November 22, 2011 at 10:52 am by Ben Duronio under Statistical Analysis
Mike Fast of Baseball Prospectus released a report on batted balls, specifically on the speed of balls leaving the bat and how likely each a ball is to land in for a hit.
It’s fascinating research, and as a saber-slanted blog we highly recommend reading it.
Fast does a good job of pointing out that the problems many of us have considered with FIP are real. Pitchers have more than no control over batted balls, so pitchers with bigger ground ball or pop-up splits need to be looked at differently than those with standard batted ball splits.
This does not mean FIP or similar defensive independent pitching stats are useless. Most pitchers don’t have those splits, and the analysis we often use to point to a positive or negative expected regression is often still correct. Even in cases such as Tim Hudson’s 2010, FIP can still be used to determine an expected regression by looking at the spread between ERA and FIP throughout his career.
No stat is perfect, and using a combination of the components of defensive independent pitching stats along with analyzing batted ball rates is the best method to accomplish what FIP attempts to do. Using FIP to glance over pitchers quickly is still a reasonable tool if you understand the pitcher being studied has relatively standard batted ball rates.