The previous post introduced some research we undertook at the behest of my fantasy star Paul Goldschmidt. In short, I kept Goldschmidt from 2012 to 2013 on the hunch that his doubles and home runs (43 and 20, respectively, in 2012) would even out some, toward a ratio closer to 1-to-1 than 2-to-1. (Huzzah, he hit 36 doubles and 36 home runs in 2013.) So I decided to follow that up with a 10-year study of whether doubles (rather, doubles plus home runs, or all extra-base hits) offered more predictive power than home runs. It was for my own fantasy purposes that I began this study, but it ballooned into something enormous thanks to my inability to accept that my hypothesis was wrong. The number of avenues I investigated was excessive, but some of them proved interesting, so I’ve started sharing. Thus ends the recap.
My hypothesis wasn’t wrong per se, just poorly worded. I thought at first that doubles and home runs (2B+HR on most of my graphs) would prove more stable over time than plain home runs, feeling that players generally displayed a consistent amount of power and that it was mostly luck who determined whether that long fly ball went over the fence or bounced just short of it. That was proven wrong; home runs are more consistent year-to-year than 2B+HR, no matter how you slice them.
If I had reworded my original hypothesis to account for aging, I would have gotten a satisfactory answer sooner. As a player enters his prime, and his power develops, home runs should eat a bigger share of his total extra-base hit pie. Here’s one visualization of that idea.
You can see the trends, but on this graph they look minor. Triples decline pretty much from the start, first counting for 10 percent of all XBH and declining with age to about five. Players get slow; that makes sense. As the triples vanish, home runs seem to replace them. However, the gains there are minimal, except at age 39 (see caption above). Doubles maintain their share of about 60 percent of all XBH.
But you are tired of hearing about dead ends. Did I find anything useful?
Doubles per home run. That’s what it came down to at the start. Goldschmidt had a more than two doubles for each home run in 2012, far too many for a player of his talent, at his age. This final graph delineates three distinct plateaus throughout a young player’s development. From ages 20 to 24, the ratio holds steady around 1.9, then it dips to around 1.7 for ages 25 to 28, then dips again to 1.6 for ages 29 and 30. So you may expect the ratio of doubles to homers to even out with age, but not as much as I originally hoped; the average ratio never dips below 1.5-to-1, except for the Bonds-skewed 39 year olds. Of course, these figures could be refined by categorizing batters by type and measuring the ratio for each type. We won’t shut the door on that possibility.
Here’s an appendix of interesting graphs that didn’t fit into the discussion above.
The mean age in this sample was 28.3 years old. The graph shows that the mode for ages is 26. The difference is explained above, in the gradual downward slope. Veterans stick around in baseball. Speed isn’t as important as in other sports, home run power lasts through the early and mid 30s, their production is generally more reliable thanks to all the data of their past seasons, and don’t forget that this window includes the PED era. PED use grinds normal aging to a halt, and sometimes can turn the process around.
Baseball is a kind climate for veterans above 30, but younger players are more profitable in almost every sense. The line representing WAR per 600 PA shows how much more valuable youngsters are than veterans, on average. Younger players are not only more productive, but cheaper, under team control, steadily improving, more handsome and likely less prone to injury. If there was enough young talent to stock major league rosters, we would already have seen that happen. Alas, the dozen or so minor and international leagues that feed into the majors are playing so far below big-league level that Juan Uribe just got two years, $15 million from the Dodgers.
This graph was from early in our process, which is why it only covers the last five years. Still, it demonstrates that the few players who are talented enough to break into the bigs at 20 and 21 are also talented enough to start the majority of games. The more traditional path to regular playing time starts at age 22 on this graph. Thirty-year-olds have the highest average games played, yet their figure is still less than 75 percent of all games in a season, underscoring the ever-present need for positional depth.
Hope you enjoyed this. These kinds of data-driven posts will be occasional features here at Midnight Baseball. We’d love any feedback and quality control you have to offer.