Explanatory Value of Exit Velocity Data

Posted on Feb 20, 2018


Beginning with the 2015 season, Major League Baseball began capturing and publishing Statcast data, which uses state of the art cameras to capture and quantify previously unquantifiable baseball activity, such as the positioning and range of fielders, and the speed (exit velocity) and angle (launch angle) at which the baseball leaves the hitter's bat. As there are now three seasons of available Statcast data, I set out to explore what explanatory power exit velocity has with respect to established measures of power and run production.

Data Capture

I used Python's Scrapy library to marry Fangraph's Advanced Batter Statistics and Batted Ball data to the Statcast Exit Velocity data for all players with more than 30 at bats for the 2015, 2016, and 2017 seasons. I filtered the exit velocity data for line drives and fly balls; popups and ground balls theoretically should have little to no bearing on power production. Statcast data is extremely granular, individual plays are available. For this exercise however, I only looked at season averages as a jumping off point. I made this choice largely because Statcast frowns upon scraping, and makes the most granular data available primarily via CSV export. This analysis was undertaken under the guise of our Web Scraping project, exporting CSV data would somewhat defeat the purpose. I intend to incorporate more granular data in future analyses. For this analysis I was able to retrieve a sample of 310 player seasons using a recursive Scrapy spider.


My initial hypothesis was that Exit Velocity on line drives and fly balls would be a statistically significant predictor of a batter's power production. (The corresponding null hypothesis was that there would be no statistical significance to Exit Velocity as an explanatory variable). I ran linear regressions of ISO (isolated power), and Home Run to Fly Ball Ratio against Exit Velocity.

ISO is calculated as SLG (Slugging percentage: total bases/plate appearances) - OBP (On Base Percentage: times on base/plate appearances); ISO attempts to measure power production by eliminating the reaching base component from offensive output, you can think of it as extra bases per plate appearance. HR/FB ratio is the percentage of balls hit in the air that become Home Runs.

The rationale here is that logically, the faster the ball travels, the more likely it is to leave the park or land for an extra base hit. Velocity is not the sole determinant of outcome, angle, direction relative to fielder positioning, dimensions of the park, wind speed and direction are some of the other variables expected to influence offensive outcomes. Therefore the expectation would be statistical significance but not an overly large coefficient of determination, as these other variables are not accounted for.

As anticipated, the simple linear regressions produced large t-values (19.6 and 16.1) indicating statistical significance with greater than 99.9% confidence, but moderate coefficients of determination (0.555 and 0.456) for HR/FB and ISO respectively. These results disprove the null hypothesis and support the expectations.

About Author

Related Articles

Leave a Comment

No comments found.

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python Python Data Analysis python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp