Sometimes I’m talking about lifting weights and someone asks me, “What’s your goal weight?” I don’t understand why I would have a goal weight.

Say I want to bench press 300 pounds. What happens when I reach 300? I just give up on the bench press now? That would be silly. If I can keep getting stronger, I should.

What happens if I fall short of my goal? Say I haven’t been able to bench more than 285.¹ Should I start eating 5000 calories a day to put on as much muscle as possible? No, I’m not going to do that, I don’t want to get fat. Realistically, if I fall short of my goal, the answer to the question of what I should change is “nothing”.

The point of a goal is to make tradeoffs between objectives. But when you set goals, you have less information about your costs than when you’re trying to implement them. At implementation time, you have new information that might change how you prioritize things, which may result in failing to achieve a goal; and that’s perfectly fine.

Sometimes a goal turns out to be easier than you thought; that doesn’t mean you should give up after you achieve it.

Sometimes a goal turns out to be harder than you thought; that doesn’t mean you should sacrifice everything else for it.

You have an implicit utility function across multiple metrics and you want to know how to prioritize those metrics, but you do that by knowing the coefficient of each metric. Say you’re the CEO of a company, and you want to increase revenue and increase NPS and you need to allocate resources to each of those. It doesn’t make sense to say “we want to increase NPS by 0.5 points” because NPS trades off against revenue (you can make your product cheaper and worse, which increases revenue but decreases NPS), so you also need to make a statement about how much you care about revenue. It would make more sense to say “$1 million of revenue is as important as 0.1 points of NPS”. And then if it turns out there’s a way to increase NPS by 0.2 at a cost of only $500,000 of revenue, then that’s a good deal and you should take it. But a flat goal of “increase NPS by 0.1” is uninstructive.

It’s common for pension funds to target an 8% investment return. That makes no sense. Your investment return is mostly determined by the market environment, which is out of your control. There are only two ways to hit an 8% return consistently: take excessive risk when forecasted returns are low (which is even worse than accepting lower returns); or intentionally hamstring your investments when forecasted returns are high (e.g. if you expect the market to earn 8% and you think you can beat the market by 2%, then you deliberately ignore your market-beating ideas so that you can hit 8% without going over). There is no reasonable utility function to which “always get an 8% expected return” is the correct decision.

In fact, standard finance theory says you should take less risk when expected return goes down (holding volatility constant). Targeting 8% return, and taking on more risk when expected return goes down, means you are actively doing the opposite of what you ought to be doing.

However, goals are okay when reality has a discontinuity. If you need to score at least 70% on a test to pass the class, then it’s reasonable to set a goal of 70%. If you’re pretty sure you can score well over 70%, then it makes sense to stop studying. If you don’t think you’ll pass, then you should study more.

Posted on Nov 24, 2025

Notes