But if your utility function is bounded, as it apparently should be then you’re one affine transform away from being able to use geometric rationality, no?
How much should you shift things by? The geometric argmax will depend on the additive constant.
I don’t think so, primarily since geometric rationality isn’t VNM rational, so we can include infinite utility functions.
But if your utility function is bounded, as it apparently should be then you’re one affine transform away from being able to use geometric rationality, no?
How much should you shift things by? The geometric argmax will depend on the additive constant.
I don’t think so, primarily since geometric rationality isn’t VNM rational, so we can include infinite utility functions.