NFL Playoff Observations

Written by

Published on January 18, 2010

The Saints picked up right where they left off - in November - blowing out the Cardinals as convincingly as they blew out the Patriots a month and a half ago. Their struggles against Atlanta, Washington, Tampa and Dallas seem like preseason games now. How strange that the game started with a 75-yard touchdown run by Tim Hightower of all players.

The Ravens had their chances to be competitive with the Colts, but made too many mistakes. Early on, Cam Cameron was calling a good game, with plenty of throws to Derrick Mason and a well-executed reverse to Mark Clayton. But once Baltimore got behind, it looked out of options - like it did a few years ago against the Colts when Steve McNair was their quarterback.

Like the Saints, the Vikings were in midseason form, getting to the quarterback and passing the ball with efficiency. The Cowboys were overhyped thanks to catching some teams at the right time (the Saints during a lull and the Eagles with major offensive line problems), and for the last six weeks Tony Romo got to play in what Scott Pianowski calls "hitter's counts." Not so Sunday when he had to play on the road from two touchdowns down with the entire Vikings defensive line in his face. Every quarterback is at the mercy of the conditions around him, though how each handles progressively more difficult ones is what separates the good from the great. I'm not sure even Peyton Manning would have had success in that spot, though.

One subplot with the Vikings is how ineffective Adrian Peterson's been for much of the season. I realize teams are probably still stacking the box, but that never seemed to stop Chris Johnson or even Peterson in years past.

Perhaps a little known side effect of the swine flu is that it destroys one's ability to kick accurately. How else can one explain the current shanking pandemic going around? Then again it's not the first time Nate Kaeding cost the Chargers a playoff win against the Jets... But to put the loss on Kaeding is to oversimplify things as Norv Turner's constant punting on fourth and short near midfield in the first half kept the Jets in the game, and New York's defense dominated in the second half. Turner's decision to go for the onside kick at the end of the game was foolish as it gave the Jets the option of going for it on 4th and 1 or kicking a game-sealing FG. In their own end, the Jets almost certainly would have had to punt, giving the Chargers more than a minute to drive down for a FG that Kaeding would no doubt have missed.

Of all the bye-week teams, only the Chargers weren't as sharp as they were during their regular-season peak, and that has to fall on Turner.

One might conclude from these games that momentum means nothing (h/t Pete Schoenke), as:

NFL's Final Four went a combined 4-9 over the final three weeks of the regular season, with two of those wins coming against teams not trying, and one against a team that had quit (Vikes over Giants). The losers this weekend went 10-2 in their last three regular season games, plus three of them won an additional game last week.

But as Pianowski wrote in our Live Blog:

Momentum is overrated, except when it isn't. Defense wins championships, except when it doesn't. It's a society of backfitting.

Sometime in the next five years, momentum will surely matter more than ever. The lesson isn't to take what happened this year and project it forward indefinitely - it's that you must discern when momentum matters and when it doesn't. I got a lot of grief for picking the Vikings (the only game I got right) because Dallas was supposedly playing so well, and I've been an advocate of emphasizing recent play. But a four or five-week sample is small, so one must make sure the results during it are convincing. Dallas' were suspect for the reasons mentioned above. This is more of an art than a science - how to know whether a short-term surge by a team is indicative of a new baseline or just a combination of good luck and temporary good focus.

There are two variables when tackling this question: (1) the size of the sample (the more games the better) and (2) the magnitude of the outlier. When Joe Mauer had 11 homers in a month after hitting 16 in two years, the sample was small, but the magnitude was very large. Could we really have said that about the Cowboys? Were their wins over the Eagles and in New Orleans that convincing, given the Saints' loss to Tampa the following week and their poor play against Washington and Atlanta the previous weeks? Did anyone really think that some team like the Steelers, Ravens or Texans wouldn't have given New Orleans a ton of problems if they played in that same spot?

I do put emphasis on recent results because teams frequently move their baselines over the course of an NFL season, and once they do, you don't want to taint your analysis with games that happened before the move. But recent results can also just be random fluctuation around an existing baseline, and you don't want to mistake that kind of movement for a real shift. Again, deciding whether the movement is random or a real shift if more of an art than a science, but it makes all the difference.

One other takeaway in all of this is not to strive for some simple-minded consistency. That I should back Dallas because of recent performance in order to be consistent with other times I've emphasized that is preposterous. Everything depends on the circumstances, and you should typically be suspicious of anyone who says "always" do x, or "never" do y. We have to find reasons to support our picks, and those reasons will necessarily include mentions of recent performance, league-wide historical data, fluctuations in player performance, etc. We must hang our hat on |STAR|something|STAR| if we're going to communicate the reasons for our picks. But don't extrapolate from that reasoning that it should apply in every circumstance. Deciding which circumstances to which one set of rules applies and which require perhaps the opposite set is the challenge.