The human brain is genuinely awful at measuring its own utility function. Read no further than the sequences in order to see just how bad people generally are at making economically optimal choices. When the choices compare two very similar benefits, with magnitudes governed by probability- say, a small chance of a lot of money, or a large chance of a small amount of money- one need only multiply (with some log scaling for sufficiently large sums of cash) to calculate the optimal decision. However, what is one to do when the choices are incredibly different in nature?

One such choice that I deal with frequently is when the dining halls at my university serve crêpes in the morning. Most of the time, when I eat a crêpe, I feel a short headache afterwards (presumably because of the sugar content). Is this headache to me worth the crêpe?

This is not a hard question for me to answer, but it is a hard question for me to answer correctly. After all, crêpes and headaches are very different things, and the choice is not for one or the other, but rather for both or neither. Finally, the things I am bargaining with happen at different times- the crêpe is in the present (relative temporally to the choice whether to eat it), and the headache is in the near future. 

The solution I have found to this problem is, in the case of the crêpe, quite elegant. We know that utility is of the same magnitude whether it is for the gain of a thing, or for the loss of that thing (assuming the thing is sufficiently small). That is, if I value a slice of cake at $3, then it should be true that I am both willing to pay $3 to buy a slice of cake, but also willing to accept $3 to give up a slice of cake I already have. 

The solution is then to compare the headache to the lack-of-crêpe-ness conferred by not eating the crêpe. If you set eating the crêpe as a baseline, and then imagine choosing between two worlds- one where you lose the crepe, and instead are presented with an alternative (say, a bagel)- and another where you are instantly afflicted with a headache. 

Instead now of having to measure a crêpe against a headache, measuring the lack-of-crêpe against the headache is much easier. Lack-of-crêpe and headache are much more similar states, in fact- both are to me mild discomforts, short and temporary in nature. And so when I picture them next to each other, it is much easier to see the subtle difference between them and pick the choice (lack-of-crêpe) that is less unpleasant.