The Problem with Stats and NFL football

Since everybody seems to be really harping on the stats to shed light on this matchup, I must caution you to not put too much stock in the stats.  Yeah, I'm a turd in the punchbowl like that.

1. There are only 16 games a season. This is WAY too small a sample size to generate statistical power. Without adequate power, you have no true predictive value.

2. It is not possible to have a reasonably controlled study. There are so many confounding variables to account for in the NFL.

3. When you try and increase sample size by looking over say 10 seasons, the confounds are even greater and they vary from player to player, team to team, game to game, season to season. How a team fared even one year before has almost no bearing on the present year.

4. They margin of difference in talent between the best team and the worst team is so slim in professional football that it takes even greater power in a study to be able to show a statistically significant difference.

5. Because it is a team game, individual stats are even less meaningful when you consider every outcome is dependent on the performance of 21 other people. As far as teams go when you rank stats, there are only 32 teams. Just because one team is ranked first and another team is ranked 20th in a particular category doesn't mean there is a significant difference between the performance of the 2 teams. We could be talking about a difference less than 25 yards rushing per game between the 1st and 20th ranked team.

6. If stats were reliable we wouldn't bother playing the games. We would just run computer simulations. We would always win our bets.

So what do all the stats mean? Not much. Or said another way, the stats mean something, but we just don't have enough of them to figure out what it is that they mean. At best, stats can be used to suggest a weak correlation between teams/players and a particular aspect of football, but not a statistically significant difference, thus they are not predictive. So when looking at matchups, remember there is not much difference between NFL teams in most performance categories, especially when comparing the two remaining teams playing in the Superbowl (is Peyton really that much better than Drew? Does anybody think that it is going to be the difference maker between these two teams?). That's the reason the two stats that correlate strongest (although not significantly) to winning during the season are time of possession and turnovers: The team with the most chances to score usually wins when you consider that every team is so evenly matched. Perhaps the most useful way to use stats in NFL football is to just compare the one thing each team does best and worst, and then see if they play into the other team's strength and weakness.  For example if team A has the #1 running game in the NFL and team B has the #32 ranked run D in the NFL, you might be able to suggest Team A matches up favorably with Team B.

So everybody knock yourselves out arguing the subtlest of statistical nuances between teams, if you like. At least it gives us something to talk about. I'll be floating in the punchbowl, following along skeptically.

This FanPost was written by a reader and member of Canal Street Chronicles. It does not necessarily reflect the views of CSC and its staff or editors.

Log In Sign Up

Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Canal Street Chronicles

You must be a member of Canal Street Chronicles to participate.

We have our own Community Guidelines at Canal Street Chronicles. You should read them.

Join Canal Street Chronicles

You must be a member of Canal Street Chronicles to participate.

We have our own Community Guidelines at Canal Street Chronicles. You should read them.




Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.