Writers of Pro Football Prospectus 2008

Most Recent FO Features

JonesJon15.jpg

» SDA: Early Playoff Elimination Round

TCU-West Virginia and Auburn-Ole Miss might as well be early playoff elimination rounds, with the losers likely knocked out of playoff contention.

06 Feb 2014

Historical DVOA Estimates

Guest Column by Andreas Shepard

As readers of Football Outsiders, you know and love the proprietary statistics like DVOA. DVOA has many wonderful qualities –- it takes into account both raw yardage and progress towards a first down, it adjusts for situation and opponent quality, and it is an excellent predictor of future performance. It does, however, have one drawback –- it requires play-by-play data, which as of now only goes back to 1989. To work around that difficulty, and to add more fuel to the perpetual argument about the average teams for now, I developed a method to estimate DVOA for seasons where play-by-play data is not yet available. I’m sharing the results here. It isn’t perfect, but it should be enough to get some good discussions going.

Before we get to the results, let me explain the method I used. (If you aren’t interested in how the sausage was made, skip ahead to the next section.) I only considered teams from 1950 to the present. That was the year of the AAFC-NFL merger, which marked a) the beginning of more reliable stat keeping and b) the end of the era where expansion teams would start up and then fold within a few seasons. For each year from 1950 on, I collected total box score statistics for each team on offense, defense, and special teams (yards, touchdowns, interceptions, fumbles, sacks, first downs, field goals, return yards, etc.). As DVOA is a per-play metric, I calculated each team’s per-play average for each of these stats. I then normalized each stat against the league average for that season. To translate these normalized per play statistics to VOA, I ran a series of linear regressions using the data from 1989-2012 (one each for rushing, passing, FG/XP, kickoffs, kick returns, punts, and punt returns). After adding opponent adjustments, these ratings are combined into composite ratings for total offense, defense, special teams, and overall.

This method produces estimated DVOA that has a correlation coefficient of 0.958 with actual DVOA. This is very high, but still not 100 percent; I can’t claim that when Football Outsiders gets its hands on the actual play-by-play data for these years that the numbers will line up exactly (although the results for 1989 and 1990 came in quite close to my estimates). And of course, even with full play-by-play data, DVOA is only an imperfect approximation of a team’s true performance. So if you disagree with the results below, just remember that these numbers are a flawed estimate of a flawed estimate of true team quality.

Two other important notes: First, 1987 numbers below include the strike data. Second, like the usual DVOA tables on the FO stat pages, all numbers here represent only the regular season.

With those preliminaries out of the way, let’s get to the results: offenses, defenses, special teams, and then total results. Actual DVOA ratings are published in standard type, with estimated DVOA for teams from 1950 through 1988 published in italics.

Best Offensive DVOA 1950-2013 (estimated 1950-1988)
Total Offense x Pass Offense x Rush Offense
Rank Team DVOA Pass Rush x Team DVOA x Team DVOA
1 NE 2007 43.5% 72.7% 19.6% CLE1 1953 75.0% STL 2000 36.5%
2 NE 2010 42.2% 67.5% 24.2% NE 2007 72.7% CLE1 1966 32.2%
3 CLE1 1953 40.2% 75.0% 9.6% GB 2011 67.6% CAR 2011 32.1%
4 SD 1982 35.9% 49.3% 18.3% IND 2004 67.6% DEN 1998 31.4%
5 KC 2002 35.4% 47.0% 29.3% NE 2010 67.5% SD 1963 31.0%
6 LARM 1951 35.0% 50.0% 21.1% WAS 1991 65.0% SF 1993 30.5%
7 DEN 1998 34.5% 52.5% 31.4% DEN 2013 60.7% SF 1954 30.0%
8 LARM 1954 34.4% 47.3% 24.2% SD 2009 59.6% KC 2002 29.3%
9 GB 2011 33.8% 67.6% 5.5% MIA 1984 57.6% DET 1990 29.2%
10 DEN 2013 33.7% 60.7% 4.3% HOIL 1961 56.3% CLE1 1963 29.0%
11 MIA 1984 33.5% 57.6% 4.2% IND 2006 55.4% KC 2003 28.9%
12 KC 2003 33.4% 43.3% 28.9% NE 2011 55.3% CLE1 1960 28.3%
13 SF 1992 33.1% 51.8% 20.7% TEN 2003 54.7% SF 1998 27.8%
14 NO 2011 33.0% 50.1% 18.5% NE 2012 53.8% SD 2006 27.2%
15 CLE1 1966 32.5% 32.7% 32.2% NE 2009 53.5% DEN 2005 26.8%
16 HOIL 1961 32.4% 56.3% 5.3% DEN 1998 52.5% CLE1 1958 26.7%
17 NE 2011 31.9% 55.3% 12.6% SF 1992 51.8% PIT 2000 26.4%
18 IND 2004 31.8% 67.6% 5.2% SF 1989 51.7% CLE1 1952 26.2%
19 KC 2004 31.6% 45.1% 24.9% MIN 1998 51.6% SD 2003 26.1%
20 SF 1993 31.4% 44.2% 30.5% SD 2013 51.3% KC 1966 25.9%

Tom Brady’s MVP seasons came while leading the two best offenses since 1950, but neither of those teams had the best estimated passing DVOA. That honor goes to Otto Graham and the 1953 Browns. They led the league in yards per attempt, completion percentage, interception percentage, and first downs per pass. Their 8.5 net yards per attempt was over 3 yards better than the league average, the largest gap ever.

On the rushing side, Marshall Faulk was really good, helping power the 2000 Rams to the best rushing performance on record. This may shock you, but Jim Brown was also really good. He played nine seasons, and the Browns finished in the top 25 in estimated rushing DVOA in five of them. Of course the rest of his team wasn’t too shabby either –- they managed to put together the second-best rushing DVOA ever the year after he retired, when running backs Leroy Kelly and Ernie Green both made the Pro Bowl.

Worst Offensive DVOA 1950-2013 (estimated 1950-1988)
Total Offense x Pass Offense x Rush Offense
Rank Team DVOA Pass Rush x Team DVOA x Team DVOA
1 HOU 2002 -43.3% -37.9% -27.4% SEA 1992 -65.3% IND 1991 -30.2%
2 SEA 1992 -41.3% -65.3% -6.6% ATL 1974 -62.7% PIT 1957 -29.2%
3 SF 2005 -40.4% -56.0% -12.2% TB 1977 -62.5% ARI 2005 -29.1%
4 TB 1977 -39.8% -62.5% -21.9% BUF 1968 -57.6% TB 1983 -27.9%
5 PIT 1965 -39.5% -57.0% -21.6% PIT 1965 -57.0% WAS 1962 -27.7%
6 OAK 2006 -37.0% -45.4% -12.8% SF 2005 -56.0% HOU 2002 -27.4%
7 CHI 2004 -36.5% -50.6% -8.5% NYJ 1976 -54.9% JAC 2013 -27.2%
8 DEN 1966 -36.2% -44.8% -26.1% DET 1959 -50.9% BAL 2013 -27.2%
9 CAR 2010 -35.8% -40.1% -20.7% CHI 2004 -50.6% PHI 1954 -26.8%
10 ARI 2010 -35.6% -46.1% -8.6% BALC 1953 -50.3% WAS 1965 -26.3%
11 NO 1997 -35.6% -41.4% -19.1% BOS 1968 -49.7% DEN 1966 -26.1%
12 BOS 1968 -34.8% -49.7% -19.0% GB 1950 -49.7% ARI 1995 -25.1%
13 ATL 1974 -34.6% -62.7% -6.1% NO 1975 -47.8% WAS 1973 -25.1%
14 TB 1976 -33.6% -45.6% -21.7% SF 1978 -46.9% PIT 1966 -24.9%
15 BUF 1968 -33.4% -57.6% -6.5% ARI 2010 -46.1% NYG 1953 -23.5%
16 IND 1991 -32.8% -29.0% -30.2% TB 1976 -45.6% WAS 1968 -23.1%
17 SF 2007 -32.2% -42.1% -3.1% JAC 2011 -45.4% PHI 1991 -23.0%
18 NYG 1951 -32.0% -36.6% -22.3% OAK 2006 -45.4% DET 2006 -22.9%
19 ARI 2012 -31.0% -30.3% -22.1% DEN 1966 -44.8% NO 1998 -22.9%
20 NYJ 1995 -30.7% -35.6% -18.3% IND 1984 -44.5% ATL 2000 -22.8%

Taking a look at the worst ever offenses, some familiar names appear. You remember that David Carr set the sack record for the expansion Texans, but his teammate Jonathan Wells also set the record for worst Rush DYAR in the last 20 years. The Texans ranked last in the league that year in basically every offensive category. However, they did not rank last in all-time passing DVOA. That dubious honor goes to the 1992 Seahawks, best remembered as the team that went 2-14 despite having defensive player of the year Cortez Kennedy. That will happen when your starting quarterbacks are Stan Gelbaugh, Kelly Stouffer, and Dan "is his brother available?" McGwire. The worst rushing offense title belongs to the 1991 Colts, who somehow managed to rush for only three touchdowns all season while averaging a mere 3.3 yards per carry. I’ll leave the Trent Richardson jokes to you.

Best Defensive DVOA 1950-2013 (estimated 1950-1988)
Total Defense x Pass Defense Rush Defense
Rank Team DVOA Pass Rush x Team DVOA x Team DVOA
1 PHI 1991 -42.4% -48.6% -34.9% TB 2002 -51.9% BAL 2000 -36.6%
2 MIN 1969 -32.9% -42.5% -19.9% PHI 1991 -48.6% PHI 1991 -34.9%
3 TB 2002 -31.8% -51.9% -8.8% MIA 1982 -48.3% SD 1998 -32.9%
4 PIT 1976 -30.1% -35.7% -25.0% MIN 1969 -42.5% MIN 2006 -30.5%
5 PIT 2008 -29.0% -32.8% -24.2% CHI 1963 -42.4% KC 1995 -30.5%
6 PIT 1974 -28.9% -41.8% -18.3% PIT 1973 -42.2% NYG 1951 -30.2%
7 BUF 2004 -28.5% -34.7% -21.9% PIT 1974 -41.8% PIT 2010 -29.0%
8 LARM 1968 -28.5% -35.2% -21.3% NYG 1963 -40.2% BAL 2008 -28.6%
9 PIT 1973 -27.9% -42.2% -16.5% MIN 1970 -38.5% TEN 2000 -27.4%
10 BAL 2008 -27.8% -27.1% -28.6% MIN 1988 -38.4% BAL 2007 -27.3%
11 CHI 2012 -26.8% -29.0% -23.2% DET 1965 -38.2% SD 2000 -26.6%
12 CHI 1985 -26.8% -35.4% -12.8% NYJ 2009 -36.5% CHI 2011 -26.5%
13 CHI 1963 -26.7% -42.4% -11.8% PIT 1976 -35.7% SF 2011 -26.1%
14 DET 1962 -26.1% -29.7% -21.7% NYG 1961 -35.6% WAS 2004 -25.7%
15 OAK 1967 -26.0% -34.5% -13.2% CHI 1985 -35.4% SD 1999 -25.7%
16 SEA 2013 -25.8% -34.3% -15.1% LARM 1968 -35.2% BAL 2006 -25.3%
17 KC 1969 -25.6% -33.0% -14.6% BUF 2004 -34.7% SF 1995 -25.2%
18 NYJ 2009 -25.5% -36.5% -13.9% OAK 1967 -34.5% PIT 1976 -25.0%
19 MIN 1988 -25.2% -38.4% -9.5% ATL 1977 -34.4% ARI 2013 -24.9%
20 BAL 2003 -25.0% -29.5% -19.9% SEA 2013 -34.3% MIN 1994 -24.7%

Shifting to the other side of the ball, the 1991 Eagles are the best defense for now and it’s not even close. The 1969 edition of the Purple People Eaters, who come in second, are closer to 35th place than they are to first. The Steel Curtain makes three appearances in the top-10 in a four-year span (and they won the Super Bowl in the fourth year). Having four Hall of Famers on one defense is a good thing.

You may be surprised to see the Eagles so far ahead of the pack –- I was too. They are generally in the conversation when people talk about all-time great defenses, but not the clear frontrunner. This is partially an issue of a difficult schedule -– their VOA without opponent adjustments is "only" -38.2% -– but that still leaves them comfortably ahead of the pack. The other major reason is their teammates. While the Eagles had an average special teams unit in 1991, that offense was terrible. Their offensive DVOA of -24.6% is the worst by far of any of the other teams with top defensive DVOAs. This set them up in more unfavorable situations than many of the other all-time great defenses. In contrast, Buddy Ryan’s other all-time great defense, the 1985 Bears, had the fifth-best offense in the league. This gave them more opportunities to shine, which helps explain why their defense only comes in 12th all-time instead of first or second like you might expect. The fact that the 1991 Eagles were still able to end up in the conversation despite their albatross of an offense is perhaps their most impressive accomplishment.

I say perhaps, because the 1991 Eagles have another remarkable thing to brag about –- their incredible balance. They had the No. 2 defense against both the pass and the run. Only the 2002 Bucs and 2000 Ravens, respectively, managed to top them. The 1998 Chargers slide into third place for run defense, which helps explain how a team that started Ryan Leaf and Craig Whelihan at quarterback managed to scrape together five wins.

Worst Defensive DVOA 1950-2013 (estimated 1950-1988)
Total Defense x Pass Defense x Rush Defense
Rank Team DVOA Pass Rush x Team DVOA x Team DVOA
1 WAS 1954 34.2% 43.4% 24.9% BALC 1981 46.8% BAL1 1950 29.5%
2 TB 1986 30.8% 40.1% 22.4% MIA 1976 45.6% NYY 1950 29.0%
3 BAL1 1950 30.2% 24.9% 29.5% WAS 1954 43.4% WAS 1959 26.8%
4 SEA 1976 29.5% 41.3% 21.9% BAL 1996 42.0% DLTX 1952 25.9%
5 BALC 1981 28.0% 46.8% 12.3% SF 1999 41.7% NYG 1968 25.8%
6 WAS 1951 26.9% 28.1% 18.3% SEA 1976 41.3% HOIL 1965 25.3%
7 MIA 1968 26.2% 29.5% 23.5% NYG 1966 40.9% WAS 1954 24.9%
8 MIN 2000 26.0% 35.2% 13.8% STLC 1969 40.9% SD 1971 24.6%
9 ATL 1966 25.8% 33.3% 18.9% BALC 1973 40.5% MIA 1986 23.9%
10 NYG 1966 25.5% 40.9% 13.2% TB 1986 40.1% KC 1976 23.5%
11 NE 1972 25.2% 37.6% 17.4% SD 1967 39.4% MIA 1968 23.5%
12 MIA 1986 25.0% 26.2% 23.9% ATL 1967 38.4% OAK 1961 23.1%
13 NYY 1951 24.9% 25.7% 18.0% NYJ 1975 38.3% WAS 1969 22.9%
14 PHI 1973 24.8% 29.8% 21.1% ATL 1996 38.0% DAL 1961 22.5%
15 NO 1977 24.5% 35.2% 19.0% ATL 1968 37.8% TB 1986 22.4%
16 DET 2008 24.3% 32.8% 17.1% NE 1972 37.6% SEA 1976 21.9%
17 SD 1974 24.3% 36.4% 15.1% MIN 1984 37.0% PHI 1973 21.1%
18 MIN 1984 23.4% 37.0% 10.5% NO 1969 36.7% ATL 1985 20.5%
19 WAS 1959 23.1% 18.6% 26.8% DET 2009 36.5% SD 1966 20.5%
20 SD 1971 22.9% 20.6% 24.6% SD 1974 36.4% CLE1 1965 20.5%

The teams that show up on the list of worst defenses are a real sorry bunch. Washington makes multiple appearances during the 1950s, when they were the last holdouts against racial integration. Not coincidentally, this period came in the middle of a 25-year playoff drought. The 1950 Colts and 1951 New York Yanks both went belly up after the season (the current version of the Colts franchise started in 1953), while the 1976 Seahawks were an expansion team that played like one. The 0-16 Lions make an appearance, as do the 1977 Saints that allowed the expansion Buccaneers their first win in 27 tries. For the full story of the extraordinarily dysfunctional 1981 Colts, who set records for most points, yards, touchdowns, and first downs allowed, see Mike Tanier’s article from a few years back. And for an even more messed up franchise, check out the Wikipedia article on the 1952 Dallas Texans, who were playing "home" games in Akron, Ohio and Hershey, Pennsylvania by the end of the season due to poor attendance and financial woes. They folded after the season, becoming the last team in NFL history that isn’t still around in some form today. (The AFL Dallas Texans, who eventually became today’s Kansas City Chiefs, are a separate franchise that started in 1960.)

Best and Worst Special Teams DVOA 1950-2013 (estimated 1950-1988)
Rank Team Best DVOA x Team Worst DVOA
1 LARM 1985 13.3% BUF 2000 -15.4%
2 NYG 1951 12.2% NYG 1984 -13.3%
3 NO 2002 12.2% CHI 1952 -12.8%
4 CHI 1967 11.2% WAS 2013 -12.0%
5 CHI 2007 11.2% NYG 1987 -11.4%
6 CHIC 1959 11.0% SEA 1997 -11.1%
7 CLE1 1994 10.1% BOS 1960 -10.7%
8 KC 1968 10.0% CLE1 1976 -10.3%
9 LARM 1984 9.8% SD 2010 -10.2%
10 CAR 1996 9.8% BUF 1976 -10.0%
11 CLE 2009 9.7% PIT 1986 -10.0%
12 DEN 1966 9.4% CIN 2002 -9.4%
13 OAK 1975 9.2% OAK 1998 -9.3%
14 DAL 1998 9.2% CHI 1997 -9.2%
15 KC 1980 9.1% TB 1992 -9.2%
16 BAL 2012 9.0% TB 1983 -9.2%
17 PHI 2001 8.9% CIN 1980 -9.1%
18 DAL 1997 8.9% STL 2004 -9.0%
19 BUF 2005 8.8% CHI 1954 -8.6%
20 MIA 2000 8.8% MIN 1977 -8.5%

The 1985 Rams are estimated to have beat out the 2002 Saints for the best special teams DVOA ever. They ranked in the top four that year in almost every facet of special teams, with All-Pro Ron Brown and his three kick return touchdowns leading the way. At the other end of the spectrum, the 2000 Bills hold on to their title of worst special teams ever. Aaron Schatz covered this extensively in the original DVOA commentary for the 2000 season, but suffice to say that Buffalo was horrendous in pretty much every way it is possible to be horrendous on special teams.

So what happens when we add it all up? This table gives the best teams in either actual or estimated DVOA. Teams that won the Super Bowl or pre-1967 NFL Championship are listed with asterisks.

Best Total DVOA 1950-2013 (estimated 1950-1988)
Rank Team Total DVOA Offense Defense ST
1 WAS 1991* 56.9% 27.2% -21.1% 8.6%
2 NE 2007 52.9% 43.5% -5.8% 3.6%
3 GB 1962* 47.9% 21.4% -20.7% 5.8%
4 GB 1961* 46.0% 27.6% -11.6% 6.8%
5 CHI 1985* 45.3% 12.5% -26.8% 6.1%
6 NE 2010 44.6% 42.2% 2.3% 4.7%
7 OAK 1967 44.3% 11.6% -26.0% 6.8%
8 PIT 1976 42.7% 10.4% -30.1% 2.1%
9 GB 1996* 41.9% 15.2% -19.3% 7.4%
10 BALC 1968 40.9% 13.0% -21.0% 6.8%
11 SF 1987 40.8% 25.0% -13.4% 2.4%
12 SF 1995 40.1% 18.6% -23.7% -2.2%
13 SEA 2013* 40.0% 9.4% -25.8% 4.8%
14 DAL 1968 39.5% 17.1% -15.7% 6.7%
15 CLE1 1953 39.2% 40.2% 5.6% 4.7%
16 KC 1968 38.8% 15.6% -13.2% 10.0%
17 MIA 1973* 38.4% 19.5% -15.6% 3.3%
18 SEA 2012 38.3% 18.5% -14.1% 5.7%
19 PIT 2004 37.6% 16.3% -18.9% 2.4%
20 KC 1969* 37.6% 7.3% -25.6% 4.6%

Of the 1638 teams since 1950, the best one played in our nation’s capital in 1991. They had incredible balance, with the offense, defense, and special teams each ranking in the top-50 all time. The recent Patriots claim a couple dubious records: best team not to win the Super Bowl (2007) and best to not win a playoff game (2010). The best team of the pre-1991 era actually maintained their performance for two seasons: the 1961-62 Green Bay Packers. Their two-year run is easily the best of any team ever.

The 1985 Bears, who are usually No. 1 or No. 2 in most "Best Team of All Time" lists, come in at No. 5. This is where we have to remember that these estimates only consider the regular season; one of the strongest arguments for the 1985 Chicago Bears as the greatest team ever is the way they stomped over three playoff opponents by a combined score of 91-10.

You’ve probably noticed that a few of the other usual suspects are absent from the list. The undefeated 1972 Dolphins are No. 38 in total DVOA. They were an excellent team, ranking No. 1 in both offense and defense that season, but were not dominant in any category. They are also held back by a weak schedule. Without opponent adjustments, they would jump up to ninth. The 1999 Rams are even more extreme. They are 41st by DVOA, but shoot up all the way to sixth if you ignore opponent adjustments. The gap between their DVOA and VOA is 10.4%, the fifth-biggest difference ever.

Several of the most dominant dynasties had consistently high DVOAs, but generally just missed the cutoff for the top 20. The 1992-95 Cowboys averaged a DVOA of 31.4% and had three teams make the top 50, peaking with 32nd place in 1992. The 1974-79 Steelers had four teams in the top 100 and averaged a DVOA of 30.5%. And most impressively, the 49ers averaged a DVOA of 28.6% for 16 years from 1983-1998, ranking in the top five for all but one season in that stretch. (They were seventh in 1990.)

Worst Total DVOA 1950-2013 (estimated 1950-1988)
Rank Team Total DVOA Offense Defense ST
1 SF 2005 -55.5% -40.4% 16.6% 1.5%
2 DET 2009 -51.6% -28.4% 17.9% -5.3%
3 WAS 1954 -51.0% -16.0% 34.2% -0.8%
4 TB 1976 -50.6% -33.6% 14.4% -2.6%
5 PIT 1965 -50.0% -39.5% 8.8% -1.7%
6 TB 1986 -49.8% -11.9% 30.8% -7.1%
7 DET 2008 -48.3% -25.3% 24.3% 1.3%
8 IND 1991 -47.8% -32.8% 13.5% -1.5%
9 STL 2008 -47.1% -28.2% 18.3% -0.6%
10 STL 2009 -45.0% -29.5% 17.2% 1.7%
11 ATL 1987 -44.0% -21.5% 19.1% -3.5%
12 BOS 1970 -43.4% -25.7% 12.5% -5.1%
13 DEN 1967 -42.7% -25.7% 16.1% -0.8%
14 ARI 2003 -42.0% -21.5% 18.3% -2.2%
15 BUF 1984 -42.0% -24.2% 11.4% -6.4%
16 SF 2004 -41.8% -21.2% 19.4% -1.2%
17 HOU 2002 -41.8% -43.3% 2.3% 3.8%
18 NYG 1966 -41.5% -13.6% 25.5% -2.4%
19 BAL1 1950 -41.5% -10.1% 30.2% -1.2%
20 OAK 1961 -41.5% -14.1% 20.8% -6.5%

Alas, all good things must come to an end. Only eight seasons later, the 49ers had become the worst team since 1950. On the surface, their 4-12 record doesn’t seem too bad, but their wins came by a combined 15 points. In second place we have the 2009 Lions, who somehow got worse by DVOA the year after going 0-16. Coming in third is Washington in 1954, who we saw at the top of the worst defense table. They had losses of 41-7, 37-7, 51-21, and 49-21 ... in the first four weeks of the season. And none of those were even their worst defeat of the year: That would come in a 62-3 demolition by the Browns in Week 7. One of the other two teams with a DVOA below -50% is, appropriately enough, the winless expansion Buccaneers, who since 1976 have allowed fans of other expansion teams to say "well, it could be worse."

Now, some of you out there may be grumbling about the top 20 list including a number of teams who didn’t win, or even play for, the championship. You have a point! The whole purpose of being the best team in a given year is so your team can take home the trophy. So, as a bonus, here is every NFL champion since 1950, ranked by DVOA.

DVOA of Champions 1950-2013 (estimated 1950-1988)
Rank Team Total DVOA Offense Defense ST
1 WAS 1991 56.9% 27.2% -21.1% 8.6%
2 GB 1962 47.9% 21.4% -20.7% 5.8%
3 GB 1961 46.0% 27.6% -11.6% 6.8%
4 CHI 1985 45.3% 12.5% -26.8% 6.0%
5 GB 1996 41.9% 15.2% -19.3% 7.4%
6 SEA 2013 40.0% 9.4% -25.8% 4.8%
7 MIA 1973 38.4% 19.5% -15.6% 3.3%
8 KC 1969 37.6% 7.3% -25.6% 4.6%
9 CLE1 1950 37.5% 14.9% -14.9% 7.7%
10 DAL 1971 36.3% 23.3% -11.3% 1.7%
11 SF 1989 36.0% 26.2% -11.5% -1.7%
12 DAL 1977 35.1% 24.8% -9.6% 0.7%
13 DAL 1992 35.0% 23.6% -9.5% 1.9%
14 PIT 1975 35.0% 13.5% -14.4% 7.2%
15 GB 1966 34.9% 20.1% -16.2% -1.4%
16 MIA 1972 34.5% 18.5% -14.8% 1.2%
Rank Team Total DVOA Offense Defense ST
17 NE 2004 34.2% 23.3% -10.7% 0.2%
18 STL 1999 34.0% 17.7% -13.5% 2.8%
19 PIT 1979 33.4% 13.9% -20.2% -0.7%
20 DAL 1995 32.7% 29.6% 0.9% 4.0%
21 BALC 1958 32.6% 19.0% -12.8% 0.9%
22 DEN 1998 32.5% 34.5% 4.3% 2.3%
23 SF 1984 32.5% 28.6% 0.5% 4.3%
24 CLE1 1954 31.9% 15.1% -15.2% 1.7%
25 TB 2002 31.6% -3.8% -31.8% 3.6%
26 NYG 1990 30.8% 10.5% -14.4% 5.9%
27 GB 1967 30.1% 8.8% -12.9% 8.4%
28 PIT 1974 30.1% -3.4% -28.9% 4.6%
29 LARM 1951 29.9% 35.0% -2.3% -7.4%
30 DEN 1997 29.6% 19.4% -5.9% 4.3%
31 NYJ 1968 29.3% 16.8% -15.6% -3.2%
32 SF 1994 27.6% 18.9% -7.5% 1.2%
Rank Team Total DVOA Offense Defense ST
33 PIT 2005 27.1% 12.0% -13.5% 1.6%
34 SF 1988 26.9% 14.2% -10.6% 2.1%
35 CHI 1963 26.4% 1.9% -26.7% -2.2%
36 PIT 2008 26.0% -1.5% -29.0% -1.5%
37 DET 1952 25.7% 9.6% -9.1% 7.0%
38 DAL 1993 24.8% 21.8% 0.8% 3.8%
39 BAL 2000 24.1% -8.1% -23.8% 8.4%
40 GB 2010 23.0% 11.5% -13.9% -2.4%
41 NO 2009 21.3% 24.3% -0.4% -3.4%
42 NE 2003 20.7% 1.2% -18.7% 0.8%
43 CLE1 1955 20.4% 17.2% -5.8% -2.6%
44 OAK 1976 18.8% 24.8% 9.4% 3.4%
45 WAS 1982 17.8% 5.7% -7.3% 4.8%
46 IND 2006 16.4% 28.5% 8.5% -3.6%
47 PIT 1978 16.2% 2.7% -12.4% 1.1%
48 GB 1965 13.8% -2.1% -15.2% 0.7%
Rank Team Total DVOA Offense Defense ST
49 DET 1953 13.5% 5.8% -6.4% 1.3%
50 SF 1981 13.3% 11.0% -5.8% -3.5%
51 LARD 1983 12.5% 0.1% -9.0% 3.4%
52 NYG 1986 11.6% 3.9% -7.9% -0.1%
53 BAL 1959 11.6% 12.5% 1.0% 0.0%
54 NYG 1956 11.3% 10.5% -4.2% -3.4%
55 PHI 1960 10.8% 10.2% 2.6% 3.2%
56 BAL 2012 9.8% 3.0% 2.2% 9.0%
57 CLE 1964 8.8% 23.0% 19.8% 5.6%
58 NYG 2011 8.4% 10.5% 2.4% 0.3%
59 NE 2001 8.0% 3.4% -1.5% 3.1%
60 WAS 1987 6.7% 10.1% 2.9% -0.5%
61 DET 1957 6.1% 0.8% -5.0% 0.3%
62 NYG 2007 1.8% -1.1% -3.8% -0.9%
63 OAK 1980 0.0% -7.7% -7.6% 0.1%
64 BAL 1970 -3.3% -1.1% 3.9% 1.6%

I’ll mostly leave the commentary to you folks, but one word on the 1970 Baltimore Colts, who come in as the only team to win the Super Bowl with negative DVOA. On the surface, they seem like a strong team, with an 11-2-1 record. However, they had the point differential of a team that would be expected to win only 9.5 games. Furthermore, they faced a weak schedule: their VOA of 5% is 8% higher than their DVOA of -3%. Add those together, and Baltimore was significantly weaker than its record.

Finally, for those curious, this last table compares projected DVOA to actual DVOA for the top 20 teams for 1989-2012. (I haven’t been able to assemble the data to create a projection for 2013 yet.)

Actual DVOA vs. Estimated DVOA, 1989-2012
Rank Team Actual DVOA Est. Rank x Team Est. DVOA Actual Rank
1 WAS 1991 56.9% 2 x NE 2007 46.3% 2
2 NE 2007 52.9% 1 WAS 1991 40.4% 1
3 NE 2010 44.6% 9 GB 1996 40.1% 4
4 GB 1996 41.9% 3 TEN 2000 39.2% 16
5 SF 1995 40.1% 5 SF 1995 38.3% 5
6 SEA 2012 38.3% 32 DEN 2012 36.6% 9
7 PIT 2004 37.6% 88 SF 1997 36.2% 49
8 DEN 2012 36.6% 6 STL 1999 35.9% 15
9 SF 1989 36.0% 10 NE 2010 35.2% 3
10 PIT 2010 35.4% 36 SF 1989 35.0% 10
11 NE 2012 35.0% 11 NE 2012 34.9% 12
12 DAL 1992 35.0% 38 DAL 1994 34.9% 17
13 NE 2004 34.2% 52 DAL 1995 34.2% 19
14 STL 1999 34.0% 8 SF 1994 33.0% 49
15 TEN 2000 33.3% 4 TB 2002 32.5% 24
16 DAL 1994 32.9% 12 IND 2004 32.3% 63
17 DAL 1995 32.7% 13 SF 1992 31.8% 29
18 DEN 1998 32.5% 31 OAK 2002 30.1% 41
19 IND 2005 32.1% 30 KC 2003 29.8% 27
20 PHI 2008 31.8% 42 DEN 1997 29.8% 33

There doesn’t seem to be a deeper pattern behind the biggest outliers, like the 2004 Steelers and the 2004 Colts. In general, those teams are just a bit better (or worse) than my estimates across all subcategories of DVOA. Passing offense and defense account for the largest chunks of the gap, but that is to be expected. Those DVOAs have a higher variation in general than rushing offense, rushing defense, or special teams. Any discrepancy in passing will have a magnified effect. As far as I can tell, the main issue is just that there are nuances that are revealed by the play-by-play data that simply are not captured in the aggregate, season long statistics I am using. I’m open to suggestions in the comments for improvements I could make to the method. Nevertheless, four of the top five teams are the same in both actual and estimated DVOA, so I’m pretty satisfied with the results overall.

Andreas Shepard is a strategy consultant and long-time reader of Football Outsiders. He lives and works in Chicago, where he is (perhaps foolishly) optimistic about Jay Cutler’s new contract.

Posted by: Guest on 06 Feb 2014

140 comments, Last at 02 Apr 2014, 3:52pm by justanothersteve

Comments

1
by tuluse :: Thu, 02/06/2014 - 3:21pm

Nice article.

One thing that really surprises me is that none of the LT lead Giant's defenses made it in the top 20, and honestly that a few more of the Bears 80s defenses didn't.

6
by Andreas Shepard :: Thu, 02/06/2014 - 3:41pm

Thanks!

To answer your questions, the Giants actually never had the best defense in DVOA from what I can tell. Their best defense was actually in 1985, when they were second to the Bears with -16.4% DVOA.

The other Bears defenses of the mid-80s were very close to the top 20 (#22 in 86, #36 in 84). Their average DVOA of -24.7% is the 3rd best 3 year run ever. #1 is the 69-71 Purple People Eaters. #2 is the 90-92 Eagles, who are skewed by their absurd numbers in 91.

8
by tuluse :: Thu, 02/06/2014 - 3:50pm

That Buddy Ryan apparently knew a thing or two about coaching defense.

49
by Michael19531 :: Fri, 02/07/2014 - 2:03am

In 1991, Buddy Ryan wasn't coaching the Eagles. He was canned after the 1990 season. Bud Carson was the Eagles d coordinator in 1991 with Jeff Fisher as perhaps his secondary coach.

95
by Shattenjager :: Sat, 02/08/2014 - 2:41am

Fisher was defensive coordinator for the Rams in 1991. He was the Eagles DC under Ryan in '89 and '90.

2
by JIPanick :: Thu, 02/06/2014 - 3:31pm

This is awesome, although I can't really comment on methodology.

Is there any way I get my hands on your estimates for teams that aren't listed here (not top 20 in something, didn't win championships)? In particular, I'd really like to see where the '78 Cowboys land.

4
by Karl Cuba :: Thu, 02/06/2014 - 3:39pm

Seconded on the request for more of the results.

11
by Andreas Shepard :: Thu, 02/06/2014 - 4:01pm

The 78 Cowboys were #1 in DVOA that year (75th all time), behind some great balance. Their offense was #1 overall, #1 passing and #3 rushing. And their defense was #3 overall, #2 against the pass and #6 against the run. Special teams were a mediocre #22 though.

3
by Karl Cuba :: Thu, 02/06/2014 - 3:38pm

Good work.

5
by Vincent Verhei :: Thu, 02/06/2014 - 3:39pm

I’ll mostly leave the commentary to you folks, but one word on the 1970 Baltimore Colts, who come in as the only team to win the Super Bowl with negative DVOA. On the surface, they seem like a strong team, with an 11-2-1 record. However, they had the point differential of a team that would be expected to win only 9.5 games. Furthermore, they faced a weak schedule: their VOA of 5% is 8% higher than their DVOA of -3%. Add those together, and Baltimore was significantly weaker than its record.

I've never looked at this before, but this was the first year after the AFL merger, and man can you see the disparity between the two leagues. Using the Simple Rating System at Pro Football Reference, the top nine teams in the league all played in the NFC. And Baltimore was only fifth in the AFC!

7
by JIPanick :: Thu, 02/06/2014 - 3:42pm

Makes me wonder, was any adjustment for AFL vs NFL in the last four years of the 60s?

15
by Andreas Shepard :: Thu, 02/06/2014 - 4:19pm

It does not include an NFL-AFL adjustment. I played around with it (using some great work that Jason Lisk did for the old PFR blog, link below), but I couldn't find something that I was happy with. One difficulty is that DVOA in each season averages to zero by design, so any penalty on the AFL teams gets balanced out by a corresponding boost on the NFL teams. And it isn't really satisfying to give "credit" to the NFL teams for staying roughly the same quality while a bunch of worse teams joined and brought the average down. I'm open to better ways of implementing an adjustment though.

http://www.pro-football-reference.com/blog/?cat=39

57
by nat :: Fri, 02/07/2014 - 8:29am

NFL-AFL is really two problems.

The first is that they were separate sets of teams that didn't play each other. Putting them in the same list is like comparing a high school division to the SEC, although not that extreme. You really can't justify treating the stats as equal.

The second is that the AFL-NFL merger was the single biggest league expansion. Suddenly a 0% DVOA moved from meaning "about the eighth best team" to "about the thirteenth best team". This problem also happens when comparing the current 32 team league to the 1970s' smaller leagues.

Merely calculating a DVOA for the top ten (or whatever number) teams each year would help in comparing teams from different size leagues. It would let you say "Team X was 15% over league average, but about average for a top ten team".

Combining that with an AFL adjustment to DVOA would make it clear that an NFL team didn't suddenly get better by including AFL stats. You might even begin to show how the AFL siphoned some talent away from the NFL as it matured.

62
by nat :: Fri, 02/07/2014 - 10:40am

Another way to look at league expansion is to look at the DVOA dilution each expansion year. For the three expansions that we have DVOA for, we can look at the average DVOA of the established teams vs. the average of the whole league. The difference is the amount those teams' DVOAs benefited from the dilution of the league, even after losing players in the expansion drafts and position in the regular draft.


Year Teams.. Dilution
1995 CAR,JAC 1.27%
1999 CLE.... 1.32%
2002 HOU.... 1.36%

Multiplied out, the cumulative dilution between 1994 and today is almost exactly 4.00% when measure this way.

Compared to the 14% increase in the number of teams, this smaller than expected dilution is a testament to the league's handling of expansion drafts to create viable new teams.

79
by Jerry :: Fri, 02/07/2014 - 7:21pm

One quibble: does the dilution die away as 5he expansion team improves?

82
by nat :: Fri, 02/07/2014 - 8:33pm

I wouldn't think so. The talent just shifts among the teams. But you still have the larger number of rosters to fill. Instead of a very weak expansion team, you end up with all teams on average being slightly weaker than they would have been without that extra team using up some of the available talent.

I haven't really worked this all out in detail. But it seems reasonable, doesn't it? When you expand the league, about sixty guys get into the league who would not have been good enough the previous year, as well as some weaker coaches. That's got to bring the average level of play down.

83
by tuluse :: Fri, 02/07/2014 - 8:44pm

Well there are other factors to consider. The population of players and coaches is growing as well, so the average level of play should go up over time.

Also, players and coaches are not equally distributed over the league. So if the effect stays, it could only have an effect on a few teams.

9
by Sisyphus :: Thu, 02/06/2014 - 3:55pm

This is vary interesting but the range for the comparison is a bit too broad and it is important to keep in mind that the comparison numbers are in relationship to the other teams in that year. The football of the 1950's and 1960's bears little relationship to the game of today. The rules were different and enforced somewhat differently as well. The offenses and defenses of those eras were constructed for the rules and other teams of their time. These teams from the early years of the "modern" NFL would have a difficult time staying on the field with most of the teams now under the current game rules. (Of course some of the best teams now might have to bring body bags along to play under the rules interpretations of the earlier periods, I am pretty sure having Butkus in his prime clothesline someone going across the middle would also deter most receivers from running crossing routes.)

12
by Rick_and_Roll :: Thu, 02/06/2014 - 4:01pm

Of course the modern era teams are better if they play head to head, but I think the way to think about things is to consider: how much better is the team than an average team of that year.

14
by Andreas Shepard :: Thu, 02/06/2014 - 4:18pm

This was my thinking also. In a perfect world we would be able to compare league quality between years and eras, but until we get there comparing relative performance in each year is a decent start.

10
by Andreas Shepard :: Thu, 02/06/2014 - 3:55pm

Let me think about whether there is a good way to post the results somewhere. There's a lot of data with 1600+ teams, so a giant data dump may not be useful.

In the meantime, I'm happy to look up specific teams for people.

48
by mehllageman56 :: Fri, 02/07/2014 - 1:53am

Great work, again. I was just curious about the 68 Colts, whether they had a higher DVOA than the Jets or Raiders from that year.

61
by Hurt Bones :: Fri, 02/07/2014 - 10:34am

Reread the Best Total DVOA 1950-2013 and you'll see that the '68 Colts, Cowboy and Chiefs all had higher DVOA than the Jets and Raiders, Yes, three teams from 1968 in the twenty best.

91
by mehllageman56 :: Sat, 02/08/2014 - 12:31am

Thanks, I kept missing that table when I was scrolling down.

69
by JIPanick :: Fri, 02/07/2014 - 12:42pm

"There's a lot of data with 1600+ teams, so a giant data dump may not be useful."

If you could get it in the right format, I think it would be.

13
by Rick_and_Roll :: Thu, 02/06/2014 - 4:06pm

Btw, nice work

16
by Hurt Bones :: Thu, 02/06/2014 - 4:25pm

Great article.

Minor nitpick

"So, as a bonus, here is every NFL champion since 1950, ranked by DVOA."

The 68 Jets and 69 Chiefs were not NFL champions, of course they won the Super Bowl. Both franchises have won a Super Bowl but never won an NFL championship.

54
by Whatev :: Fri, 02/07/2014 - 5:10am

You're gunning hard for the "Pedant of the Month" award, aren't you?

60
by Hurt Bones :: Fri, 02/07/2014 - 10:26am

It's February. I need to look forward to something.

17
by stevo :: Thu, 02/06/2014 - 4:26pm

Nice work. Your discussion on the '91 Eagles Defense gives me more insight that DVOA does more than adjusts for offense and defense in a vacuum (Offensive DVOA is adjusted for the quality of the opponent's defensive and vice versa for defense and offense). The fact that they had such a historically bad offense yet the defense gets the nod for best historic DVOA speaks volumes to the statistic.

18
by Bright Blue Shorts :: Thu, 02/06/2014 - 4:30pm

I wonder how many guesses it would take the average football fan to name the 1991 Redskins as the best team ever?

With Gibbs retiring there was no repeat as champions and of course those Cowboys teams that followed them got all the glory of the early 90s, and the Giants/49ers preceding them.

I'm guessing you'd name 25+ teams before suggesting the 1991 Redskins.

19
by Perfundle :: Thu, 02/06/2014 - 4:48pm

Seems to me the only reason for their relative anonymity is their lack of Super Bowl dominance, because they dominated everything else. First in scoring offense, second in scoring defense, 7th-largest point differential in NFL history (and of the six ahead of them, only the 1999 Rams actually won a Super Bowl), crushed their other two playoff games.

93
by SuperGrover :: Sat, 02/08/2014 - 1:41am

Not following. They dominated the Super Bowl in every way imaginable. The game was never in doubt, even after Art Monk's overturned TD turned into a muffed FG from the 8 yard line. The Skins dominated as the 37-10 score 5 minutes into the 4th shows.

94
by SuperGrover :: Sat, 02/08/2014 - 1:43am

Actually, looked at the PBP and the muffed FG was from the 2. Why the Hell didn't Gibbs go for it?

20
by dryheat :: Thu, 02/06/2014 - 5:11pm

Mark Rypien certainly loves this research. I do too, but for a different reason.

32
by tuluse :: Thu, 02/06/2014 - 7:48pm

I think they're mentioned fairly commonly as one of the most dominant teams of all time.

I do tend to hang out in unusual football circles though.

33
by JIPanick :: Thu, 02/06/2014 - 7:53pm

Coincides with my experience.

50
by Michael19531 :: Fri, 02/07/2014 - 2:14am

Eddie Epstein wrote a book, titled "Dominance" in which he ranked the best teams, using standard deviation, from 1950-2001. He rated the 85 Bears at #1 with the 91 Skins at #2.

I'm wonder where the 1983 Skins rank. During the season and into the playoffs, I, and many others, thought that that particular team was one of the greatest ever. Then they went out and layed a tyrannosaurus sized egg in the Super Bowl and were smoked by the Raiders 38-9. Despite the loss, I still believe that that 83 Skins team was just as good as any team I've ever seen. But that's just my subjective opinion.

Looking back, I wonder how the 1970 Vikings stack up. A blogger at Pro Football Reference.com argued that if the Vikings had had Fran Tarkenton at qb in 69 & 70, the Vikes would've won titles in both seasons and the "Purple People Eaters" defense would be considered the greatest D of all time. But because of their Super Bowl loss to KC and division round loss to San Fran in 1970, they probably don't get the credit that they deserve.

55
by Bright Blue Shorts :: Fri, 02/07/2014 - 6:24am

The 83 'Skins do feel like they're missing from these lists.

They were 'only' a blowout away from become back-to-back SB champions
They had a 14-2 regular season with their two losses by 1-pt each
Their offense was the highest scoring of all-time and continued to be until 98 Vikings

But the best stat about the 83 Redskins is their turnover differential ... it was something like +44 ... 61takeaways, 17 giveaways

I'm not sure whether it is a record but few teams have been close to that. This year for example Seattle led at +20

59
by Travis :: Fri, 02/07/2014 - 10:15am

The Redskins' +43 turnover differential is a record, and there's no one close. The runners-up are the 2010 Patriots and 2011 49ers, both at +28.

63
by Andreas Shepard :: Fri, 02/07/2014 - 11:22am

They were #1 in DVOA that year (29.4%), but "only" 80th all time. The offense was very good (23.3% - #1 overall, #2 passing, #1 rushing), but "only" 64th all time. The defense was 5th in the league, but the DVOA was not that impressive (-7.8%).

70
by Michael19531 :: Fri, 02/07/2014 - 3:52pm

Thanks for the feedback Andreas.

64
by Andreas Shepard :: Fri, 02/07/2014 - 11:29am

See #63 for WAS 83

The 1970 Vikings were actually not that great overall. #5 in DVOA that year, although with the 25th best defense ever. The 1969 Vikings were substantially better. They just missed the top 20 overall DVOA list, coming in at #22 (although they had the misfortune of playing #20 in the super bowl). And of course they had an even more fantastic defense in 69 than they had in 70.

Both those teams were held back by relatively average offenses, although their passing offense DVOAs were in the 7-10% range. The Fran Tarkenton Giants had a passing DVOA of 14% in 1969 and 25.5% in 1970 though, so yeah he probably would have helped.

67
by Will Allen :: Fri, 02/07/2014 - 12:11pm

Joe Kapp, for his obvious limitations, could do some things to extend a play and move the chains. The guys they had for qbs in '70 and '71, Cuozzo, Lee, and Snead (71) were really bad, although by '71 the issue was as well that the running backs were getting old, in an era where running backs were far more important.

I know DVOA doesn't show it, but I think if you could compare the number of three and outs the Vikings offense had in '70 and '71, compared to '69, the gap between the Vikings defensive performance in '69, and that of '70 and '71, may be less than one thinks.

80
by Jerry :: Fri, 02/07/2014 - 7:26pm

Wow. Gary Cuozzo. I'm remembering when he was supposed to be the next great quarterback for whoever got him from the Colts. (Not unlike a Matt Schaub or Kevin Kolb.)

68
by Will Allen :: Fri, 02/07/2014 - 12:28pm

Good grief, this article is now bringing back childhood nightmares. Look at this....

http://www.pro-football-reference.com/boxscores/197112250min.htm

boxscore from the Vikings divisional round loss to the eventual dominant Super Bowl victor Cowboys. They outgained the Cowboys 311 to 183, get 17 first downs to the Cowboys 10, but Cuozzo and Lee toss four ints to Staubach's zero, and the Vikings lose 20-12. The previous season, the Vikings had beaten, in the regular season, the eventual Super Bowl ot loser Cowboys 54-13.

It was great to be a kid and a Vikings fan from '69 through '76, but there sure were some bitter ends to seasons, without any really joyous ones.

71
by Michael19531 :: Fri, 02/07/2014 - 4:00pm

Thanks for the info about the 70 Vikes Andreas. The biggest reason that I was so curious about where the 70 Vikings ranked is because they hammered the NFC champ Cowboys 54-13. Outside of some of the recent Tom Brady (2010 vs the Jets and Bears come to mind) and the 89 Joe Montana Niners playoff games,, that 1970 game had to be one of the most comphrehensive ass kickings one good team ever gave to another.

74
by Bright Blue Shorts :: Fri, 02/07/2014 - 5:01pm

At the time it occurred the 1985 Bears 44-0 shutout of the Cowboys was rather astounding. In retrospect it could even be marked as the moment when Landry's empire began to crumble

72
by Michael19531 :: Fri, 02/07/2014 - 4:04pm

Who was #1 overall in 1970? I would've bet my paycheck that it would've been the Vikings that year. I know the Lions were good that year too, but not as good as the Vikes.

75
by Andreas Shepard :: Fri, 02/07/2014 - 5:07pm

Detroit actually comes out on top that year. Part of the reason they might not have "felt" as good is a tough schedule. They had a DVOA of 30.3%, but a VOA of only 23.7%. They were also very well-balanced - top 6 in passing and rushing offense, passing and rushing defense, and special teams. The Rams, 49ers, and Cowboys fill out the rest of the top 4 ahead of the Vikings.

That Vikings-Cowboys game was interesting (box score: http://www.pro-football-reference.com/boxscores/197010180min.htm). Dallas actually outgained Minnesota on the day, but they threw 4 INTs and had a punt blocked, and 2 of those got returned for touchdowns. So the DVOA difference in that game is probably smaller than the scoreboard would suggest.

76
by Will Allen :: Fri, 02/07/2014 - 5:18pm

The Vikings had a market advantage in the Bud Grant years, in terms of working on blocking kicks. Eventually, special teams coaching improved throughout the league, and blocking kicks became much more difficult, but for a long time the Vikings would gain victories by blocking the opponent's kicks.

77
by Will Allen :: Fri, 02/07/2014 - 5:24pm

Well, the Cowboys one td that day came on a 4th quarter drive while trailing 54-6, so DVOA may not be as instructive as usual.

86
by buzzorhowl :: Fri, 02/07/2014 - 11:29pm

As a Redskins fan since the early 80s I would blow your averages all to hell, as would pretty much all of my relatives.

But on the whole, you're probably right.

92
by SuperGrover :: Sat, 02/08/2014 - 1:38am

A few things:

1. Gibbs didn't retire after the 1991 championship, he retired after the Skins lost at SF in the 1992 playoffs with a Brian Mitchell fumble. That team was decimated by injuries.

2. As a Skins fan, I biasedly state the '91 Skins were the best team I've ever seen. The '85 Bears were close. That Skins team lost one competitive game at home to the Cowboys. That game included an onside kick, a 3rd and forever draw play for a TD, and a couple non-Skins plays that gave the Cowboys the win. The other loss was the last game of the year against Philly in which they had a 19-7 at the end of 3, rested starters, and saw that lead disappear. Keep in mind, that was a game on the road against the best defense in DVOA history.

3. The amazing thing about the Skins season was how balanced they were in all aspects of the game. The defense was very good, relying upon situational players such as Matt Millen, Kurt Gouveia, Jumpy Geathers and Monte Coleman. On offense, they had a solid receiving corps, a deep set of RBs for all scenarios (Byner/Ervins/Riggs) and a QB who was superb when given time. The ST's were also solid with real skill at kicker and PR/KR.

4. The other thing I remember from that team was that they were exceptionally healthy. The starting OL missed only 3 starts with Russ Grimm and Ed Simmons filling in ably when needed. That is exceptional considering the starting OL averaged almost 29 years of age and Grimm and Simmons were 32 and 28, respectively. In fact, the Skins started 14 players aged 28 years or older, 6 of whom made the Pro Bowl. They were an exceptionally old team that managed to stay healthy an entire season.

Great time to be a Skins fan. In fact, the last truly great season to be a Skins fan (as mentioned, the next season was a nightmare as age and injury caught up and we fumbled away a chance at a time game in Candlestick).

21
by TomKelso :: Thu, 02/06/2014 - 5:41pm

Footnote about the 1952 Dallas Texans -- they were the relocated New York (formerly Boston) Yanks, and although the franchise is considered voided by the NFL, its assets -- mostly player contracts -- were granted to the relaunch of the Baltimore Colts in 1953. That's why Artie Donovan had those stories of throwing elephant poop at the opponents to tell David Letterman -- he was a Texan before he got awarded to the Colts. The AAFC Colts (BAL1 in your listing) had gone out of business due to bad financing.

Interestingly enough, the band stayed together during this first interruption, too, and their incorporation as a separate entity is what blocked the Irsays' attempts at suing the band for trademark infringement during the NFL's version of the Babylonian Captivity.

29
by Andreas Shepard :: Thu, 02/06/2014 - 7:11pm

Learning about the insane early history of the NFL was unquestionably my favorite unexpected find from this project. A couple more fun facts about the Texans: they won their only game when they played "at home" in Akron on Thanksgiving against the Bears, who started their second string. Only 3000 people watched the game, fewer than attended a high school game earlier in the day. The Texans coach suggested that in lieu of pregame introductions they should go into the stands and shake hands with each fan.

And those Baltimore Colts that folded in 1950? They had originally started in 1946 as the Miami Seahawks, of all things. That year they were in the league against the Brooklyn Dodgers, the LA Dons, the Chicago Rockets (who became the Hornets in 1949), version 1 of the Buffalo Bills, and both the New York Yankees and the Boston Yanks. The NFL has been recycling team names for a long time.

87
by buzzorhowl :: Fri, 02/07/2014 - 11:32pm

Wow, this is great stuff. Thanks to both of you.

110
by Boballab :: Thu, 02/13/2014 - 4:45am

Good thing you didn't go back to the 1940's era or you would have run into the merged teams such as the Steagles (Steelers/Eagles), who traded owners back and forth (Yes at one time Art Rooney didn't own the Steelers but the Eagles), and the CarPit (Cardinals/Pittsburgh). However for the fun read up on that trading ownership between Bert Bell and Rooney its a classic.

120
by Jerry :: Thu, 02/13/2014 - 7:12pm

Since I just wrote this up for somebody (it's from memory):

In 1940 (I think), Art Rooney sold the team to a guy named Alexis
Thompson and bought half the Eagles from Bert Bell. According to
either Dan or Art Rooney Jr.'s autobiography, Thompson was going to
move his franchise to Boston, and the Eagles would become the
Keystoners, with homes in Philly and Pittsburgh. The other owners
didn't like that idea, though, so Thompson and Rooney/Bell ended up
swapping franchises. That means that the team in Pittsburgh is the
orgiinal Philadelphia franchise, and vice versa. There has been a
team owned by Art Rooney or his heirs in Pittsburgh every year since
1933.

22
by Will Allen :: Thu, 02/06/2014 - 6:10pm

Like most cliches "defense wins championships" has a kernel of truth, and looking at these lists shines some light on that.

23
by Will Allen :: Thu, 02/06/2014 - 6:29pm

The Super Bowl in January 1970 featured two historically great defenses. The difference in what was a one sided game was mostly 1)the Chiefs had one of the first great soccer style kickers, Jan Stenerud 2) the Vikings fumbled on a kick return and 3)the Chiefs had a gteat HOF qb, Len Dawson, and the Vikings qb was noted tequila aficionado, and appreciator of post game college band performances, Joe Kapp.

24
by Vincent Verhei :: Thu, 02/06/2014 - 6:35pm

He also, in his 70s, was still picking a fistfight in public with former teammate and former pro wrestler Angelo Mosca.

25
by Will Allen :: Thu, 02/06/2014 - 6:51pm

My favorite Joe Kapp story concerns the players having their typical Sunday night festivities, after the Sunday afternoon game, in this case, a loss by a close score. Middle linebacker Lonnie Warwick, properly consoling the loss via beverages, insists that it was his poor play which was cause for the disappointment, and Kapp, also taking the sensible approach to unhappy occasions, informs Lonnie that he is in error, and that it was he, the selfless, saddened qb, who was responsible for the loss. Each persists. It ends with Kapp and Warwick on the front lawn of whomever was lucky enough to be hosting this interesting house party, exchanging staggering, inebriated, roundhouse hooks and uppercuts at each other, widely missing, of course, until they both were on their *sses from exhaustion.

They just don't make'm like that anymore. A shame.

88
by buzzorhowl :: Fri, 02/07/2014 - 11:38pm

Another fun Joe Kapp fact, of course, is that he was the last guy before Peyton did it earlier this year to throw 7 touchdowns in one game. In the wake of that game, Tanier wrote an amazing article about him at Sports On Earth that had me on youtube looking up Joe Kapp "highlight" montages. Definitely need the scare quotes for that one. My favorite Viking-era Kapp highlight was a two-hand pass/shove that looked more like a basketball chest pass than anything a football quarterback should do. (It's at the :35 mark in this video: https://www.youtube.com/watch?v=NiVlcjWBj00)

Joe Kapp is a legend.

89
by tuluse :: Fri, 02/07/2014 - 11:52pm

Joe Kapp hightlights are the best. I could watch him all day.

It's also kind of wild watching #84 streaking down the sideline and outjumping defenders on poorly thrown balls. I feel like this is what Tebow to Randy Moss would have looked like.

90
by Will Allen :: Sat, 02/08/2014 - 12:22am

That Joe Kapp 7 td game was the first Vikings game I ever attended, as a small child. I was hooked from then on.

He once knocked a Cleveland Brown linebacker out cold, in the 1969 NFL Championship Game, when he was scrambling for yardage, the lb came in waist-high to make the tackle, and Kapp blasted him right on the chin with his knee.

96
by Raiderjoe :: Sat, 02/08/2014 - 8:44am

Yrah, he knocked out jim houston . It eas gerst how nfl filmd filmed it

26
by Dan :: Thu, 02/06/2014 - 6:54pm

Whoa, this is really cool.

One thing to keep in mind is that, any time you estimate a quantity with a regression, the estimated values will tend to be less extreme than the actual values (that's regression to the mean). You can see that in the last table, where the top 20 actual DVOAs since 1989 are higher than the top 20 estimated DVOAs since 1989.

This effect is most noticeable when you look at the very top or the very bottom of the distribution, which is presumably why recent teams (since 1989) are overrepresented on the all-time lists (best and worst teams since 1950). Although with a correlation of 0.958 it's not that big an effect.

31
by Andreas Shepard :: Thu, 02/06/2014 - 7:25pm

I actually scaled my estimates so their standard deviations matched the actual standard deviations of DVOA, for exactly this reason. So I think variation in actual performance has become variable since 1989. Why this is happening I'm not sure.

One factor might be the rule changes in the passing game, which have allowed the best quarterbacks to put up crazy good numbers while the worst quarterbacks aren't performing that much better. That would increase the performance variation.

Another reason might be free agency. Since there are more player transactions each year, that will mean that some teams will roll the dice on a bunch of free agents and have them all pan out (like the Seahawks and their incredible DLine this year), while some teams will have all their gambles fail or end up in cap hell (like the Raiders and the giant clump of dead money they dealt with this year). This again will cause their to be more spread in team quality in a given year.

45
by dmb :: Thu, 02/06/2014 - 11:57pm

I actually scaled my estimates so their standard deviations matched the actual standard deviations of DVOA, for exactly this reason.

I had the same thought as Dan, but it turns out that you were even more rigorous than I thought!

28
by dbt :: Thu, 02/06/2014 - 7:09pm

1985 Bears, I want to see weighted DVOA at the end of the season and the end of the postseason. The Bears famously scuffled the first couple weeks of the year, needing comebacks against Tampa and Minnesota, and of course dominated in the postseason.

30
by Will Allen :: Thu, 02/06/2014 - 7:21pm

Again, I think it is notable that two of the top twenty offenses were part of a team that won a championship, whereas seven of the top twenty defenses had that honor.

34
by Andreas Shepard :: Thu, 02/06/2014 - 7:56pm

I wouldn't read too much into that - I think this is actually just an artifact of where the cutoff was drawn.

If you extend both lists down to 50, you add 8 more offenses that won the championship (DAL 95, SF 84, IND 06, GB 61, WAS 91, SF 89, OAK 76, DAL 77) and only 3 defenses (BAL 2000, WAS 91, GB 62). That leaves you with 9 top 50 offense champions, 9 top 50 defense champions, and WAS 91, who is on both lists.

For what it's worth, the "average" champion has a DVOA of 25.3%: 13.2% on offense, -10.0% on defense, and 2.1% on special teams.

35
by tuluse :: Thu, 02/06/2014 - 7:56pm

Want to give us the medians too?

40
by Andreas Shepard :: Thu, 02/06/2014 - 8:41pm

They are all within a percentage point of the averages.

43
by tuluse :: Thu, 02/06/2014 - 9:33pm

Cool. Thanks.

36
by Will Allen :: Thu, 02/06/2014 - 7:58pm

Yeah, I'm not sure what, if anything, to make of it just yet.

37
by JIPanick :: Thu, 02/06/2014 - 8:01pm

The list of champions has:
8 below average offenses
13 below average defenses (1964 CLE... what?)
16 below average special teams
by my count.

38
by Will Allen :: Thu, 02/06/2014 - 8:09pm

My wild guess is that Jim Brown kept the 1964 Cleveland defense from having to be on the field very much.

39
by Andreas Shepard :: Thu, 02/06/2014 - 8:41pm

The 1964 Browns are interesting. While they weren't a great team, their division was pretty terrible, so they made it to the NFL championship game with a 10-3-1 record. There they somehow shutout the Colts 27-0 even though the Colts were the best team in the league by DVOA. The Browns were #1 in both rushing offense and #2 in passing offense that year, and while their defense was bad they certainly played well in a big moment.

58
by Jimmy :: Fri, 02/07/2014 - 9:07am

They were also missing their starting QB for a good portion of the year (and Steve Fuller sucked), they were incredibly unlucky in the only game they lost that year though. I can't quite believe that the special teams rating is positive, I watched half the season on YouTube a few months ago and that team couldn't play specials at all. That would have been my guess for why they wouldn't have been ranked top.

41
by Cythammer :: Thu, 02/06/2014 - 8:45pm

A little off topic, but that article about the 1981 Colts (which was published in 2007) included the following reader comment, which I found quite amusing considering how the 2007 season played out: "Does anyone else think that this sort of thing could happen to the Giants this year with their lame duck coach?"

42
by jamie_k74 :: Thu, 02/06/2014 - 9:24pm

I suppose it's the kind of thing that would occur many times, but the one that jumps out to me is the 2000 Steelers. 17th best rushing offence ever, but a 9-7 finish and no playoffs. I guess that's what happens when your starting QBs are Kent Graham & Kordell Stewart.

44
by Will Allen :: Thu, 02/06/2014 - 10:45pm

It's really remarkable that the 8th worst defense, the 2000 Vikings, made it all the way to the Conference Championship. Where they were beat 41-0.

46
by andrew :: Fri, 02/07/2014 - 12:41am

They weren't that historically bad until the last 3 weeks of the season when they got absolutely obliterated, backing in to the playoffs. Having fallen from 11-2 to 11-5 still got them a bye, and Moss was enough to keep them in front of the Saints... (they had beaten the Rams, and I think the Rams would have been a much worse first round opponent for the Vikings).

That being said I'm trying to remember if they had some injury at the point they fell off a cliff, or if it was just playing with leads, smoke and mirrors and playing competent offenses the last three weeks.

I do think Robert Smith was dinged if memory serves and he had been having an absolutely terrific year to that point, and wasn't himself down the stretch. That may have helped mask the defensive woes...

47
by Will Allen :: Fri, 02/07/2014 - 1:00am

It is crazy to consider how close they came to getting the #1 seed. I can't remember the tiebreakers, but if they hadn't lost a close game to the Packers in week 16, I think they would have edged into the first spot.

51
by bubqr :: Fri, 02/07/2014 - 2:18am

Very cool article. Thanks for that.

52
by The Ninjalectual :: Fri, 02/07/2014 - 4:15am

The 1991 Redskins were my first introduction to football. I had no idea what I was watching at the time!

105
by DRohan :: Mon, 02/10/2014 - 12:02am

I had a similar experience in late 70s Pittsburgh. In the first 12 months I really followed sports, the Steelers won 2 Super Bowls and the Pirates won the World Series. Impossible to appreciate at the time. I thought the "good guys" always won.

107
by justanothersteve :: Mon, 02/10/2014 - 1:47pm

That was also my experience as someone born in Green Bay in the mid-50's and being in grade school during the 60's Lombardi era.

53
by Q :: Fri, 02/07/2014 - 4:37am

96 GB was an absolute Buzz Saw. I believe Epstein's Book on Dominance had them #3 All Time and were #1 in regards to Standard Deviation among all teams.

I always wonder how much higher their rating would be if they didn't lose 3 out of their Top 4 receivers which caused a mid season losing streak where they had to start Keith Jackson at WR lol

73
by Michael19531 :: Fri, 02/07/2014 - 4:08pm

The 96 Packers are IMO, the most underrated team in history. They were great in all 3 phases. They were one of the few teams that I've seen where I knew from almost day 1 that they were going all the way.

56
by ammek :: Fri, 02/07/2014 - 7:19am

The history of variance in the NFL is an interesting topic that needs researching. In today's game, pass offense accounts for the most important part of variance. These numbers suggest that, in previous times, variance might have been more significant on defense.

Regarding the Packer championship teams, the general consensus among Packer fans is that Lombardi's 1967 team was easily the weakest of the seven. If these numbers are accurate, I think that goes to show the extent to which we underestimate special teams. The combined offense and defense of that team is nothing extraordinary, but special teams — essentially, Travis Williams — elevates it well above both the 1965 (a bit of a fluke?) and 2010 champions.

None of the three Raiders' championship teams comes out well from these numbers. I'm surprised, in particular, that the 1980 team was just average: they only lost two games by more than 6 points, they led the league in takeaways, and the AFC West that year looked about average, with two 11-win teams, two 8-win teams, and the 4-12 Seahawks. SRS judges their offense to have played better than their defense. Probably, if Football Outsiders had been around in the early-to-mid-1980s, it would have been accused of east-coast bias or irrational disrespect for Al Davis.

I'd be interested to know which was the worst Superbowl loser ever. Were the 1979 Rams worse than the 2008 Cardinals? Were the 1985 Patriots even average?

65
by clark :: Fri, 02/07/2014 - 11:57am

The 1980 Raiders looked like a fluke at the time, so I expected to see them near the bottom. They had missed the playoffs the previous two years and were below .500 in 1979. They would be below .500 in 1981. I remember the '76 and '83 Raiders as great teams, but they played a lot of sloppy football. The '76 team in particular had a long string of close wins early in the year. I remember that halfway through the season they had given up more points then they had scored... and they were 6-1.

66
by Jeff M. :: Fri, 02/07/2014 - 12:07pm

Interesting to think, with all of the "will the Seahawks be able to turn into a dynasty/mini-dynasty" that's been written, that by this metric they quite possibly already are. 2012-2013 Hawks are the third team (after the early 60s Packers and late 60s Chiefs) to repeat in the top 20.

I think it's mainly because outside of the advanced statistical community you don't see a lot of discussion of the 2012 team as even the top team of that season, much less a top 20 team all time. It'll be interesting to see, if they continue this level of performance, whether that team gets a reevaluation over time in the popular media.

84
by jacobk :: Fri, 02/07/2014 - 10:25pm

In my wildest dreams the 2012 Seahawks will eventually be seen as the 89-90 Chicago Bulls of football. The narrative would have been better if Atlanta had missed the field goal and the 49ers had been the ones to knock them out, though.

101
by GoDog :: Sun, 02/09/2014 - 2:53am

The Seahawks may or may not be on the cusp of a dynasty, but they do seem to have the right idea in their approach to the game. The top ten of all-time great combination DVOA's for the most part have punishing defenses, power running, and enough passing offense to counter the run. The '91 Redskins, '85 Bears, GB 1961-1962, even most of the great Steeler teams were of this mode. Does this speak well about this approach over a dominating offense with a so-so defense? Could be.

139
by cjfarls :: Tue, 03/11/2014 - 3:48pm

"Dynasty" is such a loaded term and I also think folks are a bit too quick to crown SEA such.

Seattle has had a great DEF for 2 years, but DEF historically is more difficult to keep dominant over long-periods. Add the fact that they will lose the advantage of having a uber-cheap QB, their top CB is gonna want to get paid, and their primary offensive weapon is an aging RB that could fall off the cliff anyday... one knee blowout and suddenly SEA is dependent on Wilson to carry the offense. Just too many and too fragile of parts when you rely on DEF dominance (where the OFF can dictate matchups against the weakest link) and running game (RB is the shortest lifespan position).

SEA certainly looks to be a good to dominant team next year and for the next few years... but justa few years ago we were saying the same thing about GB and they didn't even make the playoffs.

Such an approach could be a good way to get a dominant team for a short period, but to me the most likely path to long-term team success really remains finding a HOF QB and riding them to the playoffs for a decade-plus. If/when you can get a really good DEF to go with them, you can win a championship or 2.

141
by justanothersteve :: Wed, 04/02/2014 - 3:52pm

I normally don't comment this late after another comment. But GB has made the playoffs 2009-2013. They lost their first playoff game after a 15-1 regular season in 2011, the year after they won the Super Bowl, but that was to the eventual SB champ. I agree that dynasties are going to continue to be rare.

78
by pm :: Fri, 02/07/2014 - 5:30pm

Andreas, do you run these stat correlations on a game to game basis or do you run them using full season totals? If its the latter, couldn't you improve the ratings by doing them at a game to game level?

97
by Andreas Shepard :: Sat, 02/08/2014 - 7:29pm

These are based on full season totals. Game by game stats would definitely be an improvement, although probably not too dramatic. The biggest benefits would be a) increasing the sample size for the regressions, and b) making the opponent adjustments more precise (right now I effectively assume that teams have the same number of passing and running plays in each game). Neither of these would likely change the answer much, but they would increase the precision.

The only reason I haven't done this is I haven't been able to get the detailed stats I would need at a game-by-game level. Things like rushing and passing first downs are really tricky to find for individual games all the way back to 1950. (There's also the issue of it making my excel file even more unwieldy. Instead of 1638 rows for each team-season, I'd need 24586 rows for each team game. Since I already have 350+ columns, I am afraid that excel would go on strike.) In a perfect world though, yes I'd do it game-by-game, and include the playoffs too.

104
by Joseph :: Sun, 02/09/2014 - 11:52pm

First, Andreas, great work.
Prob. you could come out with a follow-up article in several months that details some of the changes in the data that would happen by adding the stats, at least for the teams/games/seasons for which you could obtain them.
Re: the Excel prob., couldn't you separate each decade into its own separate spreadsheet? Or do you already have a file for each season or two? Cutting and pasting the formulas is pretty easy, so I would think it isn't much of a problem to split the data into multiple files.
Also, considering the further back you go, the less data you prob. have/can obtain, you may only be able to go game-by-game back to the merger at best. I'm guessing that will make the "project" easier. Good luck--your work is definitely appreciated here.

81
by zlionsfan :: Fri, 02/07/2014 - 7:51pm

Finally, a list that has champion Lions teams in it. (BTW, the 1952 Lions should be italicized.)

Based on your comment about the '70 Lions, it seems like several Detroit teams just had, more than anything else, bad timing. The team closest to the Super Bowl ended up losing to, by this measure, the best team in NFL history. The '70 team had a really tough schedule (on p-f-r, the second-toughest in Detroit history, but well behind the '65 team). The '54 team ran into a much stronger Cleveland squad. The '62 team split with what turned out to be the third-best team in DVOA (estimated) history.

Of course, for most of the 60+ seasons, they were poorly run, poorly coached, or both ... but the times they weren't, they just turned out not to be the best. (Except for Alternate Universe 1, the '50s, when Detroit and Cleveland were powerhouses.)

85
by Will Allen :: Fri, 02/07/2014 - 10:57pm

Alex Karras, Roger Brown, Joe Schmidt, Dick LeBeau, Night Train Lane, and Co. had some huge days against Lombardi's Packers.

108
by LionInAZ :: Mon, 02/10/2014 - 5:22pm

I was very sad when Joe Schmidt resigned as Lions head coach. That was probably the best-run team in Lions history. The franchise never really recovered, even through the Barry Sanders years.

98
by Geezerman :: Sun, 02/09/2014 - 1:16am

Hey, the greatest defensive team in NFL history was coached by Rich Kotite.

Who knew?

102
by Will Allen :: Sun, 02/09/2014 - 10:06am

Well, the defense was coached by Bud Carson, who was a great, great, coordinator.

99
by Patrezi :: Sun, 02/09/2014 - 1:47am

I didn't see this but do you take into account the difficulty of the opposition? For instance so much was made this year of the Broncos' record setting season but they really had such an easy season, when you take that into account, they really were merely a very good offensive team but for instance, New Orleans and Drew Brees put up a few hundred fewer passing yards but had played several top 6 defensive teams so it was so mucn more difficult to get passing yards than it was for Denver. Really had Denver played in the NFC West or South, most likely Denver would not have made the playoffs.

100
by JIPanick :: Sun, 02/09/2014 - 2:10am

"Really had Denver played in the NFC West or South, most likely Denver would not have made the playoffs."

That's... rather unlikely.

103
by Andreas Shepard :: Sun, 02/09/2014 - 10:15am

Yep, I did include opponent adjustments. That's the primary reason, for example, that the 72 Dolphins and 99 Rams don't make the top 20. They would both be in the top 10 if you ignored their weak schedules, but accounting for that bumps them both to around #40.

106
by dmstorm22 :: Mon, 02/10/2014 - 1:13pm

Don't forget the other side was true. Those NFC West teams defensively played a very easy slate of offenses.

As evidenced in the Super Bowl, the Seahawks defense may have been more 'legit' than the Broncos offense, but the Seahawks feasted on the 2nd easiest slate of opposing offenses, while the Broncos faced the 3rd easiest slate of opposing defenses.

109
by Steve B :: Wed, 02/12/2014 - 3:03pm

Surprising to see the '78 Steelers as low as they are. Did the average rushing numbers they put up that season hurt them that much in the analysis? Really, I'm surprised their STs rating is even as high as it. Gerala (12 of 26 on FGs) was pretty bad that season.

Also interesting to see that two non-SB teams are the Steelers reps in the top 20 of the total DVOA. This will be another piece of evidence for those who feel that the '76 team was actually the best of those 70s Steelers teams.

111
by jimkiick21 :: Thu, 02/13/2014 - 11:54am

Thanks for posting...a lot of information to digest. I'm looking at the offensive ratings, and one thing that stands out to me is that the 1950 Rams appear on neither the top 20 offense list nor the top 20 passing list. While there is no 1950 Rams, the 6-5-1 Rams of 1954 appear to have the 8th best offense in history despite finishing 3rd in points that year.

The strangest numbers in the article are the rushing offense numbers. I don't claim to be a DVOA expert, but I would have to question the relevence of the stat based on the top 20 rushing offenses. Lots of weird things that perhaps a DVOA person can explain to me:

*Some of the top rushing offenses in history aren't on the list...1962 Packers, 1972 Dolphins, 1973 Bills, 1978 Patriots. The 72 Fins had 2960 yards, 26 TDs, and was 2nd in the NFL with a 4.8 average. The 62 Pack had 2460 yards, 36 (!) TDs, and led the NFL with a 4.7 average. I'm not seeing how even with a straight efficiency ranking tool that either team could finish outside the top 20.

*The 1952 Browns were their usual good selves. No Jim Brown yet and Motley was finishing up his career, leading the team with 444 yards and a TD. As a team, the Browns had 1786 yards (3rd in NFL) 12 TDs (5th), and a nice 4.5 average (1st). In 4 of their 12 games they failed to rush for 100 yards, but they are the 18th best rushing offense in history.

*The 1993 Niners are similar. No 1000 yard rusher (Watters had 950), a really good offense that finished 3rd in rushing with 2133 yards, they led the NFL in both TDs and average (26 & 4.6...Steve Young was the 2nd leading rusher with 407 yards and a nice 5.9 AVG). Seems like a fine season, but 6th best in NFL history?

*I don't think any football historian could have predicted the top rushing offense in NFL history being the 2000 Rams. 1843 rushing yards (17th in the NFL), a good 4.8 average (2nd in the NFL), and due to the strength of their offense, a fantastic 26 TDs (1st in the NFL). I'm guessing that average and TDs trump actual yards gained in these DVOA calculations, but then how do you explain teams like the 62 Packers and 72 Dolphins not being on the list at all?

113
by tuluse :: Thu, 02/13/2014 - 1:40pm

DVOA is a rate stat, so total production isn't factored in at all.

It accounts for down, distance, and field position. So a 2 yard run on 3rd and 1 is more valuable than a 2 yard run on 1st and 10.

It also accounts for strength of opponent. So a team like the '72 Dolphins that faced an extremely easy schedule gets their rating lowered.

114
by jimkiick21 :: Thu, 02/13/2014 - 1:56pm

But wouldn't it be logical for total production to be factored in to a rate stat, especially if there are so many other extemperaneous aspects already factored in?

So while the 1972 Dolphins having an easier schedule than, say, the 1952 Cleveland Browns is factored into the rushing offense ranking, shouldn't we also take into account that the Dolphins had to maintain their "rate" over 613 rushing attempts compared to the Browns' 394 rushing attempts?

I'd have to take a closer look at those top 20 rushing offenses, but it seems like very few of those teams actually led the league in rushing for that season. The 2000 Rams were 17th!

115
by tuluse :: Thu, 02/13/2014 - 2:05pm

It wouldn't be a rate stat if you factored in total production. Football Outsiders does produce another stat called DYAR that is a measure of total value, but it's used only for individuals not for teams or units.

This is a measure of how effective a certain aspect of a team was, not how often they used it.

116
by jimkiick21 :: Thu, 02/13/2014 - 2:30pm

Kind of a circular argument. I think even the most basic of rate stats, like a punt return average or a batting average, will have qualifiers so that a guy who has 1 return for 15 yards or 4 hits in 10 at bats isn't treated equally as a guy who has 30 punt returns or is batting .332 in 550 at bats.

I would think a team's rushing "rate stat effectiveness" would regress to the mean the more times they ran the ball. If you think the 2000 Rams rushing attack was more "effective" than, say, the 1962 Packers rushing attack, then there is probably something grossly wrong with the measurement of "effectiveness".

117
by tuluse :: Thu, 02/13/2014 - 3:01pm

I'm not saying it's a perfect stat. In fact I don't think such a thing exists. So DVOA does one thing, and I think it does it well. If you want a stat to do something else, well you're free to create a Usage and Defense-adjust Value Over Average.

118
by Vincent Verhei :: Thu, 02/13/2014 - 3:05pm

From the second paragraph, emphasis is mine:

"For each year from 1950 on, I collected total box score statistics for each team on offense, defense, and special teams (yards, touchdowns, interceptions, fumbles, sacks, first downs, field goals, return yards, etc.). As DVOA is a per-play metric, I calculated each team’s per-play average for each of these stats. I then normalized each stat against the league average for that season. To translate these normalized per play statistics to VOA, I ran a series of linear regressions using the data from 1989-2012 (one each for rushing, passing, FG/XP, kickoffs, kick returns, punts, and punt returns). After adding opponent adjustments, these ratings are combined into composite ratings for total offense, defense, special teams, and overall."

Certainly, if two teams have similar high rush offense DVOAs, but one has a lot more carries, then that team's rushing attack was probably more impressive and probably led to more wins. But that's not what is being measured here. This is simply which team fared best on a per-play basis.

It's also important to remember that everything pre-1989 here is an estimate of DVOA, not DVOA itself.

119
by jimkiick21 :: Thu, 02/13/2014 - 3:30pm

But its not "simply which team fared best on a per-play basis" if you are also factoring in schedule and league average...right? And wouldn't it be easier to fare better on fewer plays?

Am I missing something here?

121
by EricL :: Thu, 02/13/2014 - 7:18pm

But its not "simply which team fared best on a per-play basis" if you are also factoring in schedule and league average...right?

Yes, actually it is. That's the whole point. Taking the per-play averages, and putting them in the context of that team's league and opposition.

If you attempt to modify the averages by attempts, you're back to a counting stat. Not a rate stat. Which is fine, but it's not what DVOA is about.

122
by Jerry :: Thu, 02/13/2014 - 7:28pm

There are two possibilities:

1. In Andreas' admirable attempt to estimate DVOA from season totals, he got the estimates to a 95% correlation. It's not hard to imagine that extreme seasons would be more likely to diverge.

2. Maybe those seasons were less impressive (at least in a DVOA sense) than they look.

Until we get actual DVOA for old seasons, which may never happen, we won't know which applies.

124
by Andreas Shepard :: Thu, 02/13/2014 - 8:05pm

Checking the teams that JimKiick mentioned, here are there all-time rushing ranks (all out of 1638 team-seasons since 1950).

62 Packers: 38th (they were also 47th in '60, 48th in '61, and 53rd in '64. The Lombardi sweep was the real deal)
72 Dolphins: 145th (111th in '73)
73 Bills: 97th (to go with the 55th worst passing game ever. In 1975 the running game was 42nd all time)
78 Patriots: 76th

So all of the teams mentioned do show up as some of the best rushing offenses of all time, even if they aren't quite as high as one might have guessed. As others have mentioned, that's due to a combination of these stats measuring per-play rather than total value, and my numbers being an imperfect estimate of an imperfect stat.

The 2000 Rams, incidentally averaged a first down on over 29% of carries (10th best all time) and a touchdown on 6.8% of carries (4th best all time) while averaging 4.8 y/c against a pretty tough schedule of rushing defenses. That's a pretty great per-play resume. They had the 6th fewest carries in the league that year, which is what brings their totals down.

127
by Steve B :: Thu, 02/13/2014 - 9:16pm

So was it primarily the total number of attempts that kept those run offenses from placing higher than they did? I could also see opponent strength being a negative for the '72 Dolphins and '73 Bills.

132
by Vincent Verhei :: Sat, 02/15/2014 - 10:21pm

And wouldn't it be easier to fare better on fewer plays?

Yes. I mentioned this, specifically.

So, again, this is a per-play stat. If you want to dismiss it because it does not account for a high number of attempts, that's fine. You're welcome to do your own analysis on high-volume rushing teams and submit it as a guest column. This, after all, was a guest column. Hey, it's the off-season, we're always looking for content.

112
by Byron :: Thu, 02/13/2014 - 12:59pm

I wonder how this could be applied to the 1946-1949 seasons? I would like to see how this method considers the balance between the AAFC and NFL during those years.

123
by Andreas Shepard :: Thu, 02/13/2014 - 7:41pm

I'd love to do this, but the data gets really spotty when you go back that far. For example, not only do I not have first downs, fumbles, or sacks before 1950, I haven't even been able to find rushing attempts for the AAFC teams. That being said, I feel very comfortable going out on a limb and saying those Otto Graham Cleveland teams were quite good.

130
by Byron :: Fri, 02/14/2014 - 2:18pm

Well rushing attempts (re: numbers), fumbles and first downs are not too hard to find so I may be misunderstanding what you are looking for there. Neft & Cohen have some good team stats in this regard. The AAFC Record Manuals are also very good from a team perspective and pretty good from an individual perspective (they do not include longest rush for each player, which is annoying). I've also found the official linescores for the 1946-49 AAFC to be of some help.

I do understand that some information is very limited though. For example, there are no sack stats available although papers sometimes included a "loss while attempting to pass" statistic which could give an idea about sacks. (Sadly, I think only the San Francisco papers included this particular statistic).

125
by Steve B :: Thu, 02/13/2014 - 8:54pm

Andres,

Where did the '62 Packers, '72 Dolphins, '73 Bills and '78 Patriots place on the rushing offense list? All would seem to belong on any list of best single season rushing offenses. The '62 Packers, in particular, seem like a glaring omission given where the team placed in total DVOA and given, for one, that finished so far above league average that season in rushing TDs.

Also, what caused the '78 Steelers to place as relatively low as they did?

EDIT - Thanks, Andreas. What about the '78 Steelers?

128
by Andreas Shepard :: Thu, 02/13/2014 - 9:25pm

A couple of things weighed the Steelers down in 78. The defense declined from historically great in 73-76 (when they averaged -25.3% DVOA) to merely very good (-12.4%, 2nd in the league). On offense, their passing game was 3rd in the league (20.1%), but their running game was inefficient (-8.3%) and they ran A LOT. They dropped back to pass 405 times, but ran it 641 times, the 7th most ever.

126
by Steve B :: Thu, 02/13/2014 - 9:03pm

Another one I'm wondering about -

2000 Raiders run offense

It would seem that they'd be pretty close to the 2000 Steelers on the list. Did the Steelers edge them out primarily because they accounted for a higher % of the team's first downs and offensive TDs that season than the Raider run O did? Steelers' passing O was bottom 5 that season.

129
by Andreas Shepard :: Thu, 02/13/2014 - 9:33pm

So this one is actual DVOA, not my estimates (see http://www.footballoutsiders.com/stats/teamoff2000). The Raiders were 5th that year in rushing DVOA (170th all time).

Digging into it, while they led the league in yards, that was driven somewhat by being 3rd in the league in carries. On a per carry basis, they averaged the 3rd most yards, 2nd most TDs, and 5th most first downs. So quite good, but not an all time rushing team.

The Steelers that year quite of bit of bonus credit for playing a tough schedule of run defenses.

131
by Michael19531 :: Sat, 02/15/2014 - 2:11am

Hey Andreas. If you have a little time, could you post the estimated DVOA of the following teams?

1. The 1973-1980 Chuck Knox/Ray Malavasi LA Rams. By standard stats, the 73 version looked like a real juggernaut, but because the league didn't seed the playoff teams until 1975, the Rams played their divisional round game at Dallas where they lost. The 1975 team looked great as well.

2. The 1976-1980 Patriots. Many of my fellow Pats fans, espiecally the older ones, swear that the late 70s Pats teams were better than the Belichick/Brady teams. (For the record, I'm not one of them). Anyway, those teams were really loaded with great players like John Hannah, Mike Haynes, Leon Gray, Stanley Morgan and Raymond Clayborn just to name a few. They got hosed in 76 by a bogus roughing the passer call late in their playoff game vs the Raiders and lost. But early that season, they beat Miami, Pittsburgh and Oakland in consecutive weeks, which has to be one of the greatest 3 game runs of all time. But because of a medeling front office, the dynasty never happended.

3. The 1973-1975 Bengals. I'm mostly intrigued by how good their pass offense was. Qb Ken Anderson's standard passing effiency stats were way ahead of the field in both of those seasons, not doubt helped by o coordinator Bill Walsh. One game, in either 1973 or 74, Anderson completed what was then a single game record of 15 or 16 straight passes vs the legendary Steel Curtain in a Bengals victory. Though I was pretty young, I did know how potent their passing attack was. I also remember watching the Steelers snuff it out like a cigarette in a late season, Saturday afternoon 35-14 Steelers win in 1975. That game played out a lot like the Seahawks/Broncos Super Bowl and really made me appreciate how great the Steel Curtain really was.

Thanks very much ahead of time.

133
by Andreas Shepard :: Sat, 02/15/2014 - 10:38pm

1. The 1973 Rams were in fact a juggernaut - they just missed the table above by finishing with the 23rd best DVOA of all time (although of course the Dolphins that year were 17th). While the 1976 squad comes out as the 2nd best team in that period, they were actually great throughout. From 1966-1980, they averaged a DVOA of 20.2% (although they somehow never led the league, finishing 2nd or 3rd 7 times in that span). Early on they were led mostly by the defense and the Fearsome Foursome (peaking with the 8th best defense ever in 68), although in the mid-late 70s the offense caught up. That helped them win a record 7 straight division titles from 1973-79.

2. The 1976-1980 Pats actually come out as merely good. The 76 team is the best of the bunch, but their DVOA is only 15.1%. The other teams hovered in the 10% range. So substantially worse than the Belichick era Pats, who have been above 15% 10 out of the last 12 years, often by very large margins. That said, 76-80 was easily the best 5 year run in Patriots history up to that point. And compared to 1968-1972, when they had a DVOA worse that -40% in 3 of 5 seasons, those late 70s teams probably felt like all-time greats to any fan who had the stomach to watch them in the early days.

3. Those Bengals were quite good (averaging 18.3% DVOA from 72-76). They did in fact have a very good pass offense, and it improved every year that Bill Walsh coached QBs. Their passing DVOA was 9.0% in 71, 18.3% in 72, 30.3% in 73, 37.5% in 74, and 41.0% in 75. That last number is #67 all time, behind (among others) 7 different 49ers teams in the Montana/Young era. Turns out Bill Walsh knew what he was doing.

134
by Michael19531 :: Sun, 02/16/2014 - 2:34am

Great info Andreas. Touche on the comments about the 1970s Pats. You hit the nail right on the head lol.

135
by FBPER :: Tue, 02/18/2014 - 3:15pm

If you want, I'd be happy to put the estimated DVOA ratings on my site. I use a plugin that makes searching and sorting through big tables very easy: http://www.footballperspective.com/data-dump-schedule-adjusted-pythagore...

For example, if you type in "Mia" you get every Dolphins team, and if you type "Mia 199" you get the 10 Dolphins teams of the '90s.

136
by Andreas Shepard :: Wed, 02/19/2014 - 1:55pm

That would be great. When I was thinking about how to post this kind of stuff, your tables came to mind as the best practice. Shoot me a note ([firstname].[lastname] at gmail) and we can go from there.

137
by eagle97a :: Sat, 02/22/2014 - 8:34am

Denver 1998 was a better all around offense than its 2013 version according to this methodology. What i found interesting is Mia 1973 is rated higher here than its 1972 undefeated version.It sort of validates the players claim that 1973 was in fact better than 1972.

138
by Andreas Shepard :: Sun, 02/23/2014 - 2:43am

To those who asked to see the full data set, Chase Stuart over at Football Perspective has kindly published the full estimated DVOAs for every team from 1950-2013 at the link below. The post also includes a detailed explanation of the methodology, which goes into all of the nitty-gritty details.

http://www.footballperspective.com/estimated-dvoa-ratings-from-1950-to-2...

140
by Larry P :: Sun, 03/16/2014 - 1:24pm

Interesting stuff, and I am glad you mention playoff success. Unfortunately you are using only regular season data. As you mention the 85 Bears:

"This is where we have to remember that these estimates only consider the regular season; one of the strongest arguments for the 1985 Chicago Bears as the greatest team ever is the way they stomped over three playoff opponents by a combined score of 91-10."

IMHO, post-season success is ultimately what legacy is built upon. If there was a way for you to include post-season games in your calculations I think there would be a meaningful ranking we'd all be interested in seeing. Teams with great regular season's who then bomb in the playoffs would rank lower (as they should) and teams that make excellent playoff runs would move up in the rankings.

Just my 2 cents worth.