I did indeed try and test this formula considering a final exam. Results indicated that the model is a good predictor of trait procrastination but does not hold when considering procrastination behavior.

I suppose you have to qualify as a psychologist before you can begin to understand the theories that have been discussed here.
To say it was beyond my capabilities is an understatment however I grassped a little bit and it does make an interesting point.

Motivational Theories

First off, I agree that human behavior is, in general, complex to the point of being pretty much undefinable by a 5 variable equation, and that future research in this field should definitely include concepts from chaos theory, and much more... But I think that the daily lives of many people are constrained enough, by various responsibilities, obligations, and even human needs, that a slightly more robust version of the TMT equation could be used to give a kind of averaged prediction of a person's behavior.
After reading Steel's article, I too was noticed the problem with the graph in which socialization remained constant, but then I realized that on that same graph one could put every activity that we do in the course of a day (once again, with a slightly modified eq.) The main problem, i guess, would be in discovering what additional variables need to be added, but I am definitely intrigued by the graph in my mind; with the Utility values of various actions flowing up and down, coming into and going out of existence, and altering slightly in time.