I spent a while reading Van Trees' book on Detection and Estimation theory. Its really crazy but I find myself applying the framework of decision theory (or my interpretation of it) very frequently in my day to day life. What interests me about the framework is the notion of developing a criteria for making a final decision on the outcome of an unknown event. A criteria was developed a long time ago in the detection and estimation literature, called the "MAP Rule" (Maximizing the A posteriori Probability). Using this criteria, the optimum decision is specified for almost any scenario. The notion of optimum hangs over my head, Optimum. An ideal floating above, hovering in space.
This framework seems to be very practical: I could see it being applied in just about any scenario. Decisions are made, often with the interest of maximizing the probability of a positive result. For someone who can occasionally spend a lot of time evaluating outcomes, decision theory seems to be a potent tool. By codifying your scenario into the framework of decision theory (specifying the random elements, their probabilities and outcomes and specifying the range of decisions) the MAP rule can be applied to determine the optimum response.
My interpretation: Decision theory presents the following scenario: you are pitted with making a decision that will produce an outcome depending on the current circumstances. The full detail about the current circumstances are hidden, shrouded by the noise of imperfect observations. You must make a decision, at a specified time with a specified collection of available data. The data is your link to outside world, the world in which the outcome of each decision is a function of. Given no data, the optimum decision still exists, there is always an optimum response no matter how little information is available. However, as information becomes available, this response bends and changes with it. Plug in a couple of observations into the equation and the formula starts to carve itself and take form. The rangespace of preferable responses becomes constrained, perhaps to a solid line or perhaps to a parabola. Given a sufficient amount of information, the response will converge to a collection of points, perhaps only one. An amazing framework, all brought to life by using the MAP criteria.
Haraway and the postmodern philosophers disagree with these frameworks. Haraway would probably call the optimum decision a "God Trick." However, this optimum decision will probably never be clear cut and razor sharp: the observations are always shrouded in noise. Sometimes noise that cannot even be parameterized. I think it is a very useful tool for day to day life, as long as the noise is kept at the forefront. The noise is generated from the environment and also by the mechanisms that we use to observe the environment: our sensors that have been calibrated by the mind after so many years.
