Read on Twitter
Mitch Keller, pitcher evaluation, and the danger of confounding small edges, a thread:
You’re evaluating Mitch Keller. You look at his projections, and they range from the mid fours to the high fours with a below average K-BB: bad. But pitcher projections aren’t as accurate as hitter projections! So you march on.
Mitch Keller mirrors his fastball and curveball perfectly. The spins on each pitch go in exactly opposite directions, which has some value. Good!
How much though?
You take a look at his pitch movements, though, and he doesn’t have all that much ride on the fastball, no wiggle, changeup’s no good. Bad!
You take a look at his pitch movements, though, and he doesn’t have all that much ride on the fastball, no wiggle, changeup’s no good. Bad!
Take a look at his stuff number, which captures much of the spin-mirroring, and the bad fastball, and says: Keller is good, top 25 Stuff number among starters. Good!
But you look back at the walk totals, and the homer totals in the high minors and majors, and uh-oh: command issues. Bottom ten Command+ (min 5 projected starts) for Mitch Keller. Bad!
How do you put it all together? You need to evaluate each component & put them into a projection system, and I’m helping/willing to help anyone do that. In meantime, I intuitively sort through relative effects of each ‘edge,' & produce ranks. Back end top 75 for Keller.
Here’s a better overlay of perfect spin mirrorer Mitch Keller throwing his four-seam and curve. More on spin mirroring here https://theathletic.com/2323195/2021/01/15/spin-mirroring-identifying-fantasy-baseball-sleepers/
