For all the effort that’s gone into developing sophisticated statistical measurements of football, it remains a highly unpredictable sport. As my buddy Chase Stuart once wrote about the NFL, “we don’t know anything and we never will.” And yet, while we may not know anything for certain, we’ve learned enough that from week to week, we can make sense of some of the chaos (though not all).
With that in mind, let’s take a look at what transpired over wild-card weekend. How much did it differ from what the advanced stats would have predicted before the game? Some outcomes were easy to see coming; others illustrated just how little we can predict about a single NFL game.
What the stats saw coming
- The Chiefs ran the ball all over the Texans. KC came into its game against Houston with the league’s top rushing attack according to Football Outsiders’ DVOA metric, and the Texans boasted a decent but not great rushing D during the regular season. So it shouldn’t have come as much of a surprise that the Chiefs ran for 141 yards during their 30-0 annihilation of the Texans.
- KC’s passing game was good, not great. As well as Kansas City played in other phases of the game, its passing attack was not the most crucial element of its win. Alex Smith averaged 5.9 adjusted net yards per attempt against a team that allowed 5.4 during the regular season, so the Chiefs pretty much passed to expectations, despite the lopsided win.
- The Packers’ ground game delivered against Washington. For all the pregame chatter about Green Bay’s offensive struggles of late, the Pack ranked 10th in rushing efficiency during the regular season, closing the year with 100 or more yards in three of its final four games. Helping matters, Washington was the 11th-worst rushing D in football by DVOA. Sure enough, after 141 yards and a pair of touchdowns on the ground, Green Bay had run its way into the divisional round.
- Cincinnati’s special teams played well. The Bengals lost in excruciating fashion, but you can’t blame the special teams, which outplayed their Steeler counterparts by 1.2 expected points added (EPA) in the game. During the regular season, Cincy ranked ninth in special teams DVOA while Pittsburgh was dead-average — so in at least one regard, the game played out exactly as expected.
The biggest surprises
- Houston’s passing was horrific. The Texans ranked 22nd in passing DVOA during the regular season and the Chiefs had the NFL’s fifth-best defense against the pass, so this matchup looked lopsided before the opening toss. But Houston’s quarterback was Brian Hoyer, who had more passing success than the other three QBs the team used during the regular season. The hope was that the Texans would outplay their regular-season numbers; instead, Hoyer had the fourth-worst passing game in postseason history, an outcome no metric could have predicted.
- Minnesota’s run defense stuffed Seattle. The Seahawks get a little leeway here because they were playing on the road, in frigid conditions, with neither Thomas Rawls nor Marshawn Lynch. But they also had the NFL’s fourth-best rushing DVOA during the regular season — with a lot of the way paved by their offensive line — and they were facing a Vikings team that ranked 18th in rush defense. So it was extremely unlikely that they’d be held to 3.5 yards per carry and -5.9 expected points on the ground Sunday.
- The Chiefs destroyed Houston on special teams. Special teams play is notoriously difficult to predict — and that’s at the season level, let alone in a single game. So although the league-worst Texans specialists were facing the seventh-ranked Chiefs, that should have granted only a razor-thin edge to KC. Instead, Kansas City’s special teams were worth almost a full touchdown by EPA on Saturday, giving KC the 18th-best playoff special teams performance of the past decade.
- Cincinnati’s passing game struggled badly. This comes with an injury-related asterisk as well: Cincinnati had the league’s best passing offense during the regular season, but most of that was done before quarterback Andy Dalton was injured. However, backup QB AJ McCarron had been doing a reasonable impersonation of Dalton down the regular-season stretch, and on Saturday, he was going up against an average Steeler pass D. If McCarron hadn’t helped Cincinnati post the 34th-worst playoff passing game of the past decade by EPA, Cincinnati’s defense wouldn’t have been put in a position to hold a 1-point lead on the game’s fateful final drive.
One final note: These unlikely performances are also the most valuable. Of the 10 cases this weekend where a team added 5 or more expected points in a single phase of the game, all had less than a 30 percent probability of happening based on the teams’ regular-season numbers. Eight had a 15 percent chance or less of occurring; four had a 10 percent probability or less. Some of this can be attributed to randomness and game-to-game volatility, and some is due to individual matchups and planning.
In other words, the performances that fuel victory are often also the toughest to see coming. And with the playoff field’s Super Bowl odds becoming more tightly bunched than ever this weekend, don’t expect that to change anytime soon.