Most Read: Sports

http://www.washingtonpost.com/2010/07/06/ABMK8PP_linkset.html
Posted at 07:00 AM ET, 09/14/2011

Statistical analysis: Rex Grossman, deceptively large margins of victory and other notes


The Redskins’ victory over the Giants is reason for optimism around Washington. If you thought they were probably a 7-9 team, you might think they’re an 8-8 team or conceivably a 9-7 team after a complete win over a division rival. But Redskins fans have seen this movie before, and it doesn’t always end well.

I’ll be covering the Redskins from a statistical perspective again this year. Fair warning: The stats I use aren’t the trivial, conventional football numbers, arbitrary benchmarks, fantasy-oriented stats like total yards or touchdowns that give statistics a bad name. I’m interested in the numbers that really matter. My stats literally start with winning and work backward from there. Everyone uses statistics. Everyone. So you might as well use the good ones.

I want to know what makes a team a winner — which positions and which players impact the game most? Does defense really win championships? When should teams really go for it on fourth down? We can conjecture and debate with our coworkers around the water cooler all day long, but with statistics, we can answer questions like these with a fair degree of certainty. The techniques I use are no different from what the military uses to win in combat or what researchers use to test cures for diseases. It’s just science, and like it or not, science works.

After the jump, you’ll find some notes from Week 1.

Rex Grossman had a solid game against a depleted defense. In terms of Win Probability Added (WPA) — the total effect a player makes on his team’s chances of winning — Grossman has had nearly as many positive games and negative games in his career. But when he’s bad, he’s awful, a tendency that makes his total career numbers abysmal. He’s appeared in a total of 45 regular and postseason games, and actually has played better than he did Sunday in seven of them in terms of WPA, and in five of them in terms of Expected Points Added (EPA) — the total effect on a team’s net score difference. Sunday’s performance was a +0.19 WPA and +10.7 EPA. The point is that Grossman has played this well on previous occasions, yet still went on to be a heavy drag on his team. On the brighter side, Grossman was tenth in the league in Week 1 with a very solid 6.8 Adjusted Net Yards Per Attempt (AYPA) — yards per attempt adjusted for interceptions and sacks. (The league average is around 5.0 AYPA). The season is far too young for any conclusions, but pay attention to this stat in coming weeks, as it’s the single best predictor of team success.

Sunday’s margin of victory was deceptively large. The 14-point edge came in large part thanks to some good fortune. The major turning point in the game was rookie Ryan Kerrigan’s tipped/interception/touchdown return. It was an excellent play, and it might be a sign of more to come, but it’s not the kind of thing a team can count on week-in and week-out. The fourth-quarter strip-sack could have been a game-changing play had the subsequent return not been stopped short of a touchdown. Then the Redskins’ defense held, and blocked a field-goal attempt to preserve the 7-point lead. Later, Grossman made a bad read, resulting in a dropped near-interception.

Perhaps most critical was a gift from the Giants in the form an unnecessary roughness penalty that kept the game-clinching touchdown drive alive. On 3rd down and 9 from Washington’s 45, Grossman connected with Fred Davis for a 7-yard gain, forcing a likely punt, and giving the Giants a late-game opportunity to tie. Antrel Rolle’s penalty gave the Redskins a first down at the Giants 33. It’s not something you’ll see on the highlight reel, but that single penalty was worth a swing of 0.07 WPA, one of the largest plays of the day. I don’t mean to play woulda-coulda-shoulda here. Rather, it’s important to note just how easily any single game can swing the other way.

The offensive line’s performance is cause for concern. Of the 26 non-penalized runs Sunday, 14 of them were for 2 yards or less (excluding the goal line touchdown run). The Redskins’ run Success Rate (SR) — the proportion of plays that result in making a first down conversion or score more likely — was in the bottom third in the league at just under 30 percent. The line’s WPA and EPA numbers were equally poor. In the NFL, consistency in the running game is what makes a winner. Long break-away runs are fun and exciting, but it’s the steady threat of a 5 or 6-yard run that keeps defenses honest, and that’s the job of the line. Run SR correlates strongly with winning, far more than yards per carry or other measures of running success. This is another stat to pay attention to as the season gets rolling.

The Redskins went for it on 4th and 5 from the Giants’ 37. Was it smart? Let me answer that with a gambling question. Which would you rather have, a certain $5 dollars or a 50/50 shot at $12 dollars? Theoretically, you’re better off with the 50/50 shot at 12 bucks — effectively worth $6 dollars. That’s roughly the same logic needed when facing a fourth down in no-man’s land, outside field-goal range and too close for a punt to be worthwhile. Using recent historical probabilities for converting fourth downs and for field goal success, and using net Expected Points instead of dollars, we can establish guidelines for when to go for it. In the NFL, the 37-yard line is the absolute heart of no-man’s land, where the prospects of kicking or punting are so poor, teams are usually better off going for it on to-go distances of up to 11 yards. So yes, the call was smart.

Rookie quarterback Cam Newton had a phenomenal first game, throwing for 422 yards, 8.7 Adjusted Net Yards Per Attempt, two touchdowns and one interception. But it might be smarter to think of it this way: The Cardinals defense allowed a rookie quarterback to throw for 422 yards, 8.7 AYPA, and 2 touchdowns in his first regular season game. The Cardinals’ Kevin Kolb led all quarterbacks in Week 1 with 10.2 AYPA, so look for a pass-heavy game when the Cards come to town this weekend.

Brian Burke is former Navy pilot who has given up his F/A-18 for the less dangerous hobby of football analysis. He is the creator of Advanced NFL Stats, a website about football, statistics, and and game theory.

By Brian Burke  |  07:00 AM ET, 09/14/2011

Categories:  Statistical analysis

 
Read what others are saying
     

    © 2011 The Washington Post Company