THIS WEEK IN PRO CYCLING »

Get the VeloNews Email Newsletter FREE

  Learn More | Archive
Home » Top Stories » News and Features »

Wednesday's mailbag: Purchasing decisions and Game Theory

Published: Dec. 28, 2005

The Mailbag is a regular feature on VeloNews.com. If you have a comment, an opinion or observation regarding anything you have seen in cycling, in VeloNews magazine or on VeloNews.com, write to WebLetters@InsideInc.com. Please include your full name and home town. Letters may be edited for length and clarity.

To subscribe or not to subscribe
Editor;
There have been several times I've wanted to write, but Mr. Kessler's comment inMonday's Mailbag irritated me enough to finally put fingers to keys.To not subscribe to a magazine because of one individual is ludicrous. It's like not watching television at all because you don't like a news anchorperson. I, too, tired of the Lance/Sheryl Crow coverage. But we do live in a free and democratic country - so subscribe, don't subscribe, it's up to you. However, don't you think it's sensible to skip those articles to keep up with technological tips, equipment upgrades, team news and the other fine articles in VeloNews? I do.Let's just get over it people, and focus on the positives. I'm looking forward to a new year full of great rides and riders. CSC should be the team to beat in next year's TdF if everyone is healthy.
Terry L. Rhedin
Bremerton, WashingtonYeah, we were a touch surprised, too. Usually the single individual most angry letter writers blame for their decision not to subscribe or to cancel an existing subscription is O’Grady. – EditorThe dilemma explained
Dear Veloletters,
As an avid VeloNews.com reader, cyclist and professional economist/statistician in industrial competition (the branch that often uses game theory), I felt the strong need to respond to the recent Prisoners' Dilemma letter. A good example of the Prisoners' Dilemma can be found at Wikipedia.While the writer from Ottawa is correct that there are elements of the classic Prisoners’ Dilemma, and while it is flattering that game theory is gaining wide acceptance, a little bit of knowledge can be dangerous. Perhaps rarely has a certain bit of information been so poorly applied.The conclusion that "all cyclists are dopers" is about as far off the mark as a competition economist concluding that markets with only two players (a near monopoly) would yield competitive (low price-high quality) outcomes, which is one of the original erroneous uses of game theory asserted by those misusing game theory. This conclusion is true only in special cases.The classic Prisoners’ Dilemma will provide robust conclusions only under certain assumptions/circumstances. One of them, (start simple with a two player game) is that each player will do strictly worse if the other player "finks," which is analogous to cheating in sport, or in the case of competition, on a tacit agreement to keep prices high. The players also need to have very good knowledge about the payoffs. One has to know what the likelihoods of being caught, the benefits of doping, etc. Thus it follows that, given that one will always do worse if the other cheats, one will then, rationally decide to cheat. In sport that would be the decision to dope, or in the industrial competition sense, cut prices or increase production. So first, we see that all it takes for the Prisoners’ Dilemma to break down in the cycling application is for one player to recon he/she will do strictly worse if cheating. This could be due to exceptional talent, or very high penalties from getting caught, uncertainty about the payoffs, etc.Uncertainty matters a lot too. Think of the difference between your best and worst day in your district TT. The time difference is probably greater than the difference between your “would-be-doped time.” So one bad day or stomach flu could waste all that cheating but the chance of getting caught for top riders is only reduced a little, with out-of-competition testing.Another important element of the Prisoners’ Dilemma is that both players could conceivably "both be better off" if they could "agree" not to cheat. Here is the fundamental difference between the industrial and the sporting competition applications. In industrial competition, "agreements" to keep prices high are criminal offences under the Sherman Antitrust Act or the National Competition laws transposing the Articles of the treaty of Rome in Europe. Quite the opposite is, of course, the case in sport; it is both a sporting and often criminal offence to dope.Once we start introducing multiple players and uncertainty over payoffs, asymmetry of payoffs, as well as “repetitions” of the game, robust conclusions get messy. There is also a theorem in game theory (folk theorem) that says that almost any equilibrium can be sustained in a repeated game (a stationary point where no player wants to change her strategy). Thus, there is very little for sure we can say about cycling and cheating using game theory. Now probably some insight that we can gain from the Prisoners’ Dilemma applied to sport is that the incentive to cheat increases as the number of cheaters decreases, because the payoffs would be greater.But we still can't be sure that this would induce cheating. One must assume that if everyone was 100-percent clean, then the average Joe in peloton would have a very high payoff from cheating, weighed against the expectation-weighted penalty from possibly getting caught. The penalties might still be very high, or the average Joe might not get that much benefit. Now I have no personal experience, but I would reckon that the lantern rouge, no matter how much he/she doped, would never win the Tour.So what can one conclude from all of this? Well, certainly the idea that the Prisoners’ Dilemma tells us a lot about the percentage of dopers is unlikely to be true. It does tell us that it will be virtually impossible to wipe out all cheating, since the incentives to cheat increase as the percentage of dopers gets lower). It also tells us that a rational policy response would be to increase the penalties for cheating (which WADA and the UCI have done), add more out-of-competition testing and work to stay ahead of the drug cheats by adding money to research (which they have also done). The conclusion that vigilance is needed still holds.Finally, in my opinion, erroneously concluding that all or even most riders are cheats and then trumpeting this message actually encourages cheating. Why? Because one is removing one of the penalties for cheating: Lost Honor. When a young rider hears that he is going to be assumed to cheat whether he does or does not, one of the incentives to stay clean has been demonstrably removed.
Greg Swinand, PhD
Director/London Economics
Wicklow IrelandCheers to Brandon
Hey guys,
I just wanted to let you know that Brandon Dwight’s diaries are great – keep ’em coming. Almost none of us will get a chance to race cross in Europe, and he’s giving some great insight to the experience.Thanks dude!
Clay HarrisDitto
Hola VeloScribes,
Not that I live for fluff pieces but I am really enjoying the Brandon Dwight Diaries: Very cool reading, with some great semi-personal perspectives and some great pics.This series is a welcome distraction from all the doper pieces, Lance retrospectives and the general ennui of cycling politics. It is great to hear about the people.Thank you Dwight and thank you VeloNews.com for putting the good stuff up!
Regards,
Georges RouanWe have to admit that we didn't know what we were going to get when we made a deal with Brandon for his diaries. It's been a very pleasant surprise for us, too. - Editor
The Mailbag is a regular feature on VeloNews.com. If you have a comment, an opinion or observation regarding anything you have seen in cycling, in VeloNews magazine or on VeloNews.com, write to WebLetters@InsideInc.com. Please include your full name and home town. Letters may be edited for length and clarity.