Stephen Sutton-Brown
banner
srbrown70.bsky.social
Stephen Sutton-Brown
@srbrown70.bsky.social
Braves, heat maps, cars, tractors, and music
Creator of StuffPro and Arsenal Metrics at Baseball Prospectus
He / him
80% average Pitch Type Probability on both the four seam and slider, which takes into account usage by count and handedness www.baseballprospectus.com/player/10612...
Ryan Helsley Player Card | Baseball Prospectus
www.baseballprospectus.com
December 1, 2025 at 9:33 PM
holy shit i know so few of these, i am washed 😭
December 1, 2025 at 9:30 PM
All three of those are good because the takeaways are language agnostic (despite the latter two being written for python). Statistical Rethinking specifically has been translated into a ton of programming languages.
December 1, 2025 at 3:34 PM
My #1 is always Statistical Rethinking from Richard McElreath (xcelab.net/rm/)

"Data Science from Scratch" & "Essential Math for Data Science" were great, & I went through them at the same time. Understanding the mechanics of data science allowed me to think more creatively about applications.
December 1, 2025 at 3:32 PM
happy thanksgiving to you too, bud, happy to hear y'all are doing well. miss the old online days
November 26, 2025 at 2:59 AM
energy has moved to Cars And Bids and Bring A Trailer. I'm stewing on not owning this Mustang bringatrailer.com/listing/1973...
351-Powered, Safari-Style 1973 Ford Mustang SportsRoof
Bid for the chance to own a 351-Powered, Safari-Style 1973 Ford Mustang SportsRoof at auction with Bring a Trailer, the home of the best vintage and classic cars online. Lot #221,153.
bringatrailer.com
November 26, 2025 at 2:30 AM
PL MAX
November 24, 2025 at 1:57 AM
but enough about you and me
November 23, 2025 at 12:52 AM
that's awful, so glad y'all are all safe
November 22, 2025 at 2:39 PM
Agreed, I think I just wish the presentation was more clear in that it's a rough 2d approximation, and is ultimately representing nothing more than the difference in the observed movement angle and statcast's 2d spin axis, which is already included in all major public stuff models
November 21, 2025 at 8:01 PM
Agreed. This dot product method seems useful as a quick approximation, though no new physics are being introduced and the conclusions drawn here will be no different than if using other methods that utilize statcast's spin_axis without the 3d spin vector.
November 21, 2025 at 7:44 PM
He's assuming that the magnitude of movement along the spin_axis direction is due to magnus, and any deviation from that is nonmagnus movement
November 21, 2025 at 7:23 PM
she's a serial overwriter, which some folks love to fawn over. all fluff, little substance, zero concision
November 21, 2025 at 6:47 PM
He shared on twitter, and it's not using 3d spin vector but instead using statcast's induced horizontal and vertical movement without gravity projected onto statcast's 2d spin_axis. So [horz,vert] dot spin_axis_unit_vector
November 21, 2025 at 6:39 PM