When I launched this site a few months ago, I had only entered games dating back to 2005 into my model, and since then I have entered data from as far back as 1999. For my second blog post, I mused on some of the interesting trends I had notced in the data, including the weekly fluctuation of the national scoring average. I had suspected that scoring tended to be lower in the first few weeks of the year relative to the latter half of the season. Well that turns out to be true. For reference, here's how the national scoring average by week compares to the scoring average for the entire season:
During that same post, I remarked that the reader should "disregard 2006 for the outlier that it is." I have since come to realize that nationwide scoring for 2006 was an even larger outlier than I originally suspected. Here you can see the national scoring average plotted against a couple different projections I made:
There was obviously something different about scoring in 2006. My first suspicion was that a rule change of some sort was accountabl for this difference. A little digging led me to this press release from the NCAA outlining their 2006 rule changes. The major rule change in 2006 was the nationwide acceptance of video reviews, but that only affects DIII in the Stagg Bowl, so that can't be the culprit. There were, however, three rule changes with the intent of shortening the length of games, including:
Rule Change 1: Starting the clock on kickoffs when the foot touches the ball, not when the returning team touches the ball,
Rule Change 2: Shortening the kicking tee to 1 inch, likely limiting the number of touchbacks, and
Rule Change 3: Starting the clock when the ball is ready for play on change of possession, instead of on the snap.
According to kickpunt.com, the average hang time for a D2 kicker is about 3.5 seconds. Assuming D3 kickers are close to that standard, and with around 11 kickoffs per game (average number of TD's/game plus average number of FGA's/game + 2 for kickoffs at half), Rule Change 1 would shave under 40 seconds from a game on average. Forty seconds doesn't seem like it would make any difference in the overall scoring nationwide.
Rule Change 2 probably had similarly miniscule impact on the length of the game. DIII kickers supposedly average 60 yards on kickoffs, which would place the ball on the 5 yardline under 2006 rules. Even if we conservatively assume one less touchback per game, that would only shorten the total game by about 5 seconds. Including the effect from Rule Change 1, we're up to about 45 seconds total.
Rule Change 3 is where the bulk of the gametime was lost. According to "limited studies conducted by the Southeastern and Big Ten Conferences," the last rule change would likely save approximately five minutes per game. Combined with Rule Changes 1 & 2, games should have been about 5:45 shorter in 2006 than they had been previously.
If 2006 had followed the same trend as the other years included in my analysis, teams should have averaged 24.72 points per game. Assuming they maintained the same relative efficiency, the calculation for their expected scoring from changes in gametime would be:
56:30 / 60:00 = Expected Scoring / 24.72
Solving for Expected Scoring, we get 22.35 for the new national scoring average, and the actual scoring average for 2006 was... 23.34.
The rule changes instituted in 2006 were revoked following the season, which explains why there was only a one-season slump in scoring, but it also implies that offenses were actually slightly more efficient in 2006 than is to be expected from the long-term trend.