clock menu more-arrow no yes

Filed under:

Aging and the Astros: what do we know?

New, 13 comments


Spurred by the wave of big multi-year contracts at the winter meetings, the impact of aging on player performance  became a hot topic for sabermetric discussion.  A couple of articles by Tango, at the Book blog, and the quotes from Red Sox consultant and statistical guru Bill James, led me to wonder how this discussion affects a few Astros players.  Later in this article I will discuss the implications for Hunter Pence, Michael Bourn, and Carlos Lee, but first let's review what the data shows.

Tango, at the Book blog, used a data base of major league hitters (born between1895 and 1985)  to examine the relationship between aging and declines in Wins Above Replacement (WAR) .  Between ages 26 and 34, the average WAR by year: 4.4, 3.9, 3.4, 2.9, 2.4.  Keep in mind that this includes defense as well as offense. Between the ages of 26 and 34, the great hitter group declines exactly 0.5 wins per season, which Tango says he uses as a "rule of thumb" for age effects.

After the Red Sox entered into a 7 year/142 million contract with Carl Crawford, Tango performed another analysis, based on a data base of great fast offensive players, to examine the effect of speed on WAR decline.    His findings: speedy hitters age better than overall great hitters.  The effect is small--speed results in .05 lower decline in wins--but it adds up over the course of a long contract.  Based on average age decline for great hitters, Tango estimates that Crawford justifies a 7 year/114 million contract.  However, by including the aging advantage for speedsters, Tango concludes that Crawford is worth $28 million more, leading him to change his initial impression that the Red Sox had overpaid. 

The view that speedsters age better is contrary to the typical comments we see on baseball blogs.  I often see comments about speedsters (like Michael Bourn) which suggest that they will have short careers because "speed is the first thing to go."  While I often see that common wisdom, I haven't seen any data support for it.  And Tango's analysis may put that prediction in the urban myth category.  Why do speedsters age better?  The comments at the Book blog provide some interesting theories  Tango sees it as the same reason that a 95 mph pitcher will age better than a 88 mph pitcher--the higher velocity pitcher can sustain a greater decline and still fall within a minimum standard of performance.  Some view "speedster" as a proxy for better athletes, whose conditioning and athleticism allows the player to access a greater array of skills as they age.  Another possibility (and one I lean toward accepting) is that speedy players will continue to have more defensive flexibility as they age, while non-speedsters are forced into narrow defensive roles as they age.

This latter point is related to Bill James' explanation of the advantages of speed on the aging curve.  In the interview with ESPN, James declined to divulge the details of the Crawford analysis he performed for the Red Sox, but he provided his general views on aging, based on his previously published work.  Hitters with sub-par speed don't age well because their growing defensive limitations force them into first basemen:

THE thing that drives them out of the game is not the loss in hitting ability in absolute terms. There are dozens of 37-year-old first basemen who could still hit enough to play -- if they could play the outfield. When their speed drops below a certain level, they're no longer able to play the outfield at a decent level, no longer able to hit enough to be a cleanup hitter, and they're gone.

In James' view, speedsters who also have another good offensive skill (like OBP or SLG) are capable of staying in the game a long time because they are not limited to playing the least demanding defensive positions.  One of the several examples he provides is Craig Biggio, who had enough offensive skill to become a 20 HR hitter at age 40. Biggio is an obvious example of a speed player who continued to have enough versatility to shift between 2d base, LF, and CF in his later years.  James thinks that Crawford's power is good enough to put him in the same kind of category.

Next, let's talk about Pence, Bourn, and Lee.

We know that Michael Bourn is one of the fastest players in the game.  (Crawford, Bourn, and Bourgeois were in the same Little League outfield? wow.)  As a general matter, the results of Tango's study, indicating that speedy players age better than other players, is good news for Bourn and the Astros.  It's not hard to imagine that Bourn will be able to play center field at an above average level for many years.  I also think that his stolen base capability will stay at a high level.  Stealing bases is a skill that can be improved with experience.  As players like Biggio and Kenny Lofton have shown, the aging speed player can compensate for declining speed with greater understanding of getting jumps and knowing how to read pitchers.

The question is whether Bourn's remaining hitting skills are sufficient to allow him to be a starter, rather than a 4th outfielder, as he ages.  James suggests that another definable offensive skill is necessary to bolster the speedster as he ages.  This is similar to a previous comment I made about Bourn's career path: the length of his career may depend on how much he can improve his walk rates.  Bourn's current walk rate is decent, but not outstanding.  A good comparison is Dave Roberts (a similar lefthanded speedy outfielder), who increased his walk rates at age 30, and became a valuable lead off hitter through age 36.  Roberts retired with a .342 career OBP, but showed .360-ish OBP at ages 33 and 34.  Roberts had a career 9.9 % BB rate; Bourn had BB rates of 9.9% and 9.3% in 2009 and 2010.  Roberts' peak seasons occurred at age 32-35, when he pushed his average BB rate above 10%.  If I'm the Astros' new hitting coach, I would try to keep Bourn focused on patience and plate discipline.

The speed issue may bode well for Hunter Pence also.  Pence doesn't have the blazing speed of Bourn, but he is plenty fast and he profiles as a good defender in corner outfield positions.  Bill James included Larry Walker among his speed players, and Pence is comparable to him, in terms of late 20's speed. (Walker shows up as the 10th most similar player to Pence at age 27, according to Baseball-Reference.com.)  Furthermore, unlike Bourn, Pence has slugging ability.  Because Pence is not likely to be a fielding liability in corner outfield positions for several years, he could enjoy a reasonably good career into his 30's, so long as he can maintain a decent slugging ability.

The aren't just abstract questions for the Astros.  The Astros may have to decide over the next two years, whether to sign Bourn and Pence to multi-year contracts--or allow them to leave in free agency.

Carlos Lee is probably a contrasting example, since he is the classic non-speed player.  Lee has shown an up and down pattern of WAR from year to year.  Although he hit rock bottom last year, his decline had not been as constant as the averages in Tango's study.  I have normalized Lee's WAR per 650 plate appearances below:

Age 28 (3 year average) : 3.79 WAR

Age 29:  2.17 WAR

Age 30:  2.6 WAR

Age 31:  3.36 WAR

Age 32:  5.27 WAR

Age 33: 2.06 WAR

Age 34: -0.8 WAR

Up until last year, some of Carlos Lee's best years came in an Astros' uniform; and, somewhat unpredictability, he had his peak season at age 32.  My second observation is, "what were the Astros thinking when they signed him?"  Sure, the WAR concept wasn't around (at least in that exact form) in 2007.  But, when he was signed after his age 30 season, he had shown himself to be a 2 to 3 WAR outfielder.  That's not a bad player, but why would you offer him $105 million?

Lee's peak season in 2008 wasn't due only to his offense.  According to UZR (which is used in the WAR calculation), that was one of the few seasons in Lee's career when he was not a below average defensive player.  To a significant extent, Lee's rather abrupt decline in WAR in 2009 and 2010 is due to declining defense in LF.  Carlos Lee's WAR was 1 win below the previous season in 2009 and 2010 solely due to the year-over-year decline in defense. Lee would have a respectable positive WAR last year, if he was as good an outfielder as he was in 2008.  In that sense, Lee is an example of James' observation that declining defense is what forces good hitters into increasingly confined roles.

Based on Lee's average WAR at age 28, Tango's rule of thumb for age-related decline produces a predicted  WAR just below 1 win in 2010, which is higher than the actual -0.8 WAR for 2010.  However, because Lee has not shown a constant pattern, year to year, the application of the rule of thumb is sensitive to the starting point.  For example, if the 0.5 annual decline is applied to 2008, the projected 2011 WAR is 3.7.  I picked the starting point which produces the highest result to illustrate that Lee's uneven pattern of decline makes projections problematic.

Given that the Astros are considering Lee as a first baseman, he follows James' description of the hitter whose lack of speed and defense pushes him to the least demanding position as he ages.  The somewhat sporadic pattern of Lee's WAR over his age 28 - 34 seasons encourages me to believe that he still has enough hitting ability to post a positive WAR.  If he is shifted to first base, he most likely returns to his "typical" WAR in the low to mid-2's, and could improve even more with an offensive rebound.

  Bill James' projection system (sometimes viewed as optimistic) predicts a 2011 OPS of .804 and a wOBA of .349 for Carlos Lee.  If Lee can achieve the James' offensive projection and produce average defense at 1st base, he would post a 3.4 WAR.  If Lee's offense is just somewhat above last year's performance, he could still provide 2 WAR at first base. 

The condition that Lee provide "average defense" at first base is a big "if."  People who watched Lee at first base last year had varying reactions to Lee's defense.  The reactions ranged from terrible to mediocre to great.  The sample size for Lee's defense at first base is too small to reach conclusions either visually or through defensive metrics.  Setting aside the problems with a miniscule sample size, both DRS and UZR came up with positive defensive results for Lee at 1st base last year.  Considering that he has a full spring training to refine his defense at first base, I don't think it's out of the question to assume that he can provide average defense.

I don't intend this discussion of Lee at first base as an argument for benching Brett Wallace or sending him to AAA. That's a completely different discussion for another day.  The point is to explore the ramifications of Lee's declining LF defense as he reaches his mid-30's..