# Looking Under the Lamppost Because it is Brighter There Consider the idea of a drunkard who has lost their keys and keeps looking under the lamppost because that is where it is the easiest to see. The problem is that the drunkard does not have a good [Explanation](Explanations.md) of where his keys may have fallen. So he just looks where it is easiest: underneath the light. He is trying to predict where the keys are, but [Explanationless Prediction Is Impossible](Explanationless%20Prediction%20Is%20Impossible.md). Without a theory of where the keys may have fallen, he will just keep looking under the lamppost. This is a common problem in science. You have some quantity of interest that you want to measure, but measuring it directly is challenging. So you come up with some proxy that is *easy* to measure. You look under the lamppost. All [Scientific Measurements Use Chains of Proxies](Scientific%20Measurements%20Use%20Chains%20of%20Proxies.md), but these chains must be *criticized*, otherwise you may end up looking somewhere that has hardly anything to do with you quantity of interest. You may have actually left your keys on your bedside table at home in order to prevent driving, so not only are they not under the lamppost, they aren't even lost. --- Date: 20250105 Links to: Tags: References: * []()