Zero-Sum Foolery 3 of 4: Forecast Factory
by Sandwichman
Foolery 1: Games Theory Gamesmanship
Foolery 2: Doomsday Climate Machine
Zero-Sum Foolery 3 of 4: Forecast Factory
Long before the issue of anthropogenic climate change arrived on the doomsday agenda, Lewis F. Richardson anticipated climate modeling with his failed attempt to forecast weather numerically. His calculations predicted surface pressures 150 times higher than observed:
Paradoxically, then, the most time-consuming, precisely calculated forecast in history was also among the least accurate ever prepared by any method.
Some consolation could be had, though, from Richardson’s fantasy of the “forecast factory.” In A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming, Paul Edwards praised Richardson’s metaphors of the factory for “reach[ing] to the heart of computing as a coordinated human activity that harmonizes machines, equations, people, data, and communications systems”:
At the same time, they stand in stark contrast to today’s dominant metaphors of computation, which are mostly individual: the brain, memory, neurons, intelligence. Richardson’s forecast-factory remains a better description of the practical reality of computing. The limits of computer power, even today, stem from these human and material dimensions.
Richardson had more success investigating the mathematics of arms races. The September 1957 issue of Conflict Resolution was devoted to Rapoport’s essay on Richardson’s mathematical theory of war, an essay that Schelling described as “magnificent.”
In his assessment of the failures and successes of Richardson’s theory, Rapoport stressed the investigative rather than explanatory function of mathematical models: “Contrary to a prevalent meaning of ‘model’ in many theoretical formulations, the main function of a mathematical model is not an ‘explanatory’ one.” The distinction is fundamental to Rapoport’s profound methodological objection to the pretentious “rational” pursuit of solutions to problems and answers to questions. In Strategy and Conscience, Rapoport reaffirmed that “The important end product of such [experimental] research is not an answer but a question.”
The value of game theory, Rapoport was later to insist, lay precisely in its demonstration of the limits of supposedly rational choice. This is an insight that the strategists have either never grasped or refused to acknowledge. Philip Mirowski has been scathing in his criticism of “the strategic community” – including Schelling – for their misrepresentations of “what game theory could ever hope to do.” The strategists’ “image of game theory was one of the purest instrumentality, of the labcoated expert ‘thinking about the unthinkable.'”
Not all “unthinkable” things were eligible to be thought about, though. Some thoughts, namely Rapoport’s eloquent critique of strategic thinking, had to be castigated and dismissed as “defeatist,” “moralistic” – much as the economists feel compelled to ridicule the fallacies of those who refuse to genuflect to the prescribed articles of faith.
As Rapoport observed in his reply to Albert Wohlstetter’s bitterly dismissive commentary on Strategy and Conscience:
…the cognitive assumptions of the strategists are neither revealed truths nor self-evident facts. They are rather derivatives of a power-oriented value system, which sharply delimits the cognitive horizon of its adherents. It is high time we stopped identifying narrowness of vision with ‘realism.’ It is high time to stop calculating long enough to think awhile, perhaps even to listen to the voice of our conscience.
Rapoport’s reply compared the strategists’ assumptions to the way that 19th century political economy “conceptualized man’s economic activity in a way which made it appear inevitable that the poor must forever remain poor.” Central to that conceptualization were reverence for what eventually became known as Say’s Law and the wages-fund doctrine, which conceived the wages-fund as a zero-sum game in which trade union action to secure higher wages for one group of workers could only result in lower wages for others.
cross posted with Econospeak
I suppose I should renumber this as 3 out of 5 or perhaps number 5 in the series will be 5 out of 4! I’m working on a piece about Rapoport’s production and exchange model that uses a differential equation to arrive at some rather counter intuitive outcomes. Not being a math wizard, it is taking me some time to make sure I understand what R. was talking about.