The Rich Are Getting Richer
In the early days, there was Ithaca vs. Wittenberg, and after that, an unparalleled run by Augustana. After some turmoil, La Crosse and Rowan became powers, followed by the flagship, with a captain named Kehres. Mount Union reigned nearly unchallenged for a decade until other purple powers rose to challenge and--eventually--unseat them.
While words can tell an interesting story about the "elite" teams in Division III, numbers have their own eloquence. Right now, my data only goes back to 1999, so I can't tell you exactly how dominant those Augustana teams of the 80's were, but I can definitely tell you that Mount Union was in a class of their own for much of the first decade of this century.
No team has won a national championship since 1999 without having a rating of greater than 0.995 in my model. In my Methodology page, I say that teams who pass this threshold are legitimate title contenders. In the early 00's, Mount Union was typically the only team to maintain such a rating for most of a season, with other teams usually getting on a roll late in the season to surpass that mark.
Something interesting has been happening the last few years though, more teams are achieving ratings over 0.995, and then maintaining them.
It's pretty easy to figure out which teams were the ones to pass the 0.995 mark each season.
-- 1999-2003 was pretty much exclusively Mount Union & someone else.
-- 2004-2005 was Mount Union, and then equal part Linfield, Mary Hardin-Baylor, and the UW-Whitewater
-- 2006-2011 was UMU and UWW, and the typically one or two other teams
-- 2012-now has been basically every national semifinalist and a few others
This preseason featured seven teams with a 0.995+ rating: Mount Union, UW-Whitewater, St. Thomas, Mary Hardin-Baylor, Linfield, UW-Oshkosh, and St. John's. After their shutout win in Week 1, UW-Platteville joined the group. From the looks of things, all of these teams will likely maintain their stellar ratings for most of the season.
Are the best teams in the country actually getting better, or is something else the cause of the influx at the top?
My ratings are based on a team's opponent-adjusted point spread. The rating is each team's win probability against an average (0.500) team, which itself is a calculation based on the model's long-term standard error in predicting point spreads. A couple things could affect these ratings to explain why we're seeing more elite teams than ever: an influx of startup programs or an increased rate of blowout wins.
Startup programs are all given bad ratings (~0.030). If a lot of startup programs start flooding DIII, it would drag down the average team, and prop up the best teams. To correct for this, I have a yearly adjustment to each team's offensive rating of 0.05 points/game, which keeps the national average at 0.500 (without the adjustment, the average by now would be around 0.475). So that's probably not the problem.
Blowout wins seems like a likely culprit. If teams are more willing to run up the score now compared to ten years ago, it would definitely be a cause for these ratings to be inflated. Teams are definitely scoring more in general, and with high-tempo offenses being en vogue, it seems likely that blowout wins are becoming more common.
It could be possible for the average margin of victory to remain relatively constant, as we see above, while the frequency of blowouts increases if the number of blowouts was offset by an increase of close games. That's why I also included the standard deviation. If the number of blowouts was truly increasing we would expect to see one of these values on the rise, or both.
What's the real cause? I don't know. Maybe the "arms race" for facilities has trickled down enough into the DIII landscape such that rising tuition costs are sending more and more of the best recruits to the only schools who can afford better facilities. Maybe it's an anomaly. Regardless, I think we should just enjoy the season, playoff chase, and national tournament ahead. In my opinion, more teams playing football at an elite level is a good thing.