Wizard's Super Bowl Experiment

After reading the Wizard's very interesting newsletter article, "Super Bowl Experiment", I was left wondering, what would the results have been had one wagered $20 on all the "yes" or "over" in the descriptions. And to clarify, only those wagers with a corresponding 'no' and 'under' side.

The quick thought would be, since this year, the no/under returned a loss of 12.1%, the yes/over would have won. Is that true? And, if so, how much? Given there is vigorish(?) built in, I would not be surprised if taking the opposite approach would have won.

Are the numbers available to make these calculations?

Thanks again for documenting this for us to read.

South Point had a sheet, but the site took it down after the Super Bowl.

However, most of the Over/Under props were -110 each way.