For Good Measure: Betting on Growth versus Magnitude

Whether you are Werner Heisenberg or Janet Yellen, your field of study includes measurements of position and velocity and how they interact. Poor measurements may be unable to nail down either the one or the other (much less both), but even if only the one is measured, there is likely to be some prediction that you will be able to make. In hard-to-measure situations, consistency of error distribution can be your friend—consistent errors help you to find the message in the body of noisy data. The reader probably knows all that.