Fun with NFL Stats

I'm a statistics guy. I like using numbers. While they have shortcomings, and can be manipulated in many ways, they are able to show you trends and determine who is (or isn't) a great team in sports.

I decided to just play around a little with the NFL results from the last seven seasons (2002 through 2008), the ones since the most recent expansion and divisional realignment. I grabbed the initial statistics from the ESPN.com NFL standings page. After deleting the streak, divisional and conference record columns, I added conference, division, expected win percentage, luck, point differential, and if the team made the playoffs.

Those are mostly basic things but let me explain "expected win percentage" and "luck". Expected win percentage is just an adaptation of Bill James' Pythagorean expectation for baseball. Simply put, a teams expected win percentage is the square of their points scored divided by the sum of the square of their points scored plus the square of their points allowed. In an easier, visual form


Yes, that was created for baseball. Yes, there is a different one for basketball, where the exponent is 14 (or 16.5, depending on the formula used). I didn't find one for football so I just instituted the basic square one. If you have a link to text on a Pythagorean expectation for football, by all means send it to me or leave it in the comments below.

Luck is straightforward after explaining the expected win percentage value. Luck is merely a teams actual winning percentage minus their expected winning percentage. A negative value would mean "bad luck" and a positive "good luck". The more negative or positive, the worse or better the teams luck that season was.

So what did any of this initially pointless playtime with data reveal? Well it definitely confirmed that old saying. You know the one.. "defense wins championships". That cliche is confirmed. You have a better chance of winning a Super Bowl with a stout defense than you do an electric offense.

There are 224 individual data rows (32 teams x 7 seasons). The 1st, 9th, 11th, 18th, 22nd, 24th, 25th, 59th, 60th, 80th, 103rd, 105th, 106th, and 125th ranked offenses based on points scored throughout the regular season made it to the Super Bowl. That's an average rank of 51st. What teams actually won the game? The 18th, 24th, 59th, 80th, 103rd, 105th and 106th ranked offenses. Their average rank is 71st. The average Super Bowl loser has an offense ranked 36th.

So what about the defense? Clearly a top notch offense isn't a necessity, though it really can't hurt. How do defenses appear in this simple statistical analysis? The 1st, 4th, 7th, 12th, 14th, 16th, 17th, 28th, 29th, 57th, 60th, 134th, 146th, and 205th ranked defenses (in terms of points allowed during the regular season) made the Super Bowl during the last seven seasons. That's an average rank of 52nd, on spot lower than the offense. Huh? So maybe defenses don't win championships? Well let's look at what Super Bowl contenders came out on top. The 1st, 4th, 7th, 14th, 16th, 134th and 146th ranked defenses of the last seven seasons managed to win a Super Bowl. That's an average rank of 46th, 25 spots better than the average Super Bowl winners offenses. Super Bowl losers had defenses ranked 12th, 17th, 28th, 29th, 57th, 60th, and 205th. That's an average rank of 58th. Clearly having a top defense is more important than having a top offense.

Enough with analyzing cliches, what about luck? Everyone makes claims about a teams luck. Luck is a highly debated topic and people argue for and against it being a reason for success or failure. While not completely scientific, an easy way to determine luck in team sports is to luck at their actual win/loss record versus their expected total (explained earlier). Over the last seven seasons who have been the unlucky or lucky teams? And what did luck look like for our Super Bowl competitors?

BAD LUCK

1. 2008 Detroit Lions: -21.2%
2. 2007 Miami Dolphins: -20.9%
3. 2004 Tampa Bay Buccaneers: -18.2%
4. 2006 Detroit Lions: -18.2%
5. 2008 Kansas City Chiefs: -17.9%
6. 2005 Green Bay Packers: -17.9%
   
7. 2008 Green Bay Packers: -17.4%
8. 2002 Cincinnati Bengals: -14.7%
   
9. 2006 Jacksonville Jaguars: -14.7%
10. 2002 Jacksonville Jaguars: -14.5%

GOOD LUCK

1. 2004 Pittsburgh Steelers: 25.1%
2. 2003 New England Patriots: 19.4%
3. 2004 Atlanta Falcons: 18.4%
4. 2007 New England Patriots: 17.8%
5. 2006 Indianapolis Colts: 16.5%
6. 2003 Carolina Panthers: 15.5%
7. 2002 Green Bay Packers: 15.4%
8. 2007 Dallas Cowboys: 15.1%
9. 2006 San Diego Chargers: 15.0%
10. 2008 Miami Dolphins: 14.6%

Super Bowl Teams

• Five of the seven Super Bowl Champions had "good luck". Only the 2005 Pittsburgh Steelers and 2002 Tampa Bay Buccaneers had "bad luck" during the season. All of the Super Bowl losers had "good luck" during the regular season.
  
• The 2003 New England Patriots were the luckiest team to win the Super Bowl.
• The 2002 Oakland Raiders were the least luckiest team to lose a Super Bowl, with luck of just 0.1%.
• Of the top 10% luckiest teams (top 22), six made the Super Bowl and three of them won it.
  
• The 2003 and 2004 Super Bowls featured some of the luckiest teams of the last seven seasons. Both couldn't come out on top obviously.
  

Miscellanea

• The luckiest team to not make the playoffs was the 2005 Minnesota Vikings. They were outscored by a total of 38 points over the course of the season but still won nine games. They had a luck score of 12.1%.
• The unluckiest team to ever make the playoffs was the 2008 San Diego Chargers, they outscored opponents by 92 points over the course of the season but ended up with only a .500 record. They needed to win their final three games and Denver to lose their final three. The last game of the season pitted the team against each other, where the Chargers destroyed the Broncos 52-21.
• The 2008 New England Patriots are the only team during the analyzed time span (2002-2008) to go 11-5 and miss the playoffs.
• It has happened to three teams that went 10-6: 2003 Miami Dolphins, 2005 Kansas City Chiefs and 2007 Cleveland Browns.

So those are just some interesting things I noticed while I was aimlessly sorting and manipulating numbers this morning.