There is a common cliche related to performance analysis: "You're only as good as your last success." (Or last sale, game, deal, etc.). The idea is that we tend to judge based only on someone's immediate past performance. If you're on, you're great, but if you have a bad outcome, you're washed up. Chapter 4 of Superforecasting makes the point that even superstars are not infallible, thus, we shouldn't judge based only on immediate past performance. How do we develop a way of assessing forecasting performance that is responsive to results, but not completely driven by accurately forecasting a particular event? (As discussed in chapter 3, this is not as easy as it sounds. Given typical cognitive biases, we tend to judge as 'wrong' the weather forecaster who tells us that there is a 70% chance of rain if it doesn't rain, while this is not necessarily evidence that the forecast was incorrect). As forecasters how do we avoid being judged based only on our last forecast's prediction?