DVOA: A Near Perfect Way to Predict a Team's Final Record

Kevin DeBenedetto@KDebo9Correspondent ISeptember 11, 2009

TAMPA, FL - JANUARY 30:  The Vince Lombardi Trophy is seen during the NFC Head coach press conference prior to Super Bowl XLIII held at the Tampa Convention Center on January 30, 2009 in Tampa, Florida.  (Photo by Chris Graythen/Getty Images)

Every year,  experts, writers, and fans have made their predictions for the upcoming NFL season. For many, it seems almost like a guess, trying to understand the teams and players. Injuries, projections, and the stats that are not looked at make up a huge part of predictions though.

Rarely do you see a person make perfect predictions. There isn’t even a computer program that could simulate it perfectly. The best thing that is out their, in my opinion, is DVOA, or Defense Adjusted Value over Average.

DVOA? Explain please

Of course I will explain. Defense Adjusted Value over Average, or DVOA, is a formula written by the people at www.footballoutsiders.com. For short, what it does is evaluates a play, every play, and compares it to the league average.

The real tricky part, though, is it does NOT use 2008 (or the year prior) statistics, instead it uses PROJECTIONS for the upcoming season.

It takes things into account such as quarterback experience, injuries on the offensive line, and other things. There are many more things that go into it, and if you would like to read about it more, you can read about it at the website.

Ok, it sounds interesting, but what does it predict?

DVOA predicts the mean wins a team will get, their offensive rank, defensive rank, and special teams rank. It will also rank the teams with a percentage, based on the league average.

Each statistic gets a percentage, and that percentage is either negative or positive. That is how much better, or worse, the team will be at that statistic. Note, that like all predictions, it is not perfect. It is probably the closest I’ve seen though.

Can it predict individual games?

Good question, but the answer is no. It predicts it on an season basis, at the beginning of the season. As the season goes on though, and more stats are available, the predictions are updated. This gives a better understanding and better estimate for final records and rankings as the season goes on.

This sounds good and all, but just how accurate is it?

Of course, everyone wants results. We will look at the 2008 original predictions and the final win column for each team.


New England: Projected wins- 12 Amount of wins- 11
San Diego: Projected wins- 10.7 Amount of wins- 8
Indianapolis: Projected wins- 9.4 Amount of wins- 12
Jacksonville: Projected wins- 9.3 Amount of wins- 5
Pittsburgh: Projected wins- 8.1 Amount of wins-  12
Baltimore: Projected wins- 8.3 Amount of wins- 11
Houston: Projected wins- 7.9 Amount of wins- 8
Cincinnati: Projected wins- 7.3 Amount of wins- 4
Tennessee: Projected wins- 7.0 Amount of wins- 13
New York: Projected wins- 7.7 Amount of wins- 9
Cleveland: Projected wins- 6.1 Amount of wins- 4
Buffalo: Projected wins- 6.6 Amount of wins- 7
Kansas City: Projected wins- 6.3 Amount of wins- 2
Miami: Projected wins- 6 Amount of wins- 11
Oakland: Projected wins- 5.7 Amount of wins- 5


Philadelphia: Projected wins- 11.1 Amount of wins- 9
Green Bay: Projected wins- 10.4 Amount of wins- 6
Minnesota: Projected wins- 9.9 Amount of wins-  10
Seattle: Projected wins- 9.8 Amount of wins-  4
Tampa Bay: Projected wins- 9.8 Amount of wins-  9
Dallas: Projected wins- 8.9 Amount of wins- 9
New York: Projected wins- 8.3 Amount of wins- 12
Carolina: Projected wins- 8.4 Amount of wins- 12
New Orleans: Projected wins- 8.2 Amount of wins- 8
Arizona: Projected wins- 7.9 Amount of wins-  9
Chicago: Projected wins- 7.0 Amount of wins- 9
Washington: Projected wins- 6.6 Amount of wins- 8
St. Louis: Projected wins- 5.7 Amount of wins- 2
San Francisco: Projected wins- 5.4 Amount of wins- 7
Atlanta: Projected wins- 4.9 Amount of wins- 11

That’s good and all, but can you break that down further for me Kevin?

Within two games of their projected wins to their total amount of actual wins, there are 21 teams! If we further that and make it three games, that is 24 total teams. 24 teams within three wins of their projected wins, before the season starter. That is very well done. But too break down the teams that didn’t fall within three games, I will break them down.

Jacksonville: Four games away
They were a very injured team, losing some defensive and offensive line players.

Pittsburgh: Four games away
The hardest schedule, and they really played above expectations.

Tennessee: Six games away
Show me one person that had the Titans winning 11+ games last year, I’ll show you a crazy homer.

Kansas City: Four games away
Thigpen injured, and lost very close games.

Miami: Five games away
I think everyone remembers the wild cat and how it turned over the league.

Green Bay: Four games away
Defense underperformed throughout the whole year.

Seattle: Five games away
Hasslebeck injured throughout the year, offensive line injuries.

Atlanta: Six games away
Who predicted a Matt Ryan year like he had? Another surprise team.

As you can see, many were four games away. The teams that were far away, either shocked the league (Tennessee, Miami, Atlanta) or had injury to important players.

Kevin, you’ve shown me the light! I’m convinced. What does DVOA say about the league this year?

Ask, and you shall receive.

San Diego: 11.3 wins
Indianapolis: 10.8 wins
New England: 11.1 wins
Chicago: 10.4 wins
Pittsburgh: 9.9 wins
Jacksonville: 10 wins
Tennessee: 9.4 wins
Minnesota: 9.6 wins
Baltimore: 9.1 wins
New York Giants: 9.7 wins
Seattle: 9.5 wins
St. Louis: 8.7 wins
Philadelphia: 9.1 wins
Carolina: 8.6 wins
Dallas: 7.9 wins
Kansas City: 7.6 wins
Cincinnati: 7.2 wins
Green Bay: 7.0 wins
Washington: 7.3 wins
New Orleans: 7.6 wins
Detroit: 6.6 wins
Houston: 6.5 wins
Cleveland: 6.4 wins
Miami: 6.7 wins
Tampa Bay: 7.1 wins
Oakland: 6.7 wins
Atlanta: 6.7 wins
New York Jets: 6.3 wins
Arizona: 5.5 wins
San Fransisco: 5.4 wins
Denver: 5.1 wins
Buffalo: 5.4 wins

Break down this year for me please?

Jacksonville, Seattle, and St. Louis should surprise some people this year. We should expect major drops from Arizona, Atlanta, and Miami.

Note that these are not perfect. Do we expect St. Louis to have close to nine wins next season? Of course not, but that is what is predicted.

I would take these predictions, which got 24/32 teams within three games last year, then some "expert" predictions any day of the week.

Nonetheless, I expect a great season, and at season end will write another article about DVOA, and show these predictions and show how they do at the end of the season.

If you would like to learn more about DVOA, and understand it better, simply go on over to the website and read about it.

(Note: I do not work for www.footballoutsiders.com. I simply enjoy DVOA, and find it a great tool, and I wanted to share it with everyone)


    Brett Favre Looking to End Youth Tackle Football

    NFL logo

    Brett Favre Looking to End Youth Tackle Football

    Green Bay Packers
    via Green Bay Packers

    LAC Scout Was Why Lamar Didn't Run 40

    NFL logo

    LAC Scout Was Why Lamar Didn't Run 40

    Alec Nathan
    via Bleacher Report

    One Player on Each Team Who Is Set to Disappoint

    NFL logo

    One Player on Each Team Who Is Set to Disappoint

    Doug Farrar
    via Bleacher Report

    Does the NFL Have a Referee Problem?

    NFL logo

    Does the NFL Have a Referee Problem?

    Mike Florio
    via ProFootballTalk