Saturday, April 16, 2016

In part 1 of this series, I examined how hitters performed measured by weighed on-base average (wOBA) in relation to player age. Now I’m turning attention to the pitchers. What I discovered while evaluating how hitters performed as they aged was that they generally struggled initially at age 20 and make improvements up through age 26, where they largely plateau. After a peak at age 33, they begin a decline around age 35-36. My hypothesis was that pitchers would, while not exactly matching this curve, end up somewhere in line with the hitters, perhaps peaking and declining earlier. What I found from my assessment did not support this at all.

The initial dilemma I faced as I attempted to pool the available data was choosing the metric to evaluate. I decided to use simple earned run average (ERA) and to include an innings pit19ched floor of 60. At first, I considered looking exclusively at starting pitchers, but then felt that ignoring the value of relievers who pitched sufficient innings within a season would be a mistake, so I set the innings pitched limit to help identify these types of relievers.

ERA has its drawbacks. The results can be affected by defense and park factors, but in general, I believe we can accept the averaged results of each age group as correctly reporting the true talent level of the group. I include a graph plotting pitcher ERA as a function of age and a graph plotting the population size for each age group. Obviously, age groups with fewer pitchers results in less certainty for that group. This should be taken into consideration, as both ends of the age spectrum contain fewer pitchers.

To interpret these results, consider that the ERA scale on the vertical axis is arranged lowest to highest. This is important when comparing to the age curve of the hitters in the previous article. That curve showed an almost bell- or hill-shaped curve, peaking towards the middle and declining on either side. While this is also the case below, because ERA goes from lowest to highest, we have a rather surprising result.

To lend these results a little context, I’m going present a graph plotting the number of pitchers evaluated in each age group:

As we see, the extreme ends of the graph are quite low compared to the middle. There are few pitchers that accumulate 60 big league IP until about age 22. The same can be said after age 35. Also consider that the peak years, 25-28, may have more pitchers that were not able to be identified through some of the limitations of the Staslab database (please see the first article for details).

Now take a minute to jump back and assess the results on the ERA graph. Somewhat surprisingly, the younger pitchers between the ages of 20 and 24 have fared better than the bulk of pitchers aged 25-31. This is opposite of what was found for hitters of the same ages. But even more surprising are the results seen with pitchers beyond the age of 31, who put up average ERAs 4.00.

Conclusions

I’m hesitant to claim that Father Time has no effect on PEBA pitchers, as the second graph certainly shows a decline in the number of active pitchers with at least 60 IP in a season. The gap between age 31 and 32 is rather sharp and declines gradually from that point further to under 50 by age 36. This leads one to believe that as pitchers age, they will generally find their way out of the league, but the veterans that remained have performed well… quite well, to the point of performing better than the average pitcher. This brings into question the idea that age will necessarily impact the performance of a given pitcher. It seems a pitcher can have a very good season at advanced ages.