A data scientist walks into a bar and observes a large crowd cheering intermittently. The crowd’s eyes track events on a large TV screen, following a bouncing spherical projectile’s motion as fit actors throw it through one of two metal rings. One of these rings elicits cheers from the crowd as the object passes through, the other groans.
Puzzled, the data scientist wonders what they are drinking—and promptly orders it.
Several drinks later, our hero smells opportunity: If the Dallas Mavericks win this third game of the NBA Finals, the anticipation of a possible Finals victory for the team may boost the stock price of any Dallas-based party-consumables company. For certainly these companies’ marketers will run tactical ads to increase sales as celebrations reach fever pitch, something along the lines of:
“Our town, our team, our [product], our time!”
The data scientist goes home to plan stock trades.
Unfortunately no large publicly-traded beer manufacturers, the ideal financial instrument for such speculation, call Dallas home. The best our hero comes up with on short notice is the Dr. Pepper Snapple Group. By then the NYSE had closed for the week and the Finals finished early the following Sunday, so the data scientist placed no trades.
Undaunted, the data scientist begins post hoc analysis to evaluate the theory. The Dr. Pepper Snapple Group’s (DPS) closing price, normalized against competitors’ prices, during the week of the Finals suggests the presence of a “victory anticipation bump”:
Perhaps our hero is on to something. But is the price change significant enough to warrant investigation of causality?
Well… no. This price jump (third bar from the right on the graph below) falls within the 95% confidence interval for daily price jumps for all trading days of the NBA season; not enough an outlier to warrant further analysis. Similar results appear when DPS is normalized by Coca-Cola, and when price jumps spanning two or three days are considered.
The data scientist returns to the drawing board.