Football: Power Rankings, Sep. 19

One new feature we're rolling out this year is weekly football power rankings.

"How is that new?" you ask. "Everyone does power rankings." Well, instead of copying everyone else's approach — namely, ranking all teams from first to last mostly on feel — we will merely copy one particular approach.

Over at Fangraphs, their weekly power rankings (explained here) are initially based on fans' pre-season predictions, then as the season unfolds and more games are played, more weight is given to the actual results from that season.

We'll do the same for CIS football. The fans' predictions, submitted in our contest earlier, are summarized to serve as an aggregate pre-season prediction. Then we'll also track RPI and SRS every week, as we do for many other sports, and give it more weight as more games are played.

To be precise, each of the three measures (the fans, RPI and SRS) will receive equal weight at the beginning of the season, then the pre-season predictions will be worth proportionally less as each team plays more games. So if a team has played half of their schedule, the original pre-season weighting of one-third is reduced to one-sixth. By the end of the regular season, the rankings will be 50% based on RPI and 50% based on SRS: the schedule is 100% done and therefore 100% of the pre-season weight is gone.

To quote Paul Swydan at Fangraphs, "This system essentially acts as a built-in regression for current performance so that we don’t overreact to small samples of data, but it also allows for how teams have done to influence our evaluations of their abilities."

In CIS football terms, that means a Week 1 loss by Laval won't drop them all the way to the bottom, presuming the fans considered them a top-tier team before the season started.

These power rankings will be published every week, and will serve as a companion piece to the rest of our football coverage. Here are this week's rankings:

Next PostNewer Post Previous PostOlder Post Home


  1. If anything, this makes Concordia-Sherbrooke even closer than I would have thought. It is a good measuring stick for both teams (both lost to Laval after being close most of the game, both eeked out wins vs Bishop's. Wins over McGill and St F.X are a wash).

    Oh yeah, for the webcast join me and Moe Khan on SSN Canada.

  2. Should go from a top 10 to top 5. Why rank 1/3 of the teams. It makes no sense.

  3. Won't work :
    1- Starts with a country-wide rating.
    2- Continue with within conferences only ratings.
    3- Takes only Wins into consideration.
    4- Every team outside OUA would love to play Waterloo and York....

    Nuff said