I have all the answers, I just can't remember which questions to match them with.
Thanks for an excellent and amusing article. Mr Bayes has a lot to answer for. I have done a bit of work with decision analysis, and Monte Carlo methods are my stock in trade, so I was very tickled to see the great pure mathematician Erdos being persuaded by a computer simulation rather than by reasoning. I am a big fan for computer simulation, and the problems I work on tend to have a stochastic element.
There is no doubt that human beings have a lot of difficulty with probabilistic concepts (and so do many mathematicians). I've done some teaching in the local school of engineering. All engineers have to pass some (very applied) maths courses, which include a small amount of probability and statistics. Until a year ago they could always avoid the probs stuff in the final exam, because they had to answer, say, 6 out of 7 questions, one of which was on probability. It was found that only about 10% of the class would tackle the probs question, and not only that, a large number were not even attending the relevant lectures. Since it is plain that some grasp of probability concepts is very useful for an engineer, the exam policy has been changed, making the probs question compulsory.
The main weakness of decision analysis in real decision contexts is that you need to have some idea of the probabilities involved, and they are seldom available. For example, I've never considered trying to apply any of these methods to investment decisions, where, as you say, the main factors are emotional/psychological issues, particularly mob psychology. |
