Kyle Bland

Kyle Bland

@blandalytics.pitcherlist.com

Trying to make analytics a little less boring. All takes backed by a colorful chart Dir. of Data Analytics & Research @ https://pitcherlist.com/. He/Him

4K followers697 following1.4K posts

Kyle Bland reposted

One wonderful thing you can do with your money is give some of it to me every month, by subscribing to my frankly excellent website, hockeyviz.com/subscribe. This is wonderful in two ways, because I get your money and you get beautiful understanding of hockey. This thread has some details!

If you haven’t seen it, CJ’s site is *awesome*

CJ Turtoro@posthockey.com

ICYMI from the other site. I’ve written about hockey analytics for 15 years. In that time, I always wanted a site that dug into not only IF a player is good, but when, why, and how. Something to help writers and fans tell those stories. 3 months ago I started building it ... and today, it's live!

This, but for me / @pitcherlist.com

Micah McCurdy@hockeyviz.com

If you are a human who pays attention, you already know how sports is a mixture of joy to suffering. With math and a lot of computers, though, you can render both the the joy and the suffering crystalline. If you like that, you should subscribe, and we will both prosper. Here's to a fun season.

Kyle Bland reposted

Replying to @hockeyviz.com

If you are a human who pays attention, you already know how sports is a mixture of joy to suffering. With math and a lot of computers, though, you can render both the the joy and the suffering crystalline. If you like that, you should subscribe, and we will both prosper. Here's to a fun season.

Kyle Bland reposted

i think this needs to be broken up into divisions: - beefcake (the big bois) - boy band (tiger beat looking mfs) - romance novel protagonist (conventionally attractive) - hear me out

Snowden St.@snowden.st

who’s the hottest guy in MLB? i’m not an expert in this matters. Is it, like, Julio Rodriguez? Dansby Swanson?

You can do this on any pitcher, pitch type, or outcome you'd like! blandalytics.com/model-drilld...

Kyle Bland@blandalytics.pitcherlist.com

Seeing that a player has good Stuff is great, but it's important to make a best effort to show *how* a model gets to its outputs and/or *why* the Stuff is good. This allows models to be better interrogated and improved ("Here's why it's wrong" is better than "It's wrong") (8/8)

"Why doesn't he get whiffs if he has good Stuff?" Current models can be opaque in how they get to their conclusions. To combat this, we have developed a process to illustrate how and why a player's pitches are good/bad/neutral, based on our models. (1/8)

View full thread

Batted ball outcomes (Out, 1B, 2B, 3B, HR) can be aggregated to estimate "damage allowed on contact" (wOBAcon). Ex: Paul Skenes really limits his damage on contact with his Changeup, thanks to where he locates it and how he deploys it. (7/8)

Seeing that a player has good Stuff is great, but it's important to make a best effort to show *how* a model gets to its outputs and/or *why* the Stuff is good. This allows models to be better interrogated and improved ("Here's why it's wrong" is better than "It's wrong") (8/8)

"Why doesn't he get whiffs if he has good Stuff?" Current models can be opaque in how they get to their conclusions. To combat this, we have developed a process to illustrate how and why a player's pitches are good/bad/neutral, based on our models. (1/8)

View full thread

Using an outcome-based modeling philosophy allows us to break down which predicted outcomes are driving the pitch's value. Ex: Justin Wrobleski's Slider is net positive thanks to the value it gets from earning whiffs and avoiding balls, even with harder contact allowed. (5/8)

With pitch-level outcome estimates, we can drill deeper into "why" the model predicts a given outcome. Ex: Cristopher Sánchez's extreme Horizontal movement on the Sinker adds a ton to Called Strike%, even before you consider count and location. (6/8)

"Why doesn't he get whiffs if he has good Stuff?" Current models can be opaque in how they get to their conclusions. To combat this, we have developed a process to illustrate how and why a player's pitches are good/bad/neutral, based on our models. (1/8)

View full thread

Performing that analysis on a given pitch type's Stuff yields a breakdown highlighting what features impacted the Stuff value and by how much. Ex: Jacob Misiorowski' Curveball is so nasty because its extremely fast, while maintaining a solid velo gap to his fastball. (4/8)

"Why doesn't he get whiffs if he has good Stuff?" Current models can be opaque in how they get to their conclusions. To combat this, we have developed a process to illustrate how and why a player's pitches are good/bad/neutral, based on our models. (1/8)

At a high level, pitch models take various features and use them to value a pitch. An inexhaustive list of features includes: - Physical features: velocity, shape, spin, and release location - Location features at plate - Context features: count, pitcher/hitter handedness (2/8)

Kyle Bland reposted

A lot of my baseball career has been in analytics but many probably didn’t know I also lead our mental performance group, the psychology and cognition department. This consists of me, several researchers, and coaches who apply our findings back to the players and coaches

Just walked through a “House Above Replacement” analysis to my wife for house hunting. If you ever wondered if I can ever “turn off the nerd”, no. No I can’t.

If you aren’t on Twitter and haven’t seen Tom’s posts about OpenCommand, now’s your chance! Lots of really cool stuff can come from this project.

"Le Chonk"@cdgoldstein.baseballprospectus.com

this is a really cool article on OpenCommand, and it is completely free to read (not even a basic sub required):

Kyle Bland reposted

HockeyViz doesn't even have any ads on its free tier.

IGN@ign.com

Disney+ has altered its user agreement to clarify that ads will be allowed on all subscription tiers. https://bit.ly/4hhOjBo

Kyle Bland reposted

I assume that when whoever is making it IPOs they will reveal the full lineup implied by the existence of Tilly Norwood.