Thursday, April 14, 2011

PEDs and Baseball--The New Guy's Perspective

(Note from the Editor: I'd like to welcome Deron Pope to the site. Deron's been writing on his blog, There's a Stat for That, which caught our eye, and we're really happy to have him aboard. He's got great statistical insights on numerous sports that he'll be bringing to SportStatistics.) 

The graph below acts as a detector for Performance Enhancing Drug (PED) users.  The subsequent graphs show an obvious user in Barry Bonds; somebody who might have been using steroids in Benito Santiago; and a normal career curve for comparison--the great Richie Ashburn. We'll start with the chart showing the top 10% and the bottom 10% for adjusted career OPS+ only:

A great way to detect an abnormal career progression; the graph uses OPS+ as a proxy to measure overall batting skill.  Relative OPS+ is measured by comparing 5-year periods of a player's career.  For instance, when 32 is seen on the age axis it represents the player's performance from age 28 through age 32, and age 33 represents ages 29 through 33 and so on.  The "relative" part is introduced when all of the player's other five-year periods are indexed to the player's best five-year period.  The best 5-year period is set to equal 100 and the rest of the 5-year periods are measured accordingly.  The chart above  displays the career progression in which 80% of players fit.  A couple of things to remember when viewing the chart is that the area between the 10% lines is 80% of all players measured. Additionally, the player's performance is compared to himself, so if Player A has an 85 rating at age 32 and Player B has an 89 rating at the same age, that does not necessarily mean that Player  B was a better player; it just means Player B closer to his peak than Player A. Hit the jump to see the individual player charts.

Barry Bonds
First we'll take a look at the chart of Barry Bonds. Whether or not a federal jury decides that Mr. Bonds may or may not have lied to federal prosecutors, it's fairly obvious to most baseball fans that Barry consumed some sort of performance-enhancing substance during his  career. The peak of his career came after age 38, when most players are well into the decline of their careers--even those players that are still in the top 10% of how well they played relative to their peak years. Instead, those years were the best of his career, which just doesn't happen at those ages. He hit 73 home runs at the ridiculous age of 37, then hit 46, 45, and 45 long balls the next three years to give him the all-time home run record (at least, in some people's books). If you don't believe that Barry was using, I have a nice bridge for sale. 
Benito Santiago

Now we'll look at a more curious case--Benito Santiago. The catcher had quite a long career, spanning three decades and nine different franchises, right in the heart of the Steroid Era (Santiago played from 1986 to 2005).  He hit 217 career home runs and 920 RBIs while hitting .263/.307/.415 over the course of his nineteen seasons in the major leagues. What's curious is about the data in his mid-to-late-thirties. Santiago went from being near the bottom 10% to the top 10% between the ages of 36 and 40. I don't mean to say that Santiago was a bad hitter, but it's certainly not likely that he would suddenly have such a jump in comparative OBP+ unless he was on some sort of drugs.
Richie Ashburn
So finally we can look at Whitey, the Phillies great from the 1950s--just a few years before the advent of performance-enhancing drugs. Ashburn peaked at 30, right when a player of his ability level should, though he had a fairly steep decline between 30-34 compared to his elite peer's comparisons of themselves. Most players can play within 95% of their peak OPS+ at age 38, but Ashburn was under 90 by the age of just 34. Now that we have a few baselines set  and an idea of what to look for in analyzing the data, we can start to take a look at some other sluggers who are playing and look for evidence of steroid use. Check back soon for more PED investigative charts.

No comments:

Post a Comment