Premium and Super Premium Subscribers Get a 20% Discount at MLB.tv!
September 22, 2004
Can Of Corn
Rethinking the Defensive Spectrum
Last time, in a column about B.J. Upton's possible switch to third base, I observed that the offensive productivity gap between the average shortstop and the average third baseman was surprisingly narrow over the last five years. More than one reader raised a sensible objection to the numbers. The data sample of five years might be small enough to skew the numbers. That's a fair concern, and one I'd like to address this week.
With the help of James Click and his SQL chops, I'll examine the run gaps among all positions, and this time I'll use the entire depth and breadth of the play-by-play data, which reaches back to 1972.
Here are the average rate stats for each position from 1972 through the present:
Position AVG/OBP/SLG C .254/.319/.387 1B .277/.354/.451 2B .267/.329/.375 3B .267/.334/.419 SS .259/.311/.361 LF .278/.348/.440 CF .272/.334/.411 RF .275/.344/.447 DH .269/.345/.442
As expected, we notice a significant dip in shortstop productivity when the study is expanded to 30-plus years. To put run values to the rates, just like last time, I'll use this quick-and-dirty formula: PA*OBP*SLG*(1-OBP)/(1-AVG), with plate appearances set to 650 to simulate a full season of production for an individual player. Here's what happens:
Position Runs/650 PA C 73.3 1B 92.7 2B 73.4 3B 82.7 SS 67.9 LF 89.9 CF 81.6 RF 90.4 DH 88.8
These data etch quite a different picture than the five-year numbers used in my previous article. The shortstop position, as we would expect, is the least productive over the last 33 seasons. While the run gap between shortstop and third base since 2000 is a piddling 1.4 runs, historically it's been more than 10 times that amount. Framed another way, the gap in recent seasons has been roughly one-and-a-half runs; over the last 30-plus years, it's been roughly one-and-a-half wins. We call that a chasm.
Here's how a Bill James-inspired "productivity spectrum" would look, increasing in production from left to right for the 1972-2004 period:
Now let's see how the run values for each position compare from 2000-2004, with percent increases over 1972-2004 levels:
Position Runs/650 PA % Increase C 75.5 3.0 1B 99.6 7.4 2B 76.4 4.1 3B 84.6 2.3 SS 83.2 22.5 LF 91.1 1.3 CF 85.3 4.5 RF 90.3 -0.1 DH 89.8 1.1
Shortstops have seen a dramatic increase in productivity over the last five seasons, which isn't all that surprising considering the recent exploits of Messrs Rodriguez, Garciaparra, Jeter and Tejada. In contrast, third basemen have made very modest gains over the same period. Additionally, first basemen have widened their lead over other positions in terms of run output, which is even more damning for those organizations willing to countenance sub-optimal production from first in recent seasons.
Here then is the adjusted productivity spectrum, reflecting the last five seasons:
Shortstops have advanced two positions, and center field and third base have flip-flopped, as have left field and right field. Catcher and second base have slid a notch, while first and DH have held serve.
It's impossible to say whether the past five seasons, in terms of shortstop production, are patently anomalous or indicative of a genuine emergent trend. While the former is the most likely explanation, these new data could, as I intimated last week, be the result of advanced training and nutrition and generally bigger-bodied shortstops reducing the offensive variance relative to other positions.
As for how all this relates to Upton, the contemporary production gap between third and short is indeed quite narrow. However, it's probably unwise to justify Upton's position switch--as I did last week--based on these recent numbers. History suggests we're in a deeply aberrant period of shortstop production. If that's the case, the Rays will be squandering quite a bit of Upton's value by moving him to third. If we are in fact seeing a new standard at the position, they'll probably experience a net gain from the move. At this point, it's guesswork as to which scenario is the real one. Educated guesswork, but guesswork just the same.