In the context of decision theory, utility means something like subjective value. Utility is an abstract measure by which you express relative preferences numerically. It's the anchor that consistently attaches your gut preferences to numbers, similar to how numeric probabilities anchor your feelings about likelihood. Just like you can use probabilities to express complex expectations, you can use utility to communicate nuanced preferences.
The easiest heuristic is that probabilities represent beliefs and utilities represent preferences.
Let's go through an elementary example.
I am ordering breakfast at a restaurant. I would prefer pancakes to a croissant, and a croissant to crepes. In fact, I hate crepes. Then my utility valuations would be as follows:
- Pancakes: 100
- Croissant: 70
- Crepes: 0
Now we can use the tools of Expected Value and probability trees to think about complex and uncertain situations involving breakfast. For example, if I'm choosing between a cafe that may not have pancakes but probably will have croissants, and a diner that might have both.
You might ask, what if you don't know the relative utility of various outcomes? What if you're uncertain about your preferences and the likelihood of outcomes?
This often happens when weighing different outcomes such as love, money, or prestige. For now, don't worry about it. Subsequent articles will introduce tools to break down uncertain preferences and properly value distinct outcomes.
What Utility Isn't
Because you may be coming to this course with some prior understanding of utility based in some other context, let's dispense with what decision-theoretic utility isn't:
Utility isn't well-being. It's not wealth, and it's not happiness. A close synonym would be "expected value", with the caveat that utility refers to a subjective, first-person value. You can't trade utility between people — in fact, you can't even compare two people's utility. It would be like asking, "how many kilograms are in an inch."
Finally, utility can't be accumulated and you don't "get" utility when you get the desired outcome. What you get, obviously, is the outcome itself.
Why Use Utility?
You might be wondering why we need a concept such as this. There are two reasons: consistency and clarity.
Utility Imposes Consistency
Correctly using utility creates consistency in your decision-making. One major confusion in human decision-making is inconsistent preferences.
Example. Inconsistent preferences can occasionally trap you in negative cycles. For example, suppose that you prefer:
- Red to ble
- Green to red
- Blue to green
Then you will always be chasing the next color. This isn't a big deal if changing colors is costless, but in the real world there's typically a transaction cost, and irrational circles like this will drain your wallet without you realizing it. In gambling and betting, this is known as dutch booking.
Utility Creates Clarity
The second major reason to use utility is that it creates clarity. When considering complex life problems, there are almost always two distinct types of confusion:
- Uncertainty about your prediction of what will happen
- Uncertainty about what you want
Once you separate these types of uncertainty, your decisiveness will naturally increase. Putting consistency of subjective preference at the center of your decision-making framework sidestep many cognitive traps that smart people get stuck in. Forget the complex philosophy — a good decision is one that best suits your consistent preferences.