This is really bothering me and the more I read about it the less obvious it is. Can someone please explain the following taken from the whitepaper from Estimize:
"Our second strategy tests cumulative returns when there is a discrepancy between the Estimize-based and Wall Street-based surprises. In other words, we go long if actual earnings or revenues exceeds the Estimize consensus but falls short of the Wall Street consensus, and we got short if the opposite is true."
The market reacts to the surprise which is based on Street's number and actual EPS so when the Street estimates higher the price dives and vice versa. So how does the data from Estimize (even though more accurate) in the case of a total discrepancy help? (By total discrepancy I mean a time where the actual estimate is between Street's and Estimize's estimates.
The link to the paper: