In the sports world, no punctuation mark is as dreaded as the asterisk. Asterisks never appear in a stats column for anything good. Thanks to performance-enhancing drug (PED) usage by high-profile players like Mark McGwire, Alex Rodriguez and Jose Canseco, the last few decades have become Major League Baseball’s “asterisk era.”
Baseball is not alone in confronting PED usage in its ranks and in the public’s perception. Due in part to high-profile cases like Marion Jones and Lance Armstrong, nearly 70% of adults polled believe that track & field, cycling and weightlifting have a PED problem. Figure 1 shows data from the World Anti-Doping Agency that validates this perception. These three sports account for 44% of the “adverse analytical findings (AAF),” despite comprising only 29% of the total number of samples tested.1
Outside of the struggles with PED usage, one is hard-pressed to find any similarities between baseball and these three “Olympic” sports. Of this group, baseball is the only team sport. It is the only one played with a ball. It is the only one that has offense and defense. There is little commonality between the size, shape or demographics of the athletes.
While many authors look to cultural, societal or demographic drivers of PED usage in baseball, these theories neither distinguish baseball from sports that don’t have a PED problem; nor do they explore the common links between baseball and other sports that do.2 As we will see, the game structure of baseball, the task demands placed on the athletes and the metrics of success make it more similar to these Olympic sports than to other team field sports. The game of baseball itself creates many of the same incentives for PED use as track & field, cycling and weightlifting.
Baseball’s Task Demands
“I ain’t an athlete, lady. I’m a baseball player.” – John Kruk, 3-time All Star
Throwing the ball and swinging the bat are complex neuromuscular tasks that are very simple from metabolic and biomechanical perspectives. The ability to throw pitches past a batter or to hit pitches out of the park is the result of the power the player is able to produce – the amount of force and the speed at which it is generated. Both pitching and hitting require maximal or near maximal muscle activity for barely 1 second. To pitch or slug effectively, then, pitchers and hitters only need to train the ATP-CP metabolic pathway (Figure 2 below).
Pitching and hitting are biomechanically repetitive tasks. Athletes train to make their movements as efficient as possible, which entails minimizing the variation between in-competition work bouts. The differences between a fastball and a slider, or between a swing low-and-inside or high-in-the-zone are negligible from a gross mechanical perspective.
The pitcher who can generate more power per pitch will throw faster, and the pitcher who can do that more times per game, more games per season and more seasons per career will amass better single-season and career statistics. At the plate, the batter who can generate more torque and power will increase his bat speed, pushing outfield flyballs over the fence, and grounders and line drives past the infield. His slugging percentage will increase as will his on-base percentage, particularly if he draws more walks from pitchers who fear his long-ball prowess.
Baseball’s Game Structure
Baseball is ostensibly a team sport competed by two opposing lineups of 9 players. The game itself, though, is really a series of discrete one-on-one contests between the pitcher and the batter. These contests are non-dynamic. Once the pitcher releases the ball, he has no further impact on the batter’s action. The next event in the contest is solely determined by the batter (barring an unusual event like catcher interference). A small percentage of the time, the batter will put the ball in play and involve one or more of the other players on the field.
Baseball’s statistics reflect the reductive nature of the game.2 Pitchers are judged primarily by their defense-independent pitching statistics (DIPS) strikeouts, walks, and home runs allowed. Earned-run average is the most-reported pitching stat, but can be misleading because it is highly dependent on the other eight players on the field, and a bit of luck. DIPS reflect the pure contribution of the pitcher to preventing the opposing team from scoring runs.
Likewise, hitters are assessed based on their slugging percentage and on-base percentage, which are often combined into the OPS (on-base plus slugging). A player’s OPS has nothing to do with who is in front of or behind him in the batting lineup. It’s only about the end result of each mini-contest he engages in with the opposing pitcher. Hitters’ defensive performance in the field is ultimately a qualitative assessment by the coaching staff. Even errors, one of the few data points for fielders, are the result of subjective decisions by the official scorekeeper.
Baseball’s Evaluations and Rewards
No one should be much surprised to hear that a team’s run differential (runs scored – runs allowed) strongly and linearly correlates with winning percentage (Figure 3). Run differential also correlates nonlinearly to team revenue. This means that while a team does better financially when it is above average, it does even better than better when it is further above average.4
How important are pitching and hitting statistics to the player? Run differential can be thought of as Hitters minus Pitchers. By determining a player’s contribution to runs scored or allowed via OPS or DIPS, respectively, we can quantify his contribution to the team’s record. This also means that we can quantify each player’s contribution to team revenue.
Economist J.C. Bradbury calculated the marginal revenue product for baseball players: how much each additional run allowed (pitchers) or scored (hitters) is worth relative to league averages. Bradbury states that each point (that is, 1/1000) of on-base percentage equates to just over 3 runs over the course of a season. Working through a few examples, and accounting for the percentage of a team’s at-bats taken by a given player, Bradbury shows that Nomar Garciaparra hit for $100,000 above the team’s average revenue-per-player in 2005. Meanwhile, Garciaparra’s teammate Neifi Perez “cost” the Cubs $1.85 million relative to the average revenue-per-player.
Quantitative Juicing: Injecting Money into the Game
As we learned from reading or watching Moneyball, these are just a few of the dry, statistical inputs that go into the decisions about whether, where and for how much a baseball player will play. These statistics are the nexus of the incentives and structures that drive PED usage in baseball.
We can classify sports data into four categories: human performance, athlete performance, sports performance and derived data. Performance-enhancing drugs act upon the physiological mechanisms of human performance, resulting in improved athlete performance via greater power production. Figure 2 shows that a single energetic system underlies very short or very long activities, such as throwing, hitting or weightlifting at one extreme or endurance running and cycling at the other. These extremes of power production can be easily targeted and improved upon by PEDs such as anabolic steroids or human growth hormone in power sports, or erythropoietin (EPO) in endurance sports.
Enhanced athletic performance in these sports will show a direct and quantifiable improvement in the athlete’s sport performance and derived statistics. A 10% increase in a baseball player’s muscle mass increases by 50% the fraction of hit balls that go out for home runs. This is the result of the player generating more force with the bat, either as a function of bat weight or bat speed. The speed of a batted ball increases linearly with an increase in swing speed, and approximately logarithmically with bat weight (Figure 4). A 2-mph increase in bat speed (~5%) can send a batted ball an extra 15 feet: the average length of the warning track.
On the defensive side, a 10% increase in muscle mass can result in an extra 4-5 mph of pitch velocity. This correlates to a reduction of 0.5 ERA, which can be a difference of 10-20% whether you’re a backbencher or a starter.5
Individual statistics play such a large role in baseball because, as discussed earlier, baseball is ultimately an individual sport. An offensive threat in basketball can be double-teamed or boxed out. A quick-passing counter-attacking offense in soccer can be shut down by parking the bus in front of the net and closing down passing lanes. The only way to avoid an offensive threat in baseball is to intentionally walk the batter, which will still boost his OBP. For pitchers, defense-independent statistics are just that: defense independent. An entire team of Buckner’s behind the mound would not impact the conversation the pitcher will have with his agent, GM and accountant at contract time.
Conclusions
The more that success in sport is dependent on a single physiological system or discrete mechanical task, the more that performance in that sport can be enhanced by PED use. The more efficacious PEDs are, the greater the incentives for athletes to use them. Likewise, sports that have a non-dynamic interplay between competitors will also be more likely to confront PED use because the individual abilities and contributions of each athlete will be laid bare.
Like track & field and cycling, the game of baseball lends itself to PED use. The design of the game and the athletic tasks that form the basis of both winning games and determining a player’s financial impact are ideal for illicit enhancement. The athletic task demands are readily enhanced by illicit means, and a player’s mastery of these demands and his contribution to the team are easily quantified through meaningful statistics. Because the incentive structure to use PEDs will always be part of the game, any attempt to eliminate PEDs from baseball must affect the other side of the equation: the risks and costs. Baseball players will need to believe both that they will absolutely get caught, and that the consequences will be certain and severe enough so as to outweigh the obvious incentives.
Figure 1
| Sport | % of total samples tested | % of total AAFs | % of samples returning AAF |
| Track & field | 12.7 | 17.5 | 2.1 |
| Cycling | 12.2 | 15.8 | 1.9 |
| Weightlifting | 3.8 | 10.3 | 4.1 |
Figure 2
Figure 3
Figure 4
Notes and Further Reading
1) Because baseball is not an Olympic sport it is not under the jurisdiction of the WADA. Even so, Major League Baseball would not be subject to WADA protocols or reporting because it is a professional league independent of the Olympic and national governing bodies.
2) An easy reference for understanding baseball statistics can be found at Baseball Almanac, or Appendix B of Bradbury, note 4.
3) For an insightful look at the disparities in PED usage between American and Dominican baseball players, and the economic factors underlying these, see “What Isn’t in the Mitchell Report?” in Scorecasting: The Hidden Influences Behind How Sports are Played and Games Are Won, by Tobias J. Moskowitz and L. Jon Wertheim
4) See Bradbury’s The Baseball Economist: The Real Game Exposed.
5) If you are tired of the economics links, the academic research of Dan Russell and Terry Bahill will provide hours of enjoyment into the physics of baseball.