The Mathematical Prophet

Should we heed his word?
The Predictioneer’s Game: Using the Logic of Brazen Self-Interest to See and Shape the Future, by Bruce Bueno de Mesquita, Random House, 2009.  $27, 248 pp.
Is life a game-and nations, corporations, and individuals merely players?  Bruce Bueno de Mesquita thinks so.  In The Predictioneer’s Game, the applied game theorist takes his model of human interaction and employs it to predict the results, whether  the task is negotiating the best price for a new car or making peace between Israel and Palestine, claiming a 90% success rate.  The book focuses on neither why nor how this model makes such accurate predictions, instead attributing that success to game theory’s use of “logic and evidence.”  Rather than share that logic, the book focuses on how the model has been used.
Throughout most of the book, Professor de Mesquita insists that we can think the same way he does.  This humility is consistent with his ideas about human behavior, namely that our actions are largely shaped by the incentives we face, and we’re all quite similar.  Still, the bravado that has made de Mesquita famous, and his failures to meet his own standards of “logic and evidence” detract from the power of The Predictioneer’s Game to usher in a new paradigm in the way we think and strategize.
A Man and a Model
De Mesquita’s model treats the result of any negotiation as a function of the following attributes of the players: position, influence, salience, and desire to reach an agreement.  He, his students, or “experts,” provide numbers for each attribute for each player.  With all the necessary data, de Mesquita’s computer program iterates through multiple rounds of negotiations to see how players’ attributes change until either equilibrium is reached or it is clear that no agreement will be made.
It is easy to forget that de Mesquita created the model discussed throughout The Predictioneer’s Game; after first introducing his model, he rarely mentions that he created it, instead referring to it simply as “the model.”  De Mesquita instead largely acts as a medium between our world and the model, taking descriptions of reality as it is and returning the model’s predictions of what it will become.  The model takes on quasi-mystical properties and a personality of its own.
Anyone, de Mesquita humbly insists, can produce the numbers that the model needs as inputs with some careful thought and publicly available information.  He largely shows this by giving examples of the predictions that his students, mere NYU undergraduates, have made with their professor’s model.   The implication is that we can be predictioneers, too.
De Mesquita himself often consults experts of a particular culture or a situation to produce the assumptions to plug into the model, but he claims that the resulting predictions are barely more accurate than hastily-made assumptions.  He considers speed to be the advantage of the experts, who can quickly describe a situation drawing on their specialized knowledge.  That is where the expert advantage, in de Mesquita’s view, begins and ends: “Knowing about places, and how different they may be, is important, but, perhaps surprisingly, it is not as important as knowing about people, and how similar they are, wherever they are.”  In this broad claim about his model’s strength, he overlooks the fact that even the publicly available information he plugs in himself is largely shaped by experts.  The model will at best supplement, rather than replace, expert knowledge of regions and cultures.
His treatment of the potential Iran-Iraq partnership exemplifies the dichotomy between the man and the model.  He considers what kind of partnership five key players, al-Maliki, Khameini, Ahmedinejad, al-Sadr, and the Bonyads, would like to see between Iraq and Iran, and he conditions his answer on whether or not the US fully withdraws or maintains 50,000 troops in Iraq.  His analysis of their initial positions is convincing.  His explanations of how their positions change and the ultimate equilibria, however, only interpret the data produced by the model. Sure, it is interesting to see that Maliki and Khameini will move towards a compromise while Sadr and the Bonyands won’t, but the predictive approach sheds no light onto why this happens.
Dual Personalities
Bruce Bueno de Mesquita did not make his name by being humble; is he the bold prophet, or humble professor?  Perhaps he is both.  The introduction of Game begins with a discussion of Belgium’s King Leopold II and a comparison of Leopold’s reign in Belgium, where he passed labor-protection laws and supported public works during an economic crisis, and his reign in the Congo, where he enslaved the Congolese and extracted as much ivory and rubber as possible.  De Mesquita explains this-and, really, everyone’s behavior-as a response to incentives.  Leopold was limited by the Belgian constitution and would be deposed if he overstepped the bounds of his rule at home, but there was nothing to limit his power in the Congo.
Like Leopold, de Mesquita may be responding differently to two different sets of incentives.  As the narrator, in order to gain the reader’s trust, he is modest, matching his belief that humans are fundamentally very similar and shaped by their incentives and limitations.  As an author and a consultant seeking attention, he promotes his image as a modern prophet.
A Good Guess
Although he generally stays humble in the book, de Mesquita subtly drops hints of his prophetic power.  He tells us of the millions or even billions he passed up when he declined to help topple Anwar Sadat or to help keep Zaire’s Mobutu Sese Seko in power.  He declines to say why al-Maliki wouldn’t need to seek help from Iran because “there is a reasonable chance that Iraqi or Iranian diplomats might read this.”  If his model is only based on logic and evidence, as he so often claims, shouldn’t other game theorists be able to recreate the model?
Even if de Mesquita is unique in our time, is he unique in history, the first to be able to predict the future quite accurately?  Or is he a reincarnation of the Count de Saint-Germain, whom Voltaire considered to be all-knowledgeable.  The educated of that era believed in alchemy, Saint-Germain’s specialty, as much as the educated of today believe in modern economic analysis.  Bruce Bueno de Mesquita’s accomplishments are impressive; but claiming to use standards of logic and evidence isn’t groundbreaking, and de Mesquita’s anecdotal style and structure don’t conform to the very same standards.  Until he does so, we’ll continue to listen to the conventional experts as well.