Well, I imagine most people are going to talk about free agency and Brett Favre this week. So here’s something different. I’ll get to free agency next week, when a few more deals have been worked out and I can analyze winners and losers a bit more thoroughly. As for Favre, I’ll believe it when the season ends and Favre hasn’t played a snap. Vinny Testaverde came back last year, folks, retirement isn’t as permanent as it used to be. All right, that’s all on that topic. On to an interesting study I just did.
I noticed that this year’s playoffs included several teams that were in the 2006 postseason, but not in 2007’s, namely Washington, Tampa Bay, Jacksonville, and Pittsburgh. And several teams, such as New Orleans, NY Jets, Baltimore, and Philadelphia, had terrible seasons in 2005 and 2007 sandwiching a playoff year. I wondered if this was a recent trend, or even a trend at all – perhaps they were just noticeable because they involved such dynamic changes but it was part of the NFL averaging itself out.
So here was the hypothesis of my study: a team that had a major improvement or a major dropoff from the previous year will tend to revert to its form of two years ago. For example, the Redskins’ total wins from the past four seasons, in order of most recent to least recent: 9, 5, 10, 6. Down, up, down, and significantly in each case. The Jets are a textbook case over the same period: 4, 10, 4, 10. Six games better, six games worse, six games better. Basically, I was looking to prove that this happens all the time.
Maybe I’m just a big dork, but I found the results staggering and fascinating.
149 times from 1992 to 2007, a team’s final record was different from that of the previous year by 3 or more.
Teams continued on the same trajectory to the tune of 4 or more wins 6 times. Example: The Dolphins from 2005-07 went from 9 wins to 6 wins to 1 win. Same trajectory, minus 3, then minus 5.
Same trajectory, but only by 2-3 wins: 12 times, and one 3.5 case where a tie was involved. Example: The Panthers from 1996-98: 12 wins, 7 wins, 4 wins.
Holding steady, with a change of plus or minus only one win, or no change at all: 33 cases. Example: Seattle from 2005-07: 13, 9, 10.
Reverse trajectory, only by 2-3 wins: 45 times, including two 2.5 cases where ties were involved. Example: Dallas from 2003-05: 10, 6, 9.
Reverse trajectory, by 4 wins or more; 52 times!
Here’s the same information, in perhaps an easier-to-read form (assume negative numbers mean reverse trajectory):
4+: 6 times, or 4%
2 thru 3.5: 13 times, or 9%
-1, 0, or +1: 33 times, or 22%
-2 thru -3: 45 times, or 30%
-4: 52 times, or 35%
What does this mean? Well, it means that NFL win totals look more like random number generators than you might expect. If you generated random numbers from 0-16 over and over again, and recorded every time the number was different from the previously generated number by 3 or more, the most probable outcome would be that the following number would be different from its predecessor by 3 or more in the opposite direction. Anyone who’s actually taken a probability class want to back me up on this? It makes logical sense.
Great, says Curly or some other wiseacre. You’ve told us nothing then. Win records tend to follow probability rules.
Ah, but we’re not talking about random numbers. That’s what makes this interesting. We’re talking about football teams, with real human beings suiting up and playing games dependent on skill and very little random chance, unless the game goes into overtime and there’s a coin toss. Conventional wisdom says that a team, say the Browns, who vastly improved this past season, will build on that improvement. But my statistics say otherwise. According to the trends of the years since 1992, Cleveland is 65% likely to finish at 8-8 or worse.
As the offseason continues, soon we’ll get into predictions season, where everyone who writes about the NFL for fun or for profit will share their picks for each team’s record. You’ll find they look very similar to 2007’s final rankings, with a boost to the writer’s home team and whoever made the biggest splash in free agency (last year it was the 49ers winning the NFC West for a lot of people). But if I’m right, things will end up looking a lot more like 2006. Great news if you’re a Ravens, Jets, Chiefs, or Rams fan. Not so hot if you dig the Packers, Browns, Bucs, or (like me) Redskins.