Opinion: How Can We Evaluate Base Stealing More Precisely?

Photo Credit: John Fisher/Getty Images

Consider this scenario: Two players each have 30 stolen bases and 5 caught stealing. With an 86% success rate, both players appear to be helping their teams significantly by risking taking the extra base. However, 35 attempts is quite a small sample size. The standard deviation of the number of stolen bases in this case is 2.07. So, there’s a 2.5% chance that the base-stealer’s true success rate is actually lower than 74%, in which case the advice to that player would be to steal fewer bases. Essentially, for many players, it can be hard to tell If they should steal more or less, purely based on their stolen bases and caught stealing, since the sample sizes are small.

What if we had a metric that would give us even more confidence in telling players to steal more frequently?

A more accurate way to measure base stealing ability could be the use of xSB, which represents the probability that a player should have been safe in a stolen base attempt. The xSB calculation may consider how much the base-stealer was safe or out by. For instance, a player who gets excellent jumps that often deter catchers from attempting a throw altogether will have a higher xSB than one who has narrowly been safe on several occasions. Additionally, a base-stealer who was out by mere inches in all of his caught stealing attempts will have a higher xSB than one who was significantly out multiple times.