Reality is never wrong

Short form:Your model of reality can be wrong. But reality can never be wrong.

Listen carefully the next time you hear anyone (especially you) say that something “should not have happened” or “is crazy” or “makes no sense.”

These are just ways of saying that “my model is right and reality is wrong.”

You are batshit crazy if you insist that your model is right and reality is wrong.

Sorry. Nope.

Reality is the standard against which models are judged.

Your model is not the standard against which reality is judged.

H/T to Adam Robinson for that insight.

Long form (because this is important, and I’m trying to retrain myself)

If you believe the world operates in an orderly way according to the laws of physics, say—even if it’s probabilistic rather than deterministic—the world is the way it is because that’s what the laws + initial conditions say. So if the world doesn’t behave the way you expect, it’s your model of reality that’s wrong (does not include all the consequences of physical law) and if you say ‘what’s happening is crazy’ you’re saying, in effect “my model is right, and the idea of laws of physics is wrong.” That’s kind of nuts.

If you believe that the world operates in an orderly way—but that God sometimes jumps in and overrides the rules—the world still can’t be nuts. It’s the way it is because that’s the way God wants it. Either you don’t understand God, or She doesn’t know what She’s doing. Sorry, I’m betting on God. Again, in effect, you’re saying: “My model is right, and God is wrong.” Also kind of nuts.

To anyone who asserts that you can escape by deciding, “your mental model is that the world is nuts,” sorry, but that’s nuts. If “the world is nuts” is your model, does that mean that sensible things make no sense? Or does it mean that anything that happens is allowed by your model? In either case, your model is useless. Choosing a useless model is kind of nuts. Even a wrong model is sometimes right, but a useless model is—well, useless.

Any time that world seems nuts, it’s because your model of the world does not help you understand how what has actually happened should have happened.

But it always should have happened because it did happen.

The error is in your model, not the universe.

The correct response to the world behaving in unexpected ways is: “My model is wrong.”

You should, at a minimum, correct your model by changing your Bayesian priors, and at a maximum throw the model out in favor of one that predicted the world as well as your model (aggregate score, not every particular) and also predicted what actually happened that your model mispredicted.

​