The National Football League structures competition in such a way that it's difficult for any one team to remain better than the others for any length of time.  There is, first, attrition of players, broken down by age and sex and the length of the season.  Then there's the inverse-order draft, giving the current poor performers more favorable access to the reserve army of aspirants.  The salary cap constrains spending on high performers, and high-performing veterans can hawk their services on the free-agency market.  In addition, schedules going forward are based on current performance.  Under the current configuration of the league, each team plays six games with divisional opponents each year, four games with the teams of another conference division, and four games with the teams of a division from the other conference.  These divisions rotate predictably.  Two remaining games adjust for performance: a team that wins its division plays all three division winners within its conference.  The rationale is similar to that of a Swiss-system chess tournament, in which successful players face successively stronger and comparably successful opponents.

But Peter Ingemi suggests that the schedule strengthens the already strong teams.
If a team finishes fourth, its remaining two games will be against two fourth place teams.  If they finish second, the two remaining games will be against two second place teams, and if you finish first you are guaranteed to play every division winner in your conference.

And it is that little difference which gives these teams the edge they need.

Rather than being a disadvantage, competing with the best of the best raises their game.  It forces them to get better, to play harder and smarter.  It gives these already elite players that final incentive to raise their game to the level of a champion and allow them to do it year after year.
He's getting some push-back in the comments, and strong objections to the analogy he draws to school choice.  His idea, however, might not be completely wrong.  There's one more dimension to consider, specifically the additional practice time the playoff teams get.  (That may be why there are enough bowl games for all the six-win college teams in the country).

Thus, the division winners earn two weeks to a month of extra practice time, plus they learn something about how their opponents for next season will play, and then have to play those higher-performing opponents again.  And that old bugaboo of vulgar statisticians, disproportionate representation, appears.  According to Mr Ingemi, this year's football final four featured teams that accounted for 40% of the Super Bowl appearances in the past ten years.  For the vulgar statistician: ten years, 20 appearances, 4/20 "warrants" 20%.

And who do those Green Bay Packer fans think they are, lamenting a blown trip to the Super Bowl?  To the vulgar statistician, that's just crying with your mouth full.
Green Bay has advanced to the postseason 17 of the last 23 years (73.9%), including six of seven under Rodgers (85.7%). The Packers have 11 division titles in that time, including four straight.

Green Bay has been to six NFC championship games in that time, going 3-3 in those contests. And the Packers have reached three Super Bowls, going 2-1 in those games.
That performance is characteristic of a power rule at work, rather than a Gaussian process at work.  That power rule, however, is not good enough for the fans and the pundits.
Many organizations would trade places with the Packers in a heartbeat, and be downright giddy with two world championships since Favre became the starter in 1992. But for a team that's enjoyed legendary play at the game's most critical spot, are two titles enough?
That's not the relevant question.  It has been a difficult task for Mike Holmgren and Ray Rhodes and Mike Sherman and Mike McCarthy to do their work with the footsteps of giants behind them.

No comments: