'm just solving for A and B by batter
Averaged by pitcher. This isn't controlling for pitch type, but still some interesting names at the top and bottom. Reminder: negative values indicate the batter reduced their bat speed relative to what we'd expected given the swing's intercept point
Difference in actual vs expected bat speed by pitch type. No surprises here: breaking and offspeed pitches fool batters
a collection of unique swings when modeling bat speed as a function of intercept point assuming constant jerk (thus a function of int and int squared)
If you model swings as having a constant jerk you can get around having to use a spline fit for bat speed vs intercept point. Useful because splines and GPs take way longer to fit with batter random effects
Reposted by Stephen Sutton-Brown
a pre-game meeting but instead of ppt slides of heat maps the dodgers fo showed him y'all's tweets
gonna go out on a limb and suggest snell pitch like this all the time
Where PitchPro thinks Snell *should* target his pitches assuming he has an average miss distance of around a foot for each offering
that was the best. forgot a screw at one point and said "dadgummit" and told Georgia to wait while I go get it and she said "it ok, i gonna fix it for you, daddy"
Bragging on Shelby, she cooked here
tbh this is not much different than how my desk usually looks
no daycare today and i'm trying to work while play doh colors are being mixed around me, i need to call osha
they're saying it's the most level a light fixture has ever been
Reposted by Stephen Sutton-Brown
It just dawned on me what we had postgame.
Reposted by Stephen Sutton-Brown
one is a genius, the other’s insane
miz and vaughn pinky and the brain