Digging Under Projections

When it comes to projections, I am and have been fairly confident and content with the offensive variety. Measuring offense is by no means a solved problem, but we have very good approximations and given the sample sizes and independent nature of offense, not only can we measure it well, but it stabilizes and projects well. Honestly, I think the most exciting front in offensive analysis is further enhancement of park factors, which actually is a pervasive effect and not limited solely to hitters.

Dave Allen touched on my crusade on the pitching front already. I am a big proponent of looking at pitchers on a per-pitch level. In addition to studying pitch types and their subsequent metrics (e.g. velocity), I advocate studying the result types of those pitches. A pitcher’s strikeout rate shows a strong correlation with his swinging strike rate, and little to no correlation with his called strike rate. Correlation does not imply causation, but swinging strike rate is one of my favorite one-stop metrics to glance at when looking for either breakout or regression candidates. When the strikeouts and missed bat rates are out of whack, I look for a change in the coming year.

Offense and pitching are not the only parts of projections, though. Here at FanGraphs you are given the chance to project defensive adequacy as well. That is a tricky issue. Defensive data is among the noisiest of all facets of the game. The standard rule of thumb is that three years of defensive data is needed in order to equal the confidence level of a single year of hitting data. Naturally then, a three-year weighted average of UZR data appears to be the best way to project next year’s figures, but beware of that simplicity.

The issue is that three years is a long time and it is important to note that how a player fields at age 25 is going to be different, perhaps vastly so, than at age 28, for one example. When you get into later seasons, it can be even more treacherous. Aging curve studies in fielding done so far seem to point to fielding being a skill that peaks early in a player’s career. When projecting defense, take care to remember that one year is a poor sample and that next year is probably going to be worse than the years prior.



Print This Post



Matthew Carruth is a software engineer who has been fascinated with baseball statistics since age five. When not dissecting baseball, he is watching hockey or playing soccer.


Comments Are Loading Now!