I began using the JRI to rank teams and pick games after Week 5 of the 2008 NFL season. I felt that a reasonably large sample size of games was necessary for the formula to be useful; particularly, eliminating the first week or two would allow teams to shake off the rust of the preseason. This season, I have elected to use the JRI from the earliest possible point in order to test whether a pre-existing average performance for teams improves the accuracy of picks.
It should be noted that the JRI ranking and prediction methodologies were not finalized until several weeks after implementation. Picks for the 2009 season should more accurately reflect the JRI's current formulation.
Overall, JRI's picks were fairly accurate over the last 12 weeks of the 2008 season. JRI correctly picked the winner in 117 out of a total 183 games for a respectable 63.93% success rate. However, a closer examination of the index for these picks results in an interesting discovery.
For picks with less than 2.00 absolute index, JRI's record was 64-56 (53.33%). For the remaining picks (locks), JRI's record was a lofty 53-11 (82.81%). I believe this rather conclusively demonstrates that the team with a significant statistical advantage will likely win games. More to the point, it demonstrates the truth of football folk wisdom otherwise: any given Sunday.
