Indev

Leaving the bottom-zone model as is for the time being to get the rest of the zone built out. Current bottom zone model is around r=0.2 Top zone model is next up, currently sitting at ~r=0.1, process is going a lot faster compared to bottom zone since the methodology is built out for the most part.

Pitch framing matters just as much, if not slightly more in college ball compared to the MLB. College hitters consistently take more competitive pitches (>25% called strikes) compared to their MLB counterparts

Zone may have opened up a bit for Scherzer, but incredible game from Valenzuela. Almost a full framing run and ~6.2 strikes added

Current benchmarks for the CV framing model (v9). These are its top 10 moves out of the ~10,000 pitch library.

Clinic from their backstops

Modest but real correlation to calledstrike% w/ the current version of the model. Also interesting to note, the size of the move DOES matter according to my data, with bigger moves having a slight edge over smaller ones.

Made some pretty substantial improvements to the model; it does a good job of recognizing very good from very bad but still working on the stuff in between haha Forcing the model to recognize the load as the lowest glove point now, it just wasn't consistent enough picking up the load automatically. Stray glove detections also aren't really an issue for the data side anymore, they still appear on annotated clips sometimes but don't make it into the CSV

I really like this move, drops the hip down and hinges thru the catch very nicely

Probably a good sign if the cv model is flagging moves like this

Very very true! Carter Jensen 09/07

A few good moves from Salas in his debut. Didn't get great results but the skill is there so I think he'll be alright

With kickstand stances becoming more common in college and the MLB, i'm trying to quantify how big the increase in called strike production needs to be to offset the loss of blocking ability. Based on 2025-2026 MLB Statcast data

Table is based on MLB data; working on seeing if the relationship stays consistent at the college level but the PB/WP run value is tough to work out

Great move from Jensen. Could stand to be slightly less vertical with the glove load, notice how his glove naturally reorientates itself to closer to 45 degrees before the actual vertical move starts. Note the LKD kickstand usage👀

Haven't really seen a move at the top like this before. Thought it may have been accidental after only seeing the first move but after seeing the second it seems intentional🤷‍♂️