What do you get when you cross professional baseball ⚾️with Bayes ? Bayesball! I am super excited to announce a new paper out with @KordingLab, titled "Bayesball: Bayesian Integration in Professional Baseball Batters" https://t.co/w6JmjlP7d7
In our new paper https://t.co/uwAtDIaJwu , @skdeshpande91 and I develop a framework for estimating the optimal swing decision of batters in Major League Baseball. We take a Bayesian approach to estimating the expected runs following a swing and take decision. (5/7)
@paulgribble @KordingLab Yes, this is the version we adapted ours from. Its pretty much the same with sans serif throughout, our little dino in the bottom, and a few other minor things. I need to get it up on github for our lab anyways so i'll upload later today.
@Ronan_Denyer as it turns out, I was in Houston from 2014-2020 and spent a lot of time at astros games… trust me, it wasn’t great news to find out that all that went down. but as it relates to the paper, despite being a fan, I did analyze the data from those games..
@Ronan_Denyer you should check out the data that @adams_at compiled for those games. super cool effort deciding the audio from those games https://t.co/T7dbiZjPkA
@Ronan_Denyer unfortunately the data weren’t consistent enough to show the effects we expected. some other folks analyzed the data and found that quite a few of the bangs were wrong. so basically it’s a lot of false positives/negatives mixed in with the correct bangs/no bang.
@IntuitMachine@KordingLab@nidhi_s91@TBBake have played with YouTube data a bit more. I actually started this project with the hopes of using pose estimation on MLB youtube videos. It was more of a CV project before it was a movement project so we switched directions for that reason.
What do you get when you cross professional baseball ⚾️with Bayes ? Bayesball! I am super excited to announce a new paper out with @KordingLab, titled "Bayesball: Bayesian Integration in Professional Baseball Batters" https://t.co/w6JmjlP7d7
@pepe_uh@KordingLab Thanks, Pepe! There is definitely the issue that the pitcher's pitch selection is biased by the batters past behavior, so they are definitely not independent. Pitchers certainly try to exploit weaknesses. Situational batting and game strategy are missing from the current analysis
@tdverstynen @KordingLab @GunnarBlohm Yes! that paper is awesome. it’s really cool to see how they combined two seemingly disparate data sources in such a great way
@OliverRunswick @KordingLab@David_J_Harris Thanks! yes that would be great! feels like it is an area of serious untapped potential… at least that is what i’m banking on! 😂
@SandeepKishor13 I don’t know if the data are available for analysis, but I do think that conceptually this has to be some aspect of the playing strategy. same with tennis. I agree that racquetball is probably nearly all observation but i don’t know much about it other than playing occasionally
@richardfbetzel I think it could give us an indication whether batters tend towards being biased away/towards the prior vs maybe those that swing freely based on what they see. Makes me think of Altuve in the 18-inning marathon striking out on 3 consecutive sliders. I'm sure he had the scouting