This is just silly. What kind of micro-level change are they predicting that has a buck of upside? This doesn't imply anything insightful - it more implies that they have an overly complex model that spits out obfuscated and meaningless data, and that they get paid to periodically update and release a new number, regardless of outcome or effectiveness.