Using Machine Learning to Predict Jon Lester’s Whiff Rate

Jon Lester pitched his butt off and finished 2018 with a stellar 3.32 ERA in 180 innings. But although Lester prevented the same amount of runs as many of the game’s elite starters, the big lefty’s whiff rate was his lowest as a Cub. Why?

The answer is multifaceted and cannot be reduced to a single reason, but Lester’s curveball is one explanation. You have to flip the calendar back to 2013 to find Lester’s whiff rate on the curve as low as it was in 2018.

In order to find out why Lester’s curveball wasn’t inducing whiffs at the same rate it had over the last few years, I used a method called decision tree classification. Specifically, I considered every single Statcast metric and told my computer to find the best predictors of Lester’s curveball whiff rate.

What did the computer say are the greatest predictors of whiffs? The greatest factor for Lester’s curve was vertical release point (release_pos_z), followed by vertical movement (pfx_z). Surprisingly, spin rate wasn’t as important as other numbers. The algorithm* …Read the Rest

Source:: Cubs Insider

You may also like...