Monday, December 24, 2012


OK, let's get down to cases with the age-old question of career progression for young hitters and how more definitive patterns can be discerned from actual data. (Here we follow up on our more flamboyant previous post, where Joe P., re-stating the obvious, indulges in the type of corn-pone exposition that's simultaneously cynical and homespun.)

So here we will try to determine if young player talent levels, as measured by their OPS+ in their very earliest seasons, actually are highly preordained. By this, we mean that young players may already shake out into groups that have readily predictable categorical limits that come into play right from the start.

Bill James suggested a variant of this in the primordial days of his Baseball Abstract when he noted that the younger a player was when he excelled at the game, the longer his career was likely to be. Aside from historical aberrations where some exceptionally young (meaning teen-age) player were brought to the majors (primarily during WWII, with a bit of blip in the sixties), the pivot point for hitters in terms of defining their future career path seems to be located at age 21.

If you make it to the majors and hold your own at that point, you have a strong chance for a long, productive career.

But the level that you achieve at that age--as the chart at left strongly indicates--tends to define a series of segmented, parallel "streams of success" that for the most part remain separated rather strictly according to just how much the hitter dominates against the league at a very early age.

This long, long table displays more than a hundred hitters from 1901 to the present, sorted in descending order of their OPS+ during their age-21 season. What the viewer should be attempting to tease out of this is the frequency in which players in the five OPS+ ranges displayed here (150 OPS+ or higher at age 21, 135-149, 120-134, 110-119, and 109 and below) actually exceed the range they initially set for themselves at that very early point in their careers

(Currently active players are shown in red type.)

What will become evident from studying the players at length (and by examining the summary data for the five ranges that's displayed below the main chart) is that the percentage of improvement is at its greatest among the young players who excel the most, and it declines in a linear fashion across each gradation of quality.

While there are clearly some exceptions (and you're instructed to look across right now at the numbers for Roberto Clemente to see one of the more dramatic anomalies in baseball history), players tend to remain rather tightly slotted across their careers to the level of production they achieve at age 21.

Full disclosure: we've left off about 40 additional players in the lowest grouping, all of these being the players with the worst performance at the age of 21. You see the high preponderance of players whose names appear in bold type at the bottom of the diagram--these are players in the Hall of Fame, and we clustered them in order to see how their patterns looked. As is clear, it's only a few of these players with indifferent performance levels at age 21 who bucked the odds and evolved into Hall of Fame-caliber hitters.

Here are two summaries of the very granular data presented above. First we see how the OPS+ averages progress in each class as we go through several multi-year phases: after age 21, we measure ages 22-24, 25-27, 28-32 and 33-35. We also show the pre-age 21 data when the players have a sufficient number of plate appearances (if they have less than 300 PA in these games, we show that with an asterisk and do not record their OPS+).

The summaries show two things: 1) there are distinct "tracks" that hitters tend to cluster into as a result of their initial level of success, and 2) that the percentage of improvement toward "high peaks" decreases dramatically when the age-21 season is below 120 OPS+.

43% of hitters will show a more pronounced improvement toward a cluster of peak seasons (whether at age 22-24, 25-27, and much less frequently at 28-32) when their age-21 OPS+ is 120 or higher. Only 17% of hitters will show such pronounced improvement when their OPS+ at age 21 is less than 120.

Cesar Cedeno: a false alarm...
So what does this have to do with Joe P. and his ejaculation that young players will perform better as they approach their peak than any group of older players (even "elite" older players? Quite obviously, a good bit. We can see that in adjacent categories from age 21, the 135-149 OPS+ group vs. the 120-134 OPS+ group, that average performance levels for the former remain equal to or a bit higher in the "post-peak" period (ages 28 up) than the levels generated by the latter group in their "pre-peak" (22-24) and "peak" (25-27) periods.

The truly great players (who debut with OPS+ levels of 150 or higher) tend to be the untouchable inner-circle Hall of Famers, sustaining a level of performance even into their later years that isn't approached by any of the other OPS+ ranges. (Of course, there are a few exceptions to that rule: see Cesar Cedeno and Hal Trosky.)

Hal Trosky: felled by injuries
Joe was doubling down on Salvador Perez and Eric Hosmer vs. Josh Hamilton and Albert Pujols. Neither Perez nor Hamilton are part of the study due to the shape of their careers: Perez had less than 200 PAs as a 21-year old, while Hamilton, who was held back for several years due to susbstance abuse issue, didn't make it to the majors until age 26. Hosmer, however, did have a solid age-21 season in 2011 (118 OPS+) before struggling significantly last season.

In a group of hitters where his age-21 season is centered in the population, Hosmer is surrounded by a group of 27 hitters (including himself). OPS+ range: 115-124 for the age 21 year. Out of that group, 7 showed significant improvement. That's just under 26%. While Joe goes with his gut and proclaims that Hosmer will be "a masher," our categorical connection chart indicates that Hosmer has about a 1 in 4 chance of being a breakout hitter.

Now those aren't horrible odds, by any stretch. But unlike hitters such as Giancarlo Stanton or (in semi-recent history) Gary Sheffield, Eric didn't make a step forward during his age-22 season. He's showing difficulty against left-handed pitching. That does tend to make 2013 a pivotal year for him.

Organizing the data in this way actually quantifies the odds of breakout seasons, and takes some of the wish fulfillment away from Joe as he constructs a curious set of prognostications. As we said earlier, there is almost a morbid fascination pervading the cadre of numbers-aligned writers who count the Royals as their home team. That clearly hasn't worked well for them--certainly not in the twenty-first century, which is actually moving well into its teenage years. The response of these writers is analogously adolescent as they reach back for the comfort of Jamesian truisms to salve their wounds of fan identification, when charts such as the one above can assist them in seeing just how likely it is that the scenario they are so devoutly attempting to consummate will actually occur.