The Game Is In Theory

Game theory is promoted as a system where you can apply to any aspect of life. This area of economics has had excellent public relations for a long time. It traps people to take a simple idea and apply it to complicated situations. Its call for imagination is muddling the gap between its usefulness and how much it is being used.

Game theory is predictive of the behavior of agents who abide by that theory, and are thus worth off due to this behavior. As how a set of fables can only give insights but never advice in real life, game theory misses out on a lot of information and relevant details. This loss comes from formal language and abstractions that are far away from \(1 - \epsilon\) of the population.

The expected utility model proposed by von Neumann-Morgenstern was the basis for game theory. People still use this principle to model uncertainty even after Duncan and Howard^[1] criticized it for decades and offered better alternatives. There are many ideas in game theory that have not been fully developed. And most of them may fall in the sphere of pure intellectual analysis and not always hold practical relevance.

Game theory makes simple abstractions of strategic situations, and formalizes them into [often complex] models. It is only good at offering strategy-proof alternatives. Even so, one good real-life example of its strategic employment is in security. Many research publications have promoted its broad usage in security. Contributions from Thomas Schelling, von Neumann and John Nash while he was at RAND Corp., Milind Tambe^[2] and team towards better security are some examples. Discourse comprehension and formalizing argumentation is another unsolved interesting problem for game theory. The best human argumentation model may only be achieved by a highly performant game theoretic model.

For any decision problem, the combined optimization + control theoretic approach often offers a better alternative. MIP with a class of nonlinear convex constraints using an FSM/MPC is an effective template for this. Social and economic policies hedge on micro-economic theories with strong game theoretic considerations. Market design, however has aspects of efficiency, optimization and non-strategic based dynamics. We should stop finding solutions in game theory to problems that theory has nothing to say about.