The amount of time you spend considering the data shouldn't depend primarily on the magnitude of the decision. It should depend on your beliefs about how much a

The amount of time you spend considering the data shouldn't depend primarily on the magnitude of the decision. It should depend on your beliefs about how much another marginal unit of delay will cost you, vs. how much another marginal unit of considering data will benefit you. For instance, if you're wondering which new ballpoint pen to buy, that's trivial next to strategic decisions in war. But it still might be worth comparing reviews on Amazon.com to increase your chances of getting something ergonomic which won't annoy you in some way. Likewise, if you're Stanislav Petrov, you may have to make the most important decision in the history of mankind based on incredibly limited information, but with a large delay-dependent penalty if you just go with the flow until you can evaluate a sufficient amount of information. Contrariwise, if you slip while climbing a tree, you don't stop to work on a coherent theory of quantum gravity even if you're only a few feet off the ground, however more accurate that might make your actions; you just grab in the general direction of a branch.