11 March 2011

Brian Roberts and the Aging of Second Basemen

You can follow us on Twitter @CamdenDepot

Concern has been mounting over Brian Roberts' back as he has been held out of several practices.  He has downplayed the severity of this injury.  This typically would not be of much concern were it not for two issues:
  1. Historically, second basemen fall apart in their early thirties.
  2. Last year, Roberts was injured during Spring Training and similarly played down concerns over an injury that wound up shutting him down for the majority of the season.
In this post, I would like to establish what could be expected from a population of second basemen who performed similarly to Roberts over their Age 27 to 31 seasons and how that population performed from Age 32 to 36.  I took this population from players whose Age 27 to 31 seasons occurred from 1950 to 2005 and who generated a WAR of 12 to 22.  This group includes Bobby Grich, Davey Johnson, Robby Thompson, Chuck Knoblauch, Lou Whitaker, Dick McAuliffe, Ray Durham, Johnny Ray, Bill Doran, Willie Randolph, Ron Hunt, Damion Easley, Tom Herr, and Johnny Temple.  As a group they had an OPS+ of 109 +/- 8, a WAR of 16.6 +/- 3.3, an OBP of .360, and a SLG of .401.  Brian Roberts, during his age 27 to 31 years, had an OPS of 115, a WAR of 17.9, an OBP of .369, and a SLG of .451. 

After the jump, a run through of graphs showing what can be expected in terms of plate appearances, WAR, chance of being an average player, and chance of remaining in the Majors.

The following graph shows the average number of plate appearances for each age class, the median, and the 25th/75th percentiles:
As you can see there is a wide range among this group of fourteen players.  However, as a population you can see there is a steady decrease in plate appearances for this group of second basemen.  By the final year (age 36; one year after Brian Roberts contract has ended), the average plate appearances is primarily sustained by Lou Whittaker, Bobby Grich, Ray Durham, and Willie Randolph.  Basically, everyone else within this population has dropped out.

The next graph, goes beyond looking at playing time and focuses on performance.  Instead of plate appearances, Wins Above Replacement (WAR) is used to describe the population:
I would consider an average WAR to be 2.  The average WAR just makes the mark, but the median falls well short of it.  Brian Roberts had a WAR of 1.1 last year, which is in the lower half of this group.  By 33, the median performance is below 1 and continues through each age class.  Even more sobering is that the 75th percentile is about 1 for the Age 35 and 36 age classes.  This graph here makes one wonder why anyone would give a second basemen in his 30s anything beyond a two year deal.  The Orioles gave Roberts a four year extension arguably above market rate.  It never made sense (as I mentioned when the extension took place in 2009).

This final graph describes this population of second basemen by showing what percentage of them were above average and how many of them remained in the majors.
What might surprise some folks is that the first year of Roberts extension saw fewer than half of players like him, WAR-wise, being average in performance.  The graph above illustrates just how difficult it is for successful, above-average second basemen to maintain their performance into their 30s.  The red bars illustrate how it may appear likely that Brian Roberts will not sign another contract.  That is a sobering thought.

I ran through several regression trying to determine if I could predict Roberts collapse, but there does not seem to be a shared trait or set of trait that dictate an evaporation of ability within this subset of second basemen.  The only trait that seemed to be unique to second basemen who were successful deep into their thirties was having an above average defensive WAR as Baseball Reference calculates it.  3 of 7 players who had a positive WAR during their Age 27 to 31 seasons had a WAR above 10.  In the group who registered below average defense, only Ray Durham retained his success.  Brian Roberts, if included in this group, had the worse defense WAR with -22.  This is an awful small sample size, but it suggests not a good future for Roberts.

This really should not have been a surprise.


The Oriole Way said...

Excellent work (sorry you had to do it twice).

Spitballing here, but I think the point made regarding above average defense is a key one. The position a player plays is not random, and the population of second basemen is made up largely of players with two characteristics: 1) defensively, these players were not strong enough to play shortstop and 2) offensively, these players may not have been strong enough to play a corner position (which I think is a reasonable assumption for the good but not superstar level of player in this sample). Looked at in this light, it seems entirely reasonable that they would exhibit early aging patterns; their offensive skills would not support a move down the defensive spectrum, and their defense has already been selected against once in their career. Thus, for a player to remain useful for a long time, he would need to start aging from an above average level of defense.

Jon Shepherd said...

I think that is pretty much spot on.

domenico said...

Could you please explain why Brian never attempted surgery of his herniated disc? To an uninforned like myself, it is difficult to comprehend,

Jon Shepherd said...

I cannot speak directly for Brian Roberts, but the issues I would imagine he might have for surgery on his back would be:
1. Back surgery is typically a last alternative after other types of rehabilitation have failed because as surgeries go . . . back surgery does not have all that a remarkable success rate. You see similar aversions to surgery with things like torn labrums. The surgical fix is often not exceptionally better with respect to athletic competition.
2. It takes a while to rehab and Roberts wanted to play last year and he wanted to play this year.