We are surrounded by games of chicken. Courageous antigovernment protesters announce that they will not back down, whatever the forces of repression may do; governments (of Tunisia, Egypt, Bahrain, Libya, fill in the blank) threaten that they will clear the streets and stay in power, whatever protesters may do. Business owners (and increasingly, Republican state governors) announce that they will never accede to union demands; union leaders reply that they will not be intimidated. “We’re eyeball to eyeball,” said Secretary of State Dean Rusk in 1962, as the Cuban missile crisis passed its near-apocalyptic outcome, “and I think the other fellow just blinked.” Games of chicken all.
As games go, they can be serious, and deadly.
Whenever two parties can be described as on a “collision course,” you are very likely witnessing a game of chicken. In 1995, Congressional Republicans, emboldened by their electoral success the year before, played chicken with President Bill Clinton, threatening to shut down the U.S. government unless they got their way. Clinton countered with his own threat. A kind of train wreck occurred, but commentators agree that the damage was not mutual: Newt Gingrich and company came out worse because the American public blamed them (correctly, it appears) for the collision.
Although such encounters are the stuff of a highly technical discipline known as “game theory,” they also embrace some of the most fraught and frequent interactions that nonmathematicians encounter. Like war and generals, or politics and politicians, games of chicken are too important to be left to the game theorists alone. So here is a primer.
What happens when a chicken, instead of crossing the road, decides to run headlong into another chicken, who is similarly determined? The result could be a classic game of chicken, if certain conditions apply. For one, there must be two participants. For another, each must have two choices—in this case, swerve or go straight ahead. The decision is influenced by the perception of what the other player is likely to do, but must nonetheless be made independently, and payoffs have to be a consequence of both what each player decides and what the other one does. Most important, the payoffs must be in a particular relationship to each other.
Consider the classic scenario. Two cars speed toward each other. Each driver can do one of two things: swerve or go straight. To win, you must go straight; the one who swerves is the “chicken.” If both drivers swerve, both are “chickens,” but neither suffers relative to the other. Or, and here is the crunch, literally: If both drivers go straight, both lose, possibly forever.
It is said that games of chicken were first played by California teenagers during the 1950s, although that may simply be an urban legend. The philosopher Bertrand Russell, however, saw a gruesome parallel with nuclear brinkmanship: Each side wants the other to back down, although neither is willing to do so itself. A head-on collision beckons. Fried chicken.
When theorists analyze these interactions, they designate backing down or swerving as “cooperating,” and going straight as “defecting,” since the former results in a mutually beneficial outcome while the latter is selfish. The peculiar interest of such situations is that whereas each side is tempted to defect (go straight), thereby gaining the highest payoff, that is possible only if the other side cooperates (swerves) and accepts the lowest. And yet, if both defect, both lose out.
In game-theory terms: If driver No. 1 goes straight, he gets T, the Temptation to defect, but if, and only if, No. 2 cooperates—that is, if No. 2 swerves and thus gets S, the Sucker’s payoff of being a chicken. If both swerve, each gets R, the Reward of mutual cooperation: No one wins, but no one loses, either. The dramatic outcome, however, is P, the Punishment of mutual defection if both drivers go straight ahead.
To be a game of chicken, the payoffs must be in this relationship: T > R > S > P. Aficionados will note, however, that when T > R > P > S, another intriguing game is afoot: a Prisoner’s Dilemma, in which the worst payoff is to cooperate when the other player defects and gets T. The Prisoner’s Dilemma is especially frustrating because its logic results in both players’ getting P, which is worse than each could have gained by cooperating, i.e., R. In chicken, the worst payoff is P, the punishment of mutual defection.
Swerving may seem rational, but if you think the other fellow is a swerver, you are tempted to go straight. The rub is that he or she is thinking the same thing. Is it best, therefore, to swerve? Not if the other player also swerves, since then you could win by going straight. Should you go straight? Not if the other player also does! And so the “game” often boils down to a matter of communication, or rather manipulation: trying to get the other person to swerve.
Accept, right off, that there is no way to guarantee victory. The best either player can hope for is to improve the odds of inducing the other player to swerve. Toward that end, there are many tactics, none especially appealing. Start with reputation. If you are known as a nonswerver, perhaps because you have been “undefeated” in previous encounters, your opponent is bound to take that into account. Not surprising, national leaders have long been concerned that their country be known to stand by its commitments, just as President Richard Nixon, during the Vietnam War, was worried that the United States not be seen as a “pitiful helpless giant.”
Reputation can be burnished in several ways, like cultivating an image of being literally crazy, or perhaps suicidal. Imagine you are playing car chicken against someone who gets into his or her vehicle laughing maniacally and swearing that death is welcome. You would be crazy not to swerve! Whether actually insane or simply faking it, people get a payoff from convincing their opponent that they have taken leave of their senses.
In fact, according to H.R. Haldeman’s memoir, The Ends of Power, Richard Nixon attempted to convince the North Vietnamese that he was so crazy that his irrational threats to employ nuclear weapons must be taken seriously. To no avail. However, it is noteworthy that game theory, developed as a means of applying a high level of rationality to reveal optimum tactics and strategy, ends up—in this case at least—recommending behavior that is so irrational as to be insane.
There is also another variant on the Madman Theory: convincing the other player that you are unwilling or literally unable to swerve. The logical but nonetheless bizarre consequence suggested in the 1960s by that bizarrely logical nuclear strategist Herman Kahn is to wait until you have reached high speed and then throw the steering wheel out the window, showing the other driver you can’t swerve ... which then generates a contest to see who can throw out the steering wheel first. To be sure, as Kahn himself acknowledged: “If his opponent is not watching, he has a problem; likewise if both players try this strategy.”
Because it is deeply irrational not to swerve, the chicken players’ dilemma is identified by nuclear strategists as one of “credibility,” since nothing is to be gained by obliterating an opponent who has already attacked you, and much is to be lost (for example, livability of the planet). And so—strongly influenced by game theory, it must be noted—the likes of Mr. Kahn recommended the nuclear equivalent of tossing out the steering wheel: a policy of “launch on warning,” whereby retaliation would be assured by computers, taking the prospect of rational intervention by human beings “out of the loop.” In Stanley Kubrick’s great movie Dr. Strangelove or: How I Learned to Stop Worrying and Love the Bomb, the Soviets installed a Doomsday Machine, wired to automatically destroy the world in the event of a nuclear attack on the USSR. Their strategic misstep: They neglected to tell the Americans. (Think back to Kahn’s warning that the driver who discards his steering wheel “has a problem” if his opponent isn’t watching.)
There are other ways of convincing the oncoming driver that you aren’t going to swerve. Your determination to go straight may depend on your desire to be victorious; the more important the victory, the more risk you are willing to take. In public affairs, it is widely assumed that countries or institutions are more inclined to persevere in a game of chicken if their deepest interests are directly engaged. When Russia invaded Georgia, in 2008, no one seriously expected the United States to intervene militarily, because events there did not directly affect American security.
Another tactic: Drive a large and imposing vehicle. If an armored cement truck is confronting a VW Beetle in a game of chicken, who backs down? That goes a long way toward explaining why the Soviets turned aside during the Cuban missile crisis, since the United States was militarily superior, in both conventional forces available in the Caribbean and nuclear might. Following the near-disaster, in 1962, a high-ranking Soviet official was heard to say, “You bastards won’t be able to do this to us again,” and shortly thereafter, the USSR began to increase production of ICBM’s. Alas, everyone is less safe when two monster trucks confront each other.
Even an asymmetry in “vehicle size,” however, doesn’t always enable us to predict the outcome of chicken. In late 2002 and early 2003, the Bush administration played chicken with the regime in Baghdad, demanding that Iraq comply with U.S. demands or face war. The U.S. military was incomparably stronger than Iraq’s, so the result was never in doubt. Here was a tractor-trailer bearing down on a bicycle. Why didn’t Saddam swerve?
Maybe he didn’t believe that the United States really meant it. Maybe he clung to the hope that international opinion would intervene and grab the steering wheel of U.S. policy. Maybe he couldn’t tell which way he was supposed to turn. Because U.S. demands—for eliminating “weapons of mass destruction” and renouncing any connection with Al Qaeda—assumed offenses that didn’t exist (and couldn’t be disproved), he might have concluded that however he swerved, the United States would swerve into him. In short, maybe the Iraq war wasn’t a game of chicken after all, but rather a case in which one side wanted a collision.
Thus, all things being equal, the likely victor in a game of car chicken is the person who shows up wearing body armor and a crash helmet. The bomb-shelter craze of the 1950s was motivated, in large part, by strategists’ worries that if the Soviets had an advantage in “civil defense” (which actually was neither civil nor defense), they would take more competitive risks vis-à-vis the United States. So Americans were encouraged to build family shelters, less in hope of saving lives in the aftermath of a nuclear war than as a policy instrument. And, of course, in the hope of making the Soviets more likely to swerve.
With the cold war behind us, there is something so arcane and downright weird about these parallels between games of car chicken and the Cuban missile crisis that they evoke nervous laughter. The logic of competitive risk-taking seems absurd. Keep laughing, but remember: Games of chicken have a habit of creeping up on us. Better yet, consider the advice offered by a top-secret Defense Department computer in the 1980s movie WarGames: “The only way to win is not to play the game.”
