I like the idea of using data and analytical tools for decisionmaking.. but I will also be the first to tell you that it's easy to overinterpret numbers or analysis. Here is a football analogy. Computers and sports are now forever embedded with each other. One of the more advanced attempts to model coaching decisions with software tools has come from the computer guru's of PigskinRevolution.com
. According to them, Miami could have greatly increased their chances of beating the Steelers last week if Nick Saban had actually tossed out that challnege flag Heath's long reception and run to the endzone. Here are their numbers
in precise bayesian detail. I have to admit that I am not convinced.
A little more interesting is their analysis of Superbowl XL
.. but again, without knowing the details of their model, I am not so sure about the information added by these results. I really wonder if anyone at CMU has tried to build a model to similar ends.