That's just the psychologists approximation to behavioural science. Psychology, like every other science, is merely physics with limitations and approximations. If you want to be right, as opposed to just getting a useful approximation, you need to look at the physics of the situation.
The fact is, you are (in the consensus of scientists) quite incorrect, when stating that the person would always have bought the coke. The universe is not predetermined, according to our best theories. The state of the universe is indeterminable. When we measure a particle's position, it's velocity is not only unknown, but also impossible to know. It is possible (if exceedingly unlikely) that your entire body could spontaneously teleport a short distance, and there is no way of predicting this. A tiny change to our brain chemistry is entirely possible.
So we have a situation where no-one, no matter how intelligent they are, and how much data/time they have, can predict what we will do with 100% accuracy, and it is because of the building blocks that make up our body/brain that we can make one decision or another. Is that not the exact definition of free will?
tl;dr, Your argument is as useful, and as true as F=Ma. That is to say, very useful, but completely, provably incorrect.