Writers of Pro Football Prospectus 2008

Most Recent FO Features


» Defense and Rest Time

Do defenses really wear out over the course of a game? Do defenses benefit from long drives that give them more time to rest on the sideline? Guest columnist Ben Baldwin investigates.

08 Feb 2005

2004 Patriots vs. History

Part I: Dominance Revisited

by Aaron Schatz

I'm sure most people have had enough New England dynasty talk, but there's been less talk about where this individual Patriots team, specifically from the 2004 season, stands among the best teams in NFL history.  Peter King had his take in yesterday's MMQB, and Ned Macey has his take in the second half of this article below.  In the last couple weeks, I've found myself wondering what Eddie Epstein would say.

For those who don't know, Epstein wrote the book Dominance: The Best Seasons of Pro Football's Greatest Teams back in 2002.  It's a very good book, filled with interesting facts about some of the great teams in NFL history, and one of the few books to look at the NFL through a statistical lens.  The book started as a feature for ESPN.com that you can still read online, and was inspired by Epstein's previous book co-written with Rob Neyer, Baseball Dynasties.

Epstein chose 12 teams from NFL history for his book, using the following criteria: 1950 or later, Super Bowl (or pre-66 NFL) champion, no wild cards (adios, 1997 Broncos), and a rating over 6.00 on his "Adjusted Power Index."  API was similar to the measure Epstein and Neyer used to rank teams in Baseball Dynasties.  It consisted of adding together the number of standard deviations that distanced a team from the league average in four statistics -- points scored, points allowed, yards gained and yards allowed -- with a slight adjustment for strength of schedule.  By using standard deviations, Epstein could compare teams from the 1950's, when the dominant teams stood head and shoulders above the rest of the NFL, with teams from recent years where parity was far greater.

To give a quick example, the average NFL team in 2004 scored 343.8 points and gained 5470.3 yards.  The standard deviation was 69.8 for points, 656.2 for yards.  The Kansas City Chiefs scored 483 points, so that part of their API is 2.00.  They gained 6922 yards, so that part of their API is 2.21.  Their defense, of course, sucked, so their defensive API is -3.09.  That leaves the total (unadjusted for schedule) API as 1.11.

Since the book was published in 2002, it only includes teams up until 2001.  The 2001 Rams actually have the highest API of any team since 1950 at 8.40, but Epstein thanks the Patriots for saving him from writing an extra chapter at the last minute.  The 2000 Ravens and Titans both make the list of best defensive APIs ever. After getting 12 teams using API, Epstein picks his list of the best teams according to more subjective criteria, concluding the book by ranking the 1985 Chicago Bears as the greatest team in NFL history over a single season.

I've been wondering, where would the recent New England championship teams rank in Epstein's API?  What about the dominant defense of the 2002 Bucs?  So I figured out the API for each team of the past five years.  Unfortunately, Epstein never gives his method for schedule adjustment, so I had to approximate that based on a few teams in the book for which I have recent data handy.  To go back to our previous example, Kansas City's opponents had a winning percentage of .550 (not counting games against Kansas City) so their API ends up 0.86.

Here are the top 12 API ratings from the past five years.  Because this uses my approximation of the strength of schedule formula, it won't exactly match Epstein's ratings from the book for the 2001 Rams, 2000 Ravens, and 2000 Titans.  This list demonstrates some of the problems with the API metric, but it also shows that however much you thought the 2001 Patriots Super Bowl win was an upset, double it.

2001 STL 6.23 2.20 8.42   2004 PIT 0.42 3.71 4.13
2001 PIT 1.64 3.32 4.96   2004 DEN 1.96 2.11 4.07
2002 OAK 3.49 1.47 4.96   2004 NE 2.05 1.97 4.02
2002 TB -0.45 5.13 4.68   2000 BAL -0.31 4.19 3.89
2000 TEN 0.37 4.08 4.45   2003 DEN 1.58 1.95 3.53
2002 PHI 1.61 2.63 4.25   2003 IND 2.94 0.56 3.49

Egads.  The Rams come out as the best regular season team of all time, but no other team can top an API of 5.00, and you needed 6.00 to make it into Epstein's book.  New England doesn't even come out as the best team of 2003 or 2004.  The high rating for the 2003 and 2004 Broncos demonstrates a couple of problems with API: it takes into account neither special teams nor turnovers, two issues that the Broncos have been dealing with for two seasons.  (By the way, if my approximation of strength of schedule measure is at all close to Epstein's, the 2002 Bucs end up with the second-highest defensive API of all time, behind the 1970 Vikings.)

It seems odd that none of the last five Super Bowl teams would have even come close to making Epstein's book.  Parity has made the top NFL teams less dominant, but the fact that API uses standard deviations is supposed to solve for that problem.  I wanted to see what the ratings looked like for the top teams in another five-year period, so I checked out 1992-1996.  This period includes two teams from Epstein's book, the 1994 49ers and the 1996 Packers, and the team that Epstein admits was the last one cut from his list, the 1992 Cowboys.  The numbers will be slightly different from Epstein's since I'm approximating the strength of schedule adjustment.

1995 SF 3.60 4.14 7.74   1993 SF 5.30 -0.38 4.92
1996 GB 3.62 4.00 7.62   1996 DEN 2.80 2.05 4.85
1994 SF 5.01 1.04 6.05   1992 HOU 2.42 2.30 4.72
1992 DAL 2.62 2.81 5.43   1993 DAL 2.36 2.16 4.52
1992 SF 4.30 1.06 5.36   1995 DAL 2.55 1.89 4.44
1994 DAL 1.69 3.31 5.00   1995 KC 0.00 3.86 3.86

What a difference. First of all, this five-year period was completely dominated by two franchises, Dallas and San Francisco.  Second, six teams are at 5.00 or above.  The top teams are significantly better than the top teams of 2000-2004.

The big debate about the Patriots is this: does the fact that they win so many close games reflect an intentional style of play, or the fact that they are not as dominant as great teams of the past, or something in particular about this era of football?  This exercise would seem to give some credence to the last opinion. Either the 49ers and Cowboys of the early 90's were the two best franchises in NFL history, or parity in the last few years has increased exponentially, to the point where standard deviation measurements can't necessarily be used to compare teams between eras. With the dramatic exception of the 2001 Rams, the best teams of the past few years simply have not dominated in the same way as the other great teams of NFL history, and the rules of this era make the Patriots look less impressive than they would otherwise.

One more thought before I turn things over to Ned Macey.  I think that New England's 29-28 loss to Miami on Monday Night Football strikes their 2004 season from consideration as one of the greatest in NFL history.  The greatest team in NFL history simply does not blow a fourth quarter lead to a 4-12 team.  New England's statistics would also look better without those final 12 Miami points and the yards it took to get them, plus the two Tom Brady interceptions that gave Miami the ball for a comeback and then kept New England from one of its own.

Part II: Where Do the 2004 Patriots Rank in NFL History?

by Ned Macey

(Note: This article brings Ned Macey aboard the Football Outsiders staff.  Ned, an Indianapolis fan, will be covering the AFC South for our new offseason feature Four Downs as well as helping us with stat projects and writing articles for Pro Football Prospectus 2005. This article originally appeared in a different form on Ned's blog, Man of Leisure Sports.)

The 2001 Patriots were not a great team. According to DVOA (Defense-adjusted Value Over Average, explained here if you are reading FO for the first time), they were only the 16th best team during the regular season. In the playoffs, they won the infamous "tuck" game over Oakland, beat Pittsburgh thanks to two special teams touchdowns, and upset the highly favored Rams on a last second field goal.

As Aaron has pointed out numerous times, 2004 is the summit of a mountain that the Patriots have been climbing since Bill Belichick took over. If you look at the Patriots under Bill Belichick, they have won 5, 11, 9, 14, and 14 games. Flip the second and third years, and you would get a better idea of what was going on. Over the past five years, New England's rank in DVOA has been 22nd, 16th, 11th, 2nd, and 1st. That they picked up two Super Bowls before they reached the peak of their ability is a testament to how great a team they really are. The Patriots just became the second team to win three Super Bowls in four years. This fact, in and of itself, makes moot the whole question of whether or not the Patriots are a dynasty. The more interesting question is where the 2004 Patriots rank among the greatest teams of all time.  (Actually, of the Super Bowl era; for this survey, 1966 will represent a convenient starting point for modern football.)

Here were my criteria for considering teams: First, though I do not believe that the Super Bowl champion is necessarily the best team in any given year -- 2001 being a great example -- the greatest team of all time should certainly have been able to win the big game. Second, I included any team that appeared in multiple Super Bowls in a four year period and lost three or fewer games. Unlike Eddie Epstein's book which Aaron addresses above, I am placing an emphasis on avoiding one-year wonders. A great team should be able to sustain excellence over time. If a team did not make two Super Bowls in four years, it had to have won the Super Bowl with a record of 12-2 or better. Any time a team won multiple Super Bowls in a four year period, I included only the team that I determined to be the best among the Super Bowl winners (for example, only the 2004 Patriots).

The following eighteen teams met the above requirements:

1966 Green Bay Packers
1969 Kansas City Chiefs
1970 Baltimore Colts
1971 Dallas Cowboys
1972 Miami Dolphins
1975 Pittsburgh Steelers
1976 Oakland Raiders
1977 Dallas Cowboys
1984 San Francisco 49ers
1985 Chicago Bears
1986 New York Giants
1989 San Francisco 49ers
1991 Washington Redskins
1992 Dallas Cowboys
1996 Green Bay Packers
1998 Denver Broncos
1999 St. Louis Rams
2004 New England Patriots

I will also take a look at two teams ranked by Epstein that did not match my criteria: the 1979 Pittsburgh Steelers and the 1994 San Francisco 49ers.

That's 20 teams, so the next step was to cut the list of teams in half. I measured teams quickly in the following categories: one year win-loss record, three year win-loss record surrounding the year, five year win-loss record surrounding the year, average margin of victory in playoffs, closest playoff game, and Super Bowl wins during the five year period.

My list of ten teams starts with all teams that went undefeated or lost only a single game: 1972 Dolphins, 1976 Raiders, 1984 49ers, and 1985 Bears. Next I took the three most dominating teams by playoff margin of victory: 1989 49ers, the 1986 Giants and again the 1985 Bears. I also added the 1977 Cowboys who never played a playoff game closer than 17 points. I added the 1992 Cowboys and 2004 Patriots, as they are the only two teams to win three Super Bowls in four years. Finally, I included the 1998 Broncos, who had the best three-year winning percentage of any team already not included. These may not be the ten greatest teams of all time, but I think they have the best arguments for being the greatest team of all time.

For those fans of the teams I eliminated, I offer the following quick reasons:

  • 1966 Packers: Middle of the pack (no pun intended) on all of these indicators. No clear reason to argue for them as the greatest.
  • 1969 Chiefs: Only one Super Bowl win, the second lowest one-year winning percentage, and the lowest five-year winning percentage of any of the pre-1980 teams.
  • 1970 Colts: They are my eleventh team, but I cut to 10 mostly because they struggled in the playoffs after their 1968 incarnation was upset in the Super Bowl. 11-2-1 regular season record is not particularly dominant.
  • 1971 Cowboys: Comparing these Cowboys to the 1972 Dolphins, they had a worse one-year, three-year, and five-year record. They also won only one Super Bowl. Their playoff margin of victory is not that impressive.
  • 1975 Steelers: The Steelers won a great deal of Super Bowls, but their peak was never that high. The 1975 Steelers were their best team, but they did not dominate in the playoffs and do not compare favorably with their contemporaries when it comes to three-year and five-year records.
  • 1979 Steelers: These Steelers lost four games during the regular season. They lost two games by more than 20 points during the regular season. Their point differential was significantly lower than their 1975 incarnation eliminated above.
  • 1991 Redskins: Joe Gibbs seemed to spread out his good teams. Over a five-year period, despite winning two Super Bowls, the Redskins had the lowest winning percentage among the 20 teams.
  • 1994 49ers: These 49ers had an amazing year, but it was their first Super Bowl in five years, and their last ever. They lost to the Cowboys in the playoffs in 1992 and 1993. These teams retained similar cores throughout, but the Cowboys made three Super Bowls.
  • 1996 Packers: At 13-3, they tie for the second-worst regular season record. They also won only one Super Bowl.
  • 1999 Rams: Very similar to the 1996 Packers, who also won a Super Bowl but lost a second one or two years later when they were highly favored. The Rams also have one of the worst five-year records because they had gone from worst to first in 1999, and they have the second-lowest margin of victory in the playoffs.

With a list of 10 teams, I figure I can go into slightly more depth. The first thing I did was look at the season-long point differential for all of these teams during the regular season:

1985 Bears: 258
1984 49ers: 248
1972 Dolphins: 214
1998 Broncos: 192
1989 49ers: 189
2004 Patriots: 177
1992 Cowboys: 166
1986 Giants: 135
1977 Cowboys: 133
1976 Raiders: 113

Based on these numbers, the first team we are going to eliminate from contention is the 1986 Giants. This team was dominant in the playoffs, winning by over 27 points per game, but their three- year record is the worst in this group, and their five-year record will probably be the worst when the Patriots can drop their 2000 record after next season.

The next team to go, surprisingly, is the 1992 Cowboys. With the Patriots matching their three Super Bowls in four years, they compare favorably to the Cowboys across the board. The Patriots have a better regular season record, an equal three-year record, and a better regular season point differential. Following the Patriots' three-point win in the Super Bowl, the Cowboys have a better margin of victory in the playoffs. However, they played one great team in the playoffs, the 49ers, while the Patriots just beat the three best teams in football by both traditional record and DVOA.

Third I'm removing the 1977 Cowboys. They match up with the 1976 Raiders as great teams of the 1970s that won just a single Super Bowl. The Raiders, however, have a better one-year, three-year and five-year record. The Cowboys won their playoff games more convincingly, but the Raiders played three teams that won at least 10 games, while the Cowboys built a good deal of their margin of victory with a 30 point win over 9-5 Chicago before beating another 9-5 team.

All that being said, the fourth team to go is the 1976 Raiders. They had the worst point differential by far and only won one Super Bowl. They had an amazing three-year and five-year run, but the 1972 Dolphins are superior in both respects, and that team won two Super Bowls.

Finally, I will remove the 1984 San Francisco 49ers. They were truly an amazing team, but they do not compare favorably with the 1985 Bears. The 49ers did win two Super Bowls, but the second one was in 1981, and the 49ers suffered a 3-6 season during the strike year between the two. The Bears had a better three-year record, a better five-year record, and a better point differential during the regular season. Are the 1984 49ers one of the five best teams of all time? Perhaps, but they are directly inferior to the Bears, and therefore, we will move on without them.

That leaves us five teams: the 1972 Dolphins, the 1985 Bears, the 1989 49ers, the 1998 Broncos, and the 2004 Patriots. Let us look at each team in terms of their arguments for and against being the greatest of all time.

1972 Dolphins

The most common argument for the 1972 Dolphins is that they are the only team not to lose a game. This argument, however, is very simplistic. They played a terrible regular season schedule, teams that combined for a dismal .365 winning percentage. In the playoffs, they played quality opponents, second best record amongst these five teams, but won by the slimmest margin of any of the 20 teams consulted in this study. They won 3 games by a total of 17 points.

Some people believe that the 1973 Dolphins were actually the better team. That team went 12-2 and won their playoff games by a total of 52 points. Either way, the strength of the 1973 team, and the fact the 1971 team lost the Super Bowl, reflect strongly on the 1972 team. They have the best three-year and five-year winning percentages of any of the teams studied here. They have the second highest point differential of any of the final 10 teams. The biggest problems with their resume are their weak quality of competition and the narrowness of their playoff wins.

1985 Bears

As Aaron noted above, Epstein's book Dominance ends by naming Chicago the best team of all time over a single season. They were 15-1, losing only that memorable Monday Night game against the 12-4 Dolphins. Their quality of opposition is the second best of the five teams included here, although they never beat a team with more than 11 wins. Their playoff margin of victory is the second highest as well and third highest amongst the 20 teams. They won their playoff games by a combined 91-10.

But Chicago never made another Super Bowl. Their overall regular season records were strong: the third best overall three-year record, and fourth best five-year record amongst the twenty teams. How did they fail to make a Super Bowl? In 1984, they got blitzed by the 49ers 23-0. In 1986, they went 14-2 but lost their first playoff game to the Redskins 27-13. This game was played without Jim McMahon, but Jim McMahon was not exactly Dan Marino, and they got beat convincingly. The next season, McMahon was healthy, but the Bears fell again to the Redskins, 21-17. All told, the team went only 4-3 in the playoffs from 1984-1987.

1989 49ers

This team made the final five because of its amazing run in the playoffs. They won their three games by a total of 100 points, never winning by fewer than 27 points. The 1989 49ers were the 49ers at their peak, with Montana, Rice, and Craig all in their prime. Like the Patriots this season, they were defending Super Bowl Champions and even better in the year they repeated.

When compared with the final ten teams considered here, their regular season prowess does not match up well. They were fifth in point differential, despite playing a middling schedule, with an opponents' winning percentage of .457. Their three-year winning percentage is only 6th. Their five-year winning percentage is a more impressive fourth. The playoff run was amazing, but the best team they played was only 11-5.

1998 Broncos

This team, in my mind, is one of the least appreciated teams of all time. Coming off a season where they upset the Packers in the Super Bowl, the Broncos started 13-0 before dropping two of their final three. They returned to form in the playoffs, winning their three games by 63 points. Their overall point differential was 192 points, fourth best of the final 10 teams and best in the post-free agent era. Their opponents, however, posted only a .438 winning percentage, second worst amongst these final five teams.

Their five year winning percentage is weak, ranking only 16th out of the original 20 teams, but this is largely attributable to the retirement of Elway and the injury to Terrell Davis. If they had gone just 10-6 in 1999 -- a likely scenario if Elway and Davis had been available -- they would rise to the middle of these rankings.

2004 Patriots

The Patriots may not yet have the resume to compete with these teams. Amongst the original 20 teams, they are currently tied with four other teams for 12th in three-year winning percentage and are 19th in five-year winning percentage. Of course, the Patriots current three-year record includes their 9-7 season in 2002 and the five-year record includes their 5-11 record in 2000. Assuming a very conservative 10-6 for next season, their three-year record rises slightly, and they move to 15th in five-year winning percentage, best among the teams whose winning runs come exclusively in the free agency era. If they were to go an impressive 13-3 next year, after consecutive years of 14-2, they would move to third in three-year winning percentage and ninth in five-year winning percentage, behind only pre-1990 teams.

As for this season, the combined record of the Patriots' playoff opponents is an amazing 40-8, the most difficult Super Bowl run of all time. The 1989 49ers did not play anyone better than 11-5; the Patriots did not play anyone worse than 12-4. They are the only team on this list to beat a 15-1 team. They have a decent average playoff victory margin of 11.3 points, but this matches up poorly with the other teams on this list. Super Bowl XXXIX was also the closest playoff game of any of the twenty teams in this study.

Preliminary Conclusions

Which of these five teams is the greatest? I think it is clearly not the Broncos, who, while exceptional, are inferior to the Patriots. They had a slightly higher point differential, but their opponents had a .438 winning percentage compared with the Patriots' opponents .492 winning percentage. The Patriots repeated as champions with a better two-year record. The Patriots' three-year and five-year records are inferior for now, but with most of this year's team coming back next year, it seems likely that these numbers will improve significantly.

What we are left with are the best single year team, the 1985 Bears, the best three-year team, the 1972 Dolphins, the best long-term dynasty, the 1989 49ers, and the best free agency era team, the 2004 Patriots. This survey could clearly be made more sophisticated, but based on these limited calculations, I would go with the 1989 49ers as the NFL's greatest team. The 1985 Bears were an amazing team, but they played one team the whole season that had a record better than 11-5, and they lost. The previous year, the 49ers were a better team than Chicago, and the next two years, they failed to win a single playoff game. The 1972 Dolphins beat up on a series of bad teams and barely won the Super Bowl.

The 1989 49ers were coming off of a Super Bowl championship and the following year they went 14-2. Their playoff run is unprecedented, winning the three games by a combined score of 126-26. Over the two postseasons following the 1988 and 1989 season, they only played one game decided by fewer than 25 points. The case for these 49ers, in my mind, would be a slam dunk if they had won the Super Bowl following the 1990 season, but they dropped the NFC Championship on a last minute field goal, with defensive coordinator Bill Belichick and his New York Giants holding the 49ers to 13 points.

How the Patriots Compare

The 2004 Patriots were well below the final five teams in terms of point differential, but as stated above, they played by far the toughest schedule of any of these teams. The playoff performance of the Patriots, while less dominating, matches anyone else on this list. They played the three best teams in football and won two of the games handily.

But the 2004 Patriots have three major strikes against them. First, they got blown out by the Steelers, something that did not happen to any of the other final five teams. Second, as Aaron notes above, they lost to 4-12 Miami, by far the worst team to beat any of the teams we are considering here. The Broncos lost to an 8-8 Giants team, but no one else lost to anyone worse than 10-6. Third, I am a great believer in DVOA, and the Patriots only have the fourth-best DVOA since 2000 (behind the 2002 Bucs, 2001 Rams, and 2000 Titans). 

The Patriots just achieved something amazing, three Super Bowl victories in four years, and they did it in an era of unprecedented parity. They ran through the most difficult playoff gauntlet in history and beat three exceptional opponents in three different ways. They played an exceptional regular season schedule, teams that were 124-116 when not playing the Patriots, and went 14-2.

The Patriots are probably not the greatest team of all time. The salary cap, free agency, and expansion have diluted the talent pool. The greatest teams of the 1970s and 1980s were probably more dominant over their competition.

If you factor in the change in eras of football, however, the 2004 Patriots may prove to be the most impressive team of all time. Their record, level of consistency, and domination of top opponents are all similar to what the greatest teams were able to do while not economically constrained like the current Patriots. If they manage another extraordinary season next year, they will be equivalent in historical record with these other teams playing in eras much more conducive to continued success.

In 2001, a group of players led by a brilliant coach won a Super Bowl in one of the greatest upsets in Super Bowl history. This upstart group has developed into the most dominant team of the past 10 years, a team right that deserves a place in the discussion of the greatest football teams of NFL history. It is unfortunate that a lack of "star" players and a fluky beginning to their run are preventing many people from appreciating the greatness on display each week with the New England Patriots.

Posted by: admin on 08 Feb 2005

1 comment, Last at 07 Jul 2005, 2:19am by RamTime


by RamTime (not verified) :: Thu, 07/07/2005 - 2:19am

How do you rate blown non calls? Take the intentional grounding that was not called on the final drive of the Rams Patriots superbowl. This was intentional grounding that should have been called and not only the 10 yard penalty and loss of down but it was so blatant Brady should have been slapped along with the it.
The Patriots OC said they would have killed the clock and played for overtime if the preceding play had worked so it just makes sense that the Patriots would have killed the clock on 2nd and 20 with 29 seconds left the ball on their own 30 and no timeouts. Is there anything VOA can do to filter out blatant favortism? To refresh your memory here is that play.
another look
or this http://stlouisrams.net/download/gift_12.wmv