Optimal foraging theory predicts animals maximise energy per unit time. Real animals often take a smaller, more certain meal over a larger, riskier one. Does that refute the theory?

B. It refines it, since a starving animal cannot average over trials

The currency was slightly wrong, not the method. An animal near its energy threshold is maximising survival probability, and a run of bad luck is fatal in a way the long-run average never records. Risk-sensitive foraging is what the theory looks like once you ask what the animal is actually playing for.