Sitting in the PEW: measuring offensive success by Points-allowed Expected Win differential

By
Updated: June 21, 2013

Points-allowed Expected Win (PEW) differential can be a valuable tool for measuring offensive success.

Amidst the clouds of complex analysis we engage in here at WCSN, sometimes it’s easy to lose sight of the fact that the goal of a football game is pretty simple: score more points than the opponent.

There are essentially three ways a team can prevail in a football game: the offense can score enough points, the defense can prevent the opponent from scoring enough points, or the two units can work together in some combination of both strategies. (Yes, defenses can also score points, but that is not relevant to this discussion.)

Fans tend to be so enthralled by offensive exploits that they sometimes forget just how important defensive contributions to winning really are. Fortunately, we have our resident SportsNerds, Andy Froehle and Johnathan Wood, who never let us forget these key points for long.

History has repeatedly shown that teams not blessed with dominant offenses can still have success if their defenses are strong enough to stifle opposing offenses. The 2000 Baltimore Ravens won a Super Bowl on the shoulders of just such a defense.

Plotting the points

So just how predictive of team success is defensive performance?

In order to get a handle on this, I plotted average points per game given up by the defense against regular-season winning percentage for all NFL teams between 2008 and 2012. Those five seasons’ worth of data gave us a sample size of 160 team-seasons, more than enough for this study.

Figure 1. There is a clear negative relationship (r2 = 0.48) between average points per game given up by the defense and regular-season winning percentage. This allows us to calculate expected win percentage based on the number of points given up by the defense.

It’s obvious from even a cursory glance at this graph that there is a definite negative relationship between points per game given up by the defense and a team’s regular season winning percentage. The data is pretty closely clustered around the line, too; there aren’t many obvious outliers. In fact, a full 48 percent of a team’s win/loss record can be explained by how well the defense prevents opponents from scoring alone.

The two teams that seem to stand out as outliers are the 2011 Cleveland Browns, who only gave up 19.2 ppg but won just four games, and the 2011 Green Bay Packers, who gave up 22.4 ppg and yet won 15 games. The Browns offense that year was extremely anemic (13.6 ppg), while the Green Bay offense was extraordinarily productive (35.0 ppg).

Expected wins

Plugging average points per game given up by a team into the equation of the trend line allows us to calculate points-allowed expected wins (PEW) for that team. For example, we can see that, theoretically, a defense that only gave up 8.3 ppg should win all of its games, one that gives up 22.1 ppg should win about eight games, and a team that gives up 35.9 ppg should win no games.

What this mean, as Andy Froehle explains it, is that the defense has a measurable impact on a team’s probability of winning a game, which gives us another method by which to judge the effectiveness of an offense. Specifically, if a team achieves its expected number of wins, then its offense has performed to the standard required by the quality of that team’s defense. If the team exceeds its expected wins, the offense has been able to compensate for any weaknesses in the defense. Conversely, if the team fails to reach its expected wins, then the offense has failed to harness the full win potential offered by the defense.

In other words, once the defense has made its contribution to winning the game, it’s up to the offense to pull its weight — to take advantage of the opportunities the defense creates.

A team that exceeds its expected wins therefore has a positive PEW differential, while a team that underperforms will have a negative PEW differential.

Top 10 PEW differentials

The following table shows the top 10 PEW differentials of the past five years:

Year Team Pts/G Exp
Wins
Actual
Wins
Win
Diff
2011 Green Bay Packers 22.4 7.8 15 +7.2
2011 New England Patriots 21.4 8.4 13 +4.6
2010 New England Patriots 19.6 9.5 14 +4.5
2009 New Orleans Saints 21.3 8.5 13 +4.5
2011 New Orleans Saints 21.2 8.5 13 +4.5
2009 Indianapolis Colts 19.2 9.7 14 +4.3
2012 Indianapolis Colts 24.2 6.8 11 +4.2
2009 San Diego Chargers 20.0 9.2 13 +3.8
2008 Arizona Cardinals 26.6 5.4 9 +3.6
2008 Denver Broncos 28.0 4.6 8 +3.4

Table 1. The ten highest Points-allowed Expected Win (PEW) differentials attained by NFL teams in the past five seasons (2008-2012). PEW differential is defined as the difference between the expected wins predicted by average defensive points per game given up and the actual wins achieved. 

It’s not all that surprising to see the 2011 Green Bay Packers in the top slot. What is surprising, though, is just how high their PEW differential (+7.2) is. Based on the 22.4 ppg given up by the defense, this team could have been expected to win only about eight games — yet they won almost a full half-season’s worth of games more than that. It just goes to show how truly dominant the offense was that year.

Also not a surprise is the fact a few teams — New England Patriots, Indianapolis Colts, and New Orleans Saints — are overrepresented in this top-ten list. All of these teams have featured extremely productive offenses of late. Each of them has won about a quarter of a season’s worth of games more than expected.

In the last five years, at least, it seems that regular-season success has largely been predicated on offenses exceeding the capabilities of their defenses. Interestingly, though, there is only one Super Bowl Winner (the 2009 New Orleans Saints) on this list. Gaudy regular-season numbers are no guarantee of postseason success.

Bottom 10 PEW differeentials

For the purposes of comparison, the table below shows the bottom ten teams by PEW differential:

Year Team Pts/G Exp
Wins
Actual
Wins
Win
Diff
2012 Kansas City Chiefs 26.6 5.4 2 -3.4
2011 Miami Dolphins 19.6 9.5 6 -3.5
2010 Cleveland Browns 20.8 8.8 5 -3.8
2011 Jacksonville Jaguars 20.6 8.9 5 -3.9
2010 Carolina Panthers 25.5 6.0 2 -4.0
2009 St. Louis Rams 27.2 5.1 1 -4.0
2011 St. Louis Rams 25.4 6.1 2 -4.1
2008 Cleveland Browns 21.9 8.1 4 -4.1
2009 Washington Redskins 21.0 8.6 4 -4.6
2011 Cleveland Browns 19.2 9.7 4 -5.7

Table 2. The ten lowest Points-allowed Expected Win (PEW) differentials attained by NFL teams in the past five seasons (2008-2012). 

Quite obviously, these teams have represented the dregs of the NFL in recent years. Not one of these teams came close to having a winning record. Then again, based on their defensive performances, none of them would have been expected to have dominant seasons either, although the 2011 Miami Dolphins and 2011 Cleveland Browns should have won almost 10 games. The Browns are easily the most disappointing team on this list, having underachieved by an abysmal 5.7 games.

There is also an intriguing symmetry in these two tables: most of the teams in the top 10 overachieved by about four games, while most of those in the bottom 10 underachieved by roughly four games. Perhaps that points to the parity in this league, indicating that there is some sort of practical (even if not theoretical) limit to how drastically a team’s performance will vary from its expected win count.

With the exception of the Packers and Saints, no team from either of our favorite two divisions appears on these lists, so in the next installment of this series, we will look at how the teams of the NFC North and NFC  South have fared with respect to win differential in the last five years.

About the author(s)

Rourke Douglas Decker covers the Green Bay Packers beat for Water Cooler Sports. He resides with his family in the Twin Cities. He can be reached for questions or comments at . Connect with .

1107 comments
Rourke
Rourke

Coooool, I have internet again. It's crazy how impossible it can be to publish things when there is no power.

Look for an article by Brock Alton on Sunday, since power was down all of Saturday.

MIBearFan
MIBearFan moderator

Hawks fans are coming out of the woods.

MIBearFan
MIBearFan moderator

Go Hawks! Win it in Boston. That city needs something good to happen, so they can get over the bombing.

Kuato Face
Kuato Face moderator

Looks like the Blackhawks are on a mission to prove the existence of a Mythical but supposedly once prominent flatfish in these parts.

Mr. Horse
Mr. Horse

Headline of the week, from Science Daily

Three Centaurs Follow Uranus Through the Solar System

aciddragon
aciddragon

i guess i need to break down and buy a Fiat 

MIBearFan
MIBearFan moderator

Baking a homemade strawberry pie. Yumm!

Kuato Face
Kuato Face moderator

I am building a Warhammer 40K squad in Fallout New Vegas. Pretty fun. I love the unlimited followers mod.

Mr. Horse
Mr. Horse

Who would not want a game worn Ahmad Carroll jersey? And only US $149.99 on eBay.

Pat Fenis, Esq.
Pat Fenis, Esq. moderator

Going to watch the Crew again today for a bachelor party

Kuato Face
Kuato Face moderator

Archer really cracks me up.

aciddragon
aciddragon

 like Bears fan slow, or Viking fan slow?

aciddragon
aciddragon

 nice! keep it up and you'll make somebody a great wife someday 

Kuato Face
Kuato Face moderator

 this sentence has me reeling with confusion.

Kuato Face
Kuato Face moderator

Best line. "I hate surprises. Except for surprise fellatio."

MIBearFan
MIBearFan moderator

  

Or I'll eat like a fucking king without all the hassle.

Mr. Horse
Mr. Horse

  

Good God.  I hope not.  Some fool must have bought this when he was headed to the HOF for sure during his rookie season, before anybody had seen much.

aciddragon
aciddragon

   uh-huh, that's why you bandwagon with the Blackhawks and now the D-Backs? i saw the facebook posts 

Classic Nardo
Classic Nardo

  I'm not bitter at all - I just call it like I see it.

aciddragon
aciddragon

 now that's just mean. just cause the Bears haven't done anything of note in 28 years doesn't mean you gotta be bitter 

aciddragon
aciddragon

   Lions fans aren't really slow, they're manic. 

aciddragon
aciddragon

 that too you Croc wearing hipster

MIBearFan
MIBearFan moderator

  

I thought you said footwear.

aciddragon
aciddragon

 no you are wrong as usual, much like in your choice of football teams 

MIBearFan
MIBearFan moderator

  

Obviously, me. You're...

GGGGAAAAAAYYYYYYY!!!!!!!!!!!!

aciddragon
aciddragon

 i think we both know which one of us is more correct 

Kuato Face
Kuato Face moderator

 I was sort of thinking that is what you meant but I could not fathom why a bachelor party would be at a baseball game.