Correlations of NFL statistics to Wins

So the most simple "net points per drive" comes out on top, no surprise given that raw net points correlates well with team wins, and indeed the basic drive chart measures such as Touchdown%, yards per drive, and yards per play are all solid.

The play-by-play findings tell you one thing pretty clearly: it's a passing game, and the way to win is to pass more effectively than your opposition, whether that is measured by overall effectiveness (our EFR number), the passing play success rate, or the passing big gain plays (defined as 20+ yards for passes, and remember we include such things as pass interference plays which often go for 20 or more yards). Now Bud Goode, among others, has of course for years been talking about the "killer stat" of net passing yards per play.

Then we find a couple of down-by-down stats rounding out the list: namely net play success rate on 3rd down and 2nd down respectively.

You might be wondering if regular season success translates reasonably well to the ultimate goal -- taking home the Super Bowl trophy. Here's a look at the past six NFL champions and how they stack up in the key correlators:

Some interesting points emerge -- Touchdown Percentage has been the greatest single predictor of the high correlators, with all of the teams in the top four on the season. On the other hand, the passing net big gain percentage has been less stellar, suggesting that while this stat may be important for accruing regular season wins, a team that depends too heavily on the big play may be sketchy come the post-season. Finally, net 2nd down play success rate has seen four of the six prior winners being top ranked -- probably about the best top rank pick success of any stat you could find anywhere -- who says third down is the key down?

Indeed, in looking into the 2nd down stats further, in the two years the winner did not come from the top ranked team (and of course there are those, us among them, who would say that the Patriots were beaten by the Raiders in that infamous 'tuck' playoff game) in 2001 the top ranked 2nd down PSR team was the runner-up St. Louis Rams, while the Denver Broncos were top rated in 2003 but fell brutally to the Colts in the opening round. Maybe it's just a quirk, but we'll give an update on the #1 2nd Down PSR team at the end of week seventeen.

Of course, one of the great things about the NFL is that the game evolves from year to year -- the strategies employed successfully in the prior season may now have been coutered, new wrinkles and rule changes may have come into the game. So as a final look we'll run the correlations for the 2003 season alone against our top ten overall stats, recognizing that with a much smaller sample size, the 95% CI ranges are much broader:

And for those of you who can't get enough numbers to review, the following table shows the correlations for the six year span for all the numbers featured in our three main stat classes -- drives, play-by-play, and down-by-down: