Stolen Base Model

MLB stolen base predictions for today

Today's slate ranked by stolen base likelihood, built for SB props and for the DFS bonus.

What goes into a steal projection

Four things decide a stolen base, and only one of them is the runner.

Runner: sprint speed and, more predictive than speed, how often he actually attempts. Plenty of fast players do not run, and attempt rate is the input that separates a threat from a fast guy who stays put.

Pitcher: time to the plate. A starter who takes 1.5 seconds from first movement to the catcher's glove gives up bases that a 1.25 second delivery does not, regardless of who is behind the plate. Pickoff frequency and whether he is left-handed both matter here too.

Catcher: pop time and arm strength. The gap between the best and worst catcher arms in the league is worth a meaningful share of attempts on its own.

Game state: score, inning and base situation. Teams do not run down six in the eighth, and they run more with a base open in a close game.

Props versus DFS

For a stolen base prop the model output is a probability you can compare directly to the book's price, and the market is thinner here than it is on home runs, so mispricing is more common. For DFS the steal is a bonus on top of the base projection, which means a runner is worth targeting when his hitting projection already stands up. Chasing steals from a bat you would not otherwise roster is a losing trade.

Frequently asked questions