Editors note: This story is part of a Feature "The Doping Dilemma" from the April 2008 issue of Scientific American.
A game theory model of doping in cycling applies to other sports as well, particularly baseball. For expert insight, I spoke with Lance Williams, an investigative reporter with the San Francisco Chronicle and co-author (with Mark Fainaru-Wada) of Game of Shadows, a revelatory book about how BALCO, the Bay Area Laboratory Co-Operative, supplied baseball players and other athletes with performance-enhancing drugs. When I outlined my ideas about how game theory explains massive cheating in sports, Williams backed up virtually every point: “Athletes have a huge incentive to dope. There are tremendous benefits to using the drugs, and there is only a small chance that you will get caught. So depending on your sport and where you are in your career, the risk is often worth it. If you make the team, you’ll be a millionaire; if you don’t, you’ll probably go back to driving a delivery truck.”
Once the top competitors in a sport begin cheating, the rule breaking cascades down through the ranks until an entire sport is corrupted. In Williams’s estimation, based on wide-ranging reporting and numerous interviews with athletes, coaches, trainers, drug dealers and drug testers, between 50 and 80 percent of all professional baseball players and track-and-field athletes have been doping. Given that reality, Williams explained, for many athletes not only is doping not cheating, it is necessary. To illustrate his point, Williams cited Charlie “the Chemist” Francis, coach of the sprinter and (briefly) the 1988 Olympic gold medalist in the 100-meter run, Ben Johnson, who was busted for doping and stripped of his medals. The doping was “completely self-defensive,” Francis told Williams. “It was cheat or lose.”
How can leagues and governing bodies change the incentives in the game matrix for baseball? Williams suggested stronger penalties against both individual athletes and entire teams, along with stiffer legal sanctions. Ironically, Williams and Fainaru-Wada personally felt legal heat for refusing to reveal their sources to federal authorities: “We stood to serve more time behind bars [both faced 18-month sentences] than any of the steroid dealers convicted in the BALCO conspiracy and any of the athletes we reported on, including [longtime San Francisco Giants slugger] Barry Bonds.” You know a system is corrupt when the messenger is shot and the gun makers walk.
See what we're tweeting about
3 Comments
Add CommentI would expect the doping to start with the competitors that are 'on the bubble' since the top competitors are more likely to win w/o the help and also have more to lose from getting caught. As the lower athletes increase in performance, then they force more of their competitors to follow suit.
Reply | Report Abuse | Link to this(...based strictly on the pure economic explanation)
Does this make sense to anybody else?
That may be how it begins in a sport. Twenty years ago one of the telltale signs that a track athlete had begun doping was that they suddenly would jump from an elite also-ran to a World/Olympic champion or record-holder. Ben Johnson was such an athlete. But once it takes hold, the non-doping athletes who were formerly good enough to be champions suddenly aren't good enough any more. Carl Lewis comes to mind. I would bet, based on his classic sprinter's body-type and his career progression, that he was the last Olympic sprint champion that didn't dope. Every single one since Lewis looks more like Arnold Schwarzenegger than Jesse Owens. Lewis likely represented the pinnacle of non-artificially enhanced human performance.
Reply | Report Abuse | Link to thisWhat is missing from this game theory analysis is the phenomenon of the effect of social perception of the crime versus the punishment - and the ancillary increase in the risk/reward for taking bribes to cover for a primary player.
Reply | Report Abuse | Link to thisThe reason that more draconian penalties don't work is that when the general societys perception that the penalty for doing something wrong is too harsh, people cover up for others more readily. This also feeds another crime of corruption, taking bribes, because the higher cost of being caught creates a more lucrative market, and changes the game for those players and there is also a better rationalization for helping the primary player. Thus, looking at one set of players in isolation in an artificially bounded system doesn't necessarily correspond to how things would work in the real world. This lack of holistic game modeling makes many a prediction fail. These are systems, not isolated entities.
An excellent and simple example of this sort of phenomenon is the drug war. Every increase in penalty increases the risk premium, which attracts players, which in turn increases the penalty, and an we go. Now 30% of the Baja police force in Mexico is estimated to be working for the narcos, and all customs inspectors on the Mexican side ignore "merchandise" going north. How often is payment extracted for letting things pass through? I think the real question is how often there isn't payment.
Model the whole system of players and associates and you will probably come to quite different conclusions.