The Right Move vs the Rational Move
If you’re playing poker, there are two ways to evaluate performance after a hand is played:
- Did you win? Was it the right move?
- Based on the information you had, did you act in the most optimal way? Was it the rational move?
Now, by the definition of the rational move being ‘optimal’, in the long-run these two answers should be Yes or No about the same number of times.
But you should try to optimise the second method, because next time you’re playing you have a better guarantee that you’ll be correct. And you can adjust your model so that what it says is optimal is actually optimal, if you realise that your decision-making model was saying giving you the wrong answer. But you can’t really do anything if you know how many times you were right. That doesn’t tell you how to act now.
And that’s what truth is, to me. A model which describes what will happen in the world well. Not whether the answer was correct.
There are two job interviewees, and one has an umbrella and the other doesn’t, and Carl says to you “The person with the umbrella will definitely get the job because she has an umbrella”. Carl isn’t involved in the interviewing process, he doesn’t know who will get the job, he’s just telling you based on the umbrella. You are involved in the interviewing process, and you know that the umbrella candidate is going to get the job. But not because she has an umbrella, she’s just the stronger candidate. Is Carl’s statement true? Does he have knowledge?
Not really. His model is faulty. If he relies on the umbrella-test for more interviews he will be wrong and we will realise he doesn’t have knowledge. But he still doesn’t have knowledge now. He doesn’t “lose knowledge” because of the world. He just doesn’t have it.
That’s obviously a simple explanation, but it carries to other situations.
A good doctor isn’t a doctor who gets the treatment right. It’s a doctor who has a good reason to give that treatment. That’s why that’s what we look for when suing for negligence.
A good investor isn’t someone who made a lot of money, or predicted the GFC. It’s someone who articulated good reasons before the event happened about why it would happen, and rebutted reasons it wouldn’t happen.
A good vote isn’t a vote for someone who doesn’t make the world bad. It’s a vote for someone who can articulate good reasons why President A is more suitable than President B, and can rebut reasons that President A is not less suitable than President B.
Post your own examples below.