Aging and the MLB Life Cycle

Fun fact: We all get older and every one of us will peak at what we do. Little by little, year by year, we start to suck a little more at what we do as our strength, dexterity and mental fitness degrade once we get over that hump. We have no idea how to statistically analyze the time when a pretend sports writer starts to decline, so this won’t be an introspective piece. However, we will at least take a crack at it with baseball players.

When we decided to have a look at how aging impacts major leaguers, we assumed we wouldn’t be the first kids in the pool. Not surprisingly, Bill James was the among the first to dive in. Prior to his “Looking For The Time,” the presumably anecdotal suggestion was that the peak came somewhere around 30 with the prime years considered between 28-32. James blew that notion out of the water suggesting the peak was in fact 27 using something called VAM, short for Value Approximation Method. More recently on Baseball Prospectus, JC Bradbury decided to take a different look with a different methodology which not surprisingly yielded different results, these more in line with the traditional model. With a Google search, you can find a number of rebuttals and hem-hawing that becomes increasingly complex and less readable to the average human being.

It is worth noting that we’re not looking to write the gospel on when a player peaks in the post steroids era. Rather, we wanted to create a rough estimate that allows for application to other problems and questions folks might have like the value of 1 WAR and what type of contract a player has earned. Our methodology is pursposefully simplistic. It was our opinion that if you want to look at how players age over the course of a normal career, we needed to isolate players who have aged in the majors.

We developed our sample from players who reached the age of 35 in any season between 2007-2014, separating the fielders from pitchers. From there, we mapped their WAR year by year in that player’s entire career and took the average. As we’ve already done other work using WAR to determine value and all around performance, we used it as the basis for this piece as well. We also found it to be the most complete statistic in determining all around performance. To remove those playing God and messing with their bodies’ natural processes, we omitted the proven or self-admitted performance enhancers as their post-thirty upticks would clearly have distorted the true aging process. I’m sure there were a few juicers who got past us and made it into the sample, but I’d rather not get too much into the speculation game and finger pointing that makes the old school look like witch hunters right now.

We looked at 140 position players with enough data points to track development from the ages of 22 to 39. In terms of WAR, field players in our sample peaked at the age of 29 with a sweeping curve that rose to that point and faded off on the other side. Bats enjoyed an five year plateau from about the age of 26 until 30 where they hovered around the two win line. The decline gets steep from 32 to 35 and again after the age-36 year. Between age-32 and age-33, the percent decline in WAR for a player was greatest at a 20.6% drop from 1.65 WAR to 1.31 WAR.

Pitchers weren’t nearly as tidy. Collectively, those in the sample (114 pitchers, both starters and relievers) had their best performance at 28, but their performance also spiked upward at 24, 31 and then was a bumpy road after 35. Starters had two sharp negative inflection points — from age-31 to age-32 they collectively experienced a 20.3% drop in WAR then from age-34 to age-35 a steep decline of 30.2% in year-to-year WAR.

Broken down further, relievers had a tendency to peak a little later with an almost indiscernible difference between 29 and 30 while starters still top out at 28. Unlike position players, pitchers seem to go out on a high note with a slight uptick during their 38 year. Of course, a pitchers’ life span in our sample was shorter than the position players, shaving a year off each end.

It’s worth noting that players don’t follow these trajectories naturally. Plenty of players showed decline in years where the rest of the sample improved, while players improved when the sample declined. Different players peaked at different years. However, I feel like you could use this as a rough means of projecting how a player’s career could play out. Application from this study will be much cleaner for the position players than it will be for pitchers, though that’s not to say that it can’t work for both. For example, it will be difficult to assume every single single pitcher will improve between their age-30 and age-31 season. We’ll work out the kinks over time and really refine our approach, but expect other pieces built around this one. The samples need to get bigger, especially for the pitchers and we’ll continue to update our research every year. We’ll also try to eliminate that 2007 season with time.

One thing that we’ve already used the system to determine is how much a win cost in the 2015 free agency season. Looking at contracts awarded to free agent fielders on the open market and projecting those players along a normal development curve, future wins should cost approximately six million dollars per win.

It will interesting to see this concept in action with the Pirates in 2015 in the form of A.J. Burnett. In the study, pitcher WAR actually increased from age-37 to age-38 (Burnett’s 2015 age) from 0.8 to 1.5 WAR, albeit with only 5 players still as starters in our study.

There are definitely flaws with our methods and we understand that. Again, we’re not seeking anything absolute, because in the end we can’t achieve that anyway. As was the case with James and Bradbury, there are a number of ways to look at aging and this is just another one.