Building a Super Bowl Winner: Offense

by Scott Kacsmar

In Part I of our look at Super Bowl teams in the DVOA era, we looked at some of the most common traits of those championship clubs. Building at least one dominant unit (an offense or defense in the top three of the DVOA ratings) was an important factor. But is one side of the ball more important? Today we are focusing on the offensive side of the ball.

"Offense wins championships" has never caught on as a saying, but offense does tend to win a lot of the games necessary to get to those championship opportunities. We observed through DVOA that more No. 1 offenses (six) have won a Super Bowl since 1989 than No. 1 defenses (five). Top-ranked offenses (four) have even lost more Super Bowls than top-ranked defenses (two) in that time. But obviously a great offense by itself cannot get the job done. Dan Marino never won a Super Bowl in Miami, while the Air Coryell Chargers and run and shoot Oilers never even got to one.

Consistently strong offenses do usually make it to at least one Super Bowl. Since 1989, there have been 12 offenses that ranked in the top five in DVOA for at least three consecutive seasons. Nine of those teams appeared in at least one Super Bowl, with seven winning a championship. The 2009-2013 Patriots and 2012-2014 Broncos did not win a Super Bowl at their dominant peaks, but both won one in their very next seasons with offenses ranked out of the top five. So a lot of the same pieces were still present, though Denver's offensive decline last year is well known. As for the three offenses that failed to reach the big game, two of them -- the 2000-2002 49ers and 2001-2005 Chiefs -- partially fell victim to that more defensively-dominated era in the early 2000s that we discussed in Part I. Dick Vermeil's Chiefs may have had as much offensive talent as his Greatest Show on Turf Rams did, but the defense was never there and the AFC started to be controlled by the Patriots, Colts, and Steelers. That also played a factor in the third example, the 2008-2011 Chargers, never getting past the AFC's divisional round.

It certainly pays to build a great offense, and we know from countless studies that offensive success is more sustainable from year to year than defensive success. But how does one go about acquiring the pieces to build and sustain a great offensive unit? It would be crude to suggest "sign a Hall of Fame quarterback" as the first part of the strategy when there are only 25 modern-era (since 1945) Hall of Fame quarterbacks in existence. An offensively-driven team really needs that caliber of quarterback, but even when we add Kurt Warner, Tom Brady, Peyton Manning, Ben Roethlisberger, Eli Manning, Drew Brees, and Aaron Rodgers to the list of Hall of Famers, that is 32 players over the span of 70 years. Meanwhile, teams in 2016 are preparing to start Trevor Siemian, Blaine Gabbert, Shaun Hill, and Case Keenum in Week 1.

No one said winning a Super Bowl was easy, but it is easy to spot great offense when you see it.

Super Bowl-Winning Offenses

Our first table looks at the last 27 Super Bowl winners in the DVOA era (since 1989). Their rank in offensive DVOA along with their rank in the previous season is included, as well as the rankings in both passing and rushing. Also included are where the offense ranked in weekly variance, ordered from most consistent to least consistent that season, and the average schedule of defenses faced, ranked from hardest to easiest. Top-three units are in bold.

Two-Year Outlook

The average offensive rank is 7.8, as only eight teams ranked outside of the top 10, and interestingly enough, all of them have come since 2000. The four offenses ranked 20th or worse were paired with what are most often considered (along with the 2013 Seahawks) the best defenses in the NFL in this time: 2000 Ravens, 2002 Buccaneers, 2008 Steelers, and 2015 Broncos. No offense ranked lower than Denver at 25th last season, but that also represents the biggest single-season fall as the Broncos had still been the third-ranked unit in 2014. The decline of the offensive line and unresolved losses of Julius Thomas and Wes Welker hurt, but the biggest problem was Peyton Manning's final season being his worst -- a far cry from the peak he managed to sustain from 1999 to 2014.

Nineteen of these teams had been a top-10 offense in the previous season, including 14 top-five finishes, and that will likely be 20 once we add the 1988 49ers. On average, these units deviated 5.2 spots in the rankings from the previous season. We again see that bizarre period around 1999-2003 producing unique results with the more defensively-driven winners. The 1999 Rams, 2000 Ravens, and 2001 Patriots are the only offenses to rank worse than 14th in previous-season DVOA. Baltimore won with defense, but did benefit from drafting running back Jamal Lewis and signing tight end Shannon Sharpe in 2000. The 2001 Patriots of course replaced an injured Drew Bledsoe with Tom Brady at quarterback, but it was actually the 2004 Patriots that made the second-biggest jump in the rankings, climbing 11 spots after adding Corey Dillon to the backfield.

But no one can compare to the massive improvement the Rams made in 1999, moving up 22 spots to the No. 4 offense. (Yes, not the No. 1 offense; opponent adjustments were huge.) While Isaac Bruce (1994) and Orlando Pace (1997) were already there, the Rams made three huge moves. Marshall Faulk was acquired via trade with Indianapolis and became the best dual-threat back in the NFL. Torry Holt was drafted with the No. 6 pick and formed a great duo with Bruce for years to come. The biggest move was a stroke of luck. Trent Green was supposed to be the starting quarterback for this team, but his season ended after a hit by Rodney Harrison in the preseason. The Rams rallied around Kurt Warner, who was on the roster in 1998, but had no success to speak of unless we are counting the Arena Football League. Warner had a dream season, throwing 41 touchdowns and claiming MVP and Super Bowl MVP honors. This was the start of one of the greatest three-year runs on offense in NFL history. It was only made possible by adding three Hall of Fame talents to the starting lineup in 1999, one by complete accident. That's not a realistic blueprint to follow.

Variance and Schedule

Weekly variance did not prove to be a factor for these teams. There were just as many consistent offenses in the top 10 (10) as there were inconsistent offenses ranked 20th or worse (10). The 2012 Ravens (30th) were the least consistent offense, but sure put it together in the playoffs. The 1990 Giants (second) were the most consistent offense despite losing Phil Simms late in the season.

The average schedule faced was pretty much average at 14.1, though there were more teams (11) that played a top-10 schedule than teams (four) that played a bottom-10 schedule. A year after Denver won the Super Bowl in 1998 despite playing the fourth-toughest schedule (and the highest listed here), the 1999 Rams had it the easiest with the last-ranked schedule.

Passing vs. Rushing

Fourteen of these offenses were able to rank in the top 10 in both passing and rushing DVOA for good balance. However, the average passing offense (7.4) was better than the average rushing offense (9.7), which is to be expected when we preach the importance of the passing game in dictating NFL outcomes. Eleven teams featured a top-two passing attack, while only five teams finished that high with the running game. This is not to say the running game is irrelevant, because only five of the 27 teams finished in the bottom half of the league in rushing DVOA. However, that does include the 25th-ranked 2003 Patriots and 28th-ranked 2002 Buccaneers, the lowest units in this table.

The best rushing teams share something in common: great running backs (Emmitt Smith and Terrell Davis) that were heavily relied upon behind very talented offensive lines in smart schemes. Smith's 1993 and 1995 seasons may have been the best of his career, as he won MVP and Super Bowl MVP in 1993 after a two-game holdout to start the season, and set career-highs in 1995 with 1,773 rushing yards and 25 touchdowns. Davis was incredibly productive for Denver in 1997-98, leading the league in DYAR both years, achieving a 2,000-yard season, and winning an MVP in addition to a Super Bowl MVP. As I highlighted in his Hall of Fame case three years ago, when the playoffs are included, Davis had 951 carries in those two seasons and the two highest rushing totals in NFL history. No running back has ever meant more to a Super Bowl-winning team than Davis, who did it twice. This is in stark contrast to the other great rushing team, the 2009 Saints, who were very efficient with a strong interior line, but had no individual back even rush for 800 yards that season. In the playoffs, the running game was dominant against Arizona, but afforded Drew Brees very little production (39 carries for 120 yards) in the NFC Championship Game and Super Bowl. Tracy Porter's two massive interceptions were more significant for New Orleans than the No. 1 rushing attack.

Alas, those run-heavy approaches with Smith and Davis are from a different era of NFL football, but some teams will still try to win that way this year. Adrian Peterson will have to be great for Minnesota with Norv Turner as offensive coordinator. Turner was Smith's offensive coordinator in 1991-93, but unlike in Dallas, Peterson is not playing with Troy Aikman, Michael Irvin, and the best offensive line in the league. Even the current Dallas team thinks it has the right setup with its offensive line, rookie back Ezekiel Elliot, and a healthy Dez Bryant, but will have to start the season with fourth-round rookie quarterback Dak Prescott after another Tony Romo back injury. Rarely do rookies pay dividends for championship offenses.

Timing Is Everything

Examples like Dallas show the importance of timing in the NFL, and trying to match the windows for players with a variety of experience levels. The Cowboys acquired Irvin, Aikman, and Smith in the 1988-1990 drafts, so they were able to grow and hit their primes at the same time. Right now, Dallas has a young line and Bryant in his prime, but a rookie quarterback and a 34-year-old Jason Witten nearing the end of his career. Even if Prescott is the future, no rookie quarterback has ever started a Super Bowl.

The following table shows each Super Bowl winner's primary quarterback, leading rusher, leading receiver (by yards), and head coach. The "XP" is how many seasons that person had been with the team he won a ring with at the time. Members of the Hall of Fame are in orange, likely future Hall of Famers are in yellow, and rookies are in green.

(Note: the importance of the offensive line is not lost on us. The problem is the time crunch with the regular season about to start, so I did not have time to complete an offensive line study. Unlike the skill positions, we really cannot just focus on one offensive lineman as the whole starting unit needs analyzed. Due to the nature of the position, our data is really just limited to career experience (draft status, seasons played, games started) and adjusted line yards. It is worth noting how the 2008 Steelers, 2011 Giants, 2013 Seahawks, and 2015 Broncos have been accused of having the worst offensive line in the league those years. Three of those teams had a No. 1 defense as well as quarterbacks capable of managing pressure in unique ways. Not much of a study is needed to conclude that the quality of Super Bowl-winning offensive lines was much better in the '90s than in today's game. This is something we could approach at a later date with more time and care.)

Overview

These four-man groups had been together for an average of 4.9 seasons when they won the Super Bowl. The 2014 Patriots (9.0) had more than a year's worth of experience than the next closest team, and that is even with Jonas Gray as the leading rusher instead of playoff rushing leader LeGarrette Blount, who was only in his second season with the Patriots.

I know I keep stressing how unusual 1999-2003 was, but the results back it up. The four shortest collections of talent belong to the 2000 Ravens (1.3), 2002 Buccaneers (1.8), 1999 Rams (3.0), and 2003 Patriots (3.3). The 2001 Patriots (3.5) are tied for fifth.

No team acquired all four people in the same year, though there were seven cases of getting three in one year. They were not always by plan -- Baltimore never intended to start Trent Dilfer, and the Packers traded for Brett Favre in 1992 to backup Don Majkowski. The 1995 Broncos really struck gold by hiring Mike Shanahan, finding Terrell Davis in the sixth round, and watching Rod Smith become one of the greatest undrafted wide receivers in NFL history. New Orleans stretched out success even longer with the hiring of Sean Payton, the signing of Drew Brees and stealing seventh-round draft pick Marques Colston in 2006. Baltimore had an instant rebuilding plan by hiring John Harbaugh and using its first two draft picks on Joe Flacco and Ray Rice in 2008. Finally, Seattle put itself on the right track towards a DVOA dynasty with the hiring of Pete Carroll, second-round selection of Golden Tate, and in-season trade for Marshawn Lynch in 2010. It also helped to draft Russell Okung, Earl Thomas, and Kam Chancellor that year as well.

Quarterbacks

Wow, I just pegged 21 of the 27 teams with a Hall of Fame quarterback. Even if you are understandably down on Eli Manning, it is not like Joe Flacco and Russell Wilson are closed cases. In fact, Wilson's case will likely be stronger than Eli's very soon, but it is only 2016. Brad Johnson, Trent Dilfer, and Mark Rypien stand out here, and Phil Simms was replaced by Jeff Hostetler due to injury for the stretch run for the 1990 Giants.

Quarterbacks had the most experience with their team (6.3 seasons), but a lot of that has to do with John Elway and Tom Brady winning Super Bowls so late in their careers after also getting to three early ones. They are the only quarterbacks in NFL history to start at least five Super Bowls.

Twenty-one of the quarterbacks were in at least their fourth season with the team when they won. Only Trent Dilfer won in his season debut, and he did not even open the 2000 season as Baltimore's starter. He ended up replacing Tony Banks. Again, that period of 1999-2002 was responsible for four of the six quarterbacks to win within their first two seasons on a team. Since then, only Ben Roethlisberger and Russell Wilson pulled it off as sophomores on similarly built teams: vertical passing offenses bolstered by a commitment to the running game and great defense.

As much as some fan bases (Denver in particular this year) believe their team can win without a stellar quarterback, this table says otherwise. Manning was clearly not himself last year, but the DNA of a HOFer was still there. How many quarterbacks can pull off that 80-yard touchdown drive in Week 2 in Kansas City when the Broncos were down 24-17? It is impossible to say how the rest of the season would have progressed had Denver lost, but if every other game result had been the same, the Broncos would have slipped from the No. 1 seed to the No. 5 seed. It is hard to imagine any Super Bowl run then, and the Broncos led the NFL with seven games won by scores in the fourth quarter or overtime, including five against playoff teams. Brock Osweiler helped contribute to that too, but while the two Denver quarterbacks made plenty of mistakes last season, they still moved the ball (4,216 passing yards) and made those timely plays. Good luck expecting Trevor Siemian, a seventh-round pick with one kneeldown on his regular-season resume, to do the same this year.

Running Backs

They say running back is the quickest position to learn, and the numbers support that here with the leading rusher averaging just 3.3 seasons in the league. That includes the only two rookie players in the table, Jamal Lewis and Joseph Addai. We already touched on the greatness of Smith, Davis, and Faulk, but one look at this table should make it abundantly clear that a HOF-caliber back has not been required to win a Super Bowl. I am not even that confident that Marshawn Lynch will get in, but he is the closest case since 2000. Lynch and Rice are the only 1,000-yard rushers on the last eight Super Bowl winners.

Seven leading rushers were able to win a ring in their team debut compared to one quarterback, one receiver, and three coaches. We decry teams after the draft for using really high picks on running backs, and this table really shows why that is the case in today's NFL. The level of running back production necessary to win a Super Bowl can be found in players of a lesser caliber.

Receivers

The quality of player is higher here than with the running backs, though Parcells and Belichick sure loved to test that with Stephen Baker and David Givens sticking out like sore thumbs. Personally, I think Gary Clark was better than teammate Art Monk. Rod Smith and Anquan Boldin wouldn't be bad additions to Canton. Greg Jennings really went downhill after leaving Green Bay. Shannon Sharpe and Rob Gronkowski are the only tight ends on the list, and Gronkowski is getting close to first-ballot lock status after just six seasons.

Antonio Freeman, Deion Branch, and Victor Cruz are the only leading receivers who were just in their second NFL season, but Cruz blew away expectations after catching zero passes as an undrafted free agent in 2010. He exploded in 2011 with 1,536 receiving yards, joining Michael Irvin (1995) as the only receivers in NFL history to win a Super Bowl during a 1,500-yard receiving season.

In updating an old study on the top four receivers (in yards) for all 50 Super Bowl winners, I came up with the following results:

• Just over three-quarters (76.0 percent) of the players were homegrown, meaning they were acquired through the draft or as an undrafted free agent before playing for any other team.
• High draft picks paid off: 48.0 percent of the players were drafted in the first or second round. Only 15 players were undrafted (9.9 percent).
• Out of the 200 player-seasons studied, only eight belonged to rookies, including two teams with multiple players (Lynn Swann and John Stallworth on the 1974 Steelers, and Travis Taylor and Jamal Lewis on the 2000 Ravens). All of the rookies were third or fourth receiving options, except for Charlie Brown, who led the 1982 Redskins with 690 yards in the strike-shortened season.
• Torry Holt (1999 Rams) was the only rookie to exceed 700 receiving yards on a Super Bowl winner. He had 788 yards.
• The 2002 Buccaneers are the only Super Bowl winner to have its top four receivers acquired through trade and free agency. The only other team with more than two such players is the 2015 Broncos (three), who signed Emmanuel Sanders, Owen Daniels, and Jordan Norwood in free agency.
• Twenty-nine of the 50 Super Bowl winners were led in receiving by a player they drafted in the first four rounds, including Gary Clark (twice) in the supplemental draft.
• Only four No. 1 receivers were acquired through free agency: Plaxico Burress (2007 Giants), Shannon Sharpe (2000 Ravens), Todd Christensen (1983 Raiders), and Don Maynard (1968 Jets). Trades produced six No. 1 receivers: Anquan Boldin (2012 Ravens), Keyshawn Johnson (2002 Buccaneers), Paul Warfield (1972-73 Dolphins), Roy Jefferson (1970 Colts), and Carroll Dale (1966 Packers).

Free agents and rookies are rarely the tickets to a championship. Most teams draft players that develop in their system for a few years before putting it together for a Super Bowl.

Head Coaches

Expect Jimmy Johnson, Mike Holmgren, Mike Shanahan, Bill Cowher, and Tom Coughlin to get some Hall of Fame consideration down the road, but right now Bill Belichick projects to have more Super Bowl wins by a HOF coach than the rest of this group combined for this period of time.

George Seifert was the only rookie head coach to win right away, and the 1989 49ers were really the pinnacle of that team's run. However, he had been the defensive coordinator from 1983-88, and not enough credit is given to how great that San Francisco defense was every year too. Bill Walsh was the mastermind in building the 49ers machine, but Seifert did not slow them down. Still, he gets somewhat viewed as a guy who won with someone else's team, though maybe not as much as the other two coaches to win in their debut season. Jon Gruden is viewed as taking Tony Dungy's team and great defense all the way with Tampa Bay in 2002. Gary Kubiak stepped into a pretty loaded Denver team last year that had been close already and then watched the offense fall off a cliff, but defensive coordinator Wade Phillips brought the goods on his side of the ball. Barry Switzer and Mike Tomlin won in their second seasons, and it could be said they won with the teams Jimmy Johnson and Bill Cowher helped grow.

No one here took longer than Cowher (14 years) to win his first championship. The next closest was five years, which makes this pretty damn interesting. That means no other coach since 1989 has been able to win his first championship if he has been in the same city for more than five seasons. If we take this back to the beginning of the Super Bowl, then we find similarly stunning results, although it is difficult to treat coaches who started their careers before the Super Bowl equally with today's coaches. But Vince Lombardi, Weeb Ewbank, and Hank Stram all won a championship within the first five seasons of their careers, then collectively won the first four Super Bowls before the 1970 merger.

Eight coaches first won with their second team (or third in Carroll's case). On average, the coach had been with his team for 4.0 seasons when winning his first championship. Among the 31 Super Bowl-winning head coaches, only Chuck Noll (six), John Madden (eight), Tom Landry (12), and Cowher (14) needed more than five seasons on one team. Landry was tasked with an expansion team in 1960, and it should be noted that he lost Super Bowl V and won Super Bowl VI, so even that example is not nearly as long as it looks. Cowher really stands out, as does Madden given the Raiders' success before and after he was the coach.

What does this say about someone like Marvin Lewis, who is going into his 14th season with Cincinnati without a single playoff win? Not hopeful. Ron Rivera (Carolina) and Jason Garrett (Dallas) are going into their sixth full season on the job. Did Rivera just miss the best opportunity he'll ever have in losing Super Bowl 50? Time will tell, but 27-of-31 is rather daunting. This gives some support to the heavy turnover among coaches.

Quick Changes in Offensive DVOA

Change in personnel is almost a necessity to expect change in on-field performance. Practically no team brings back the same roster and staff from one year to the next, but some changes are obviously more impactful than others.

What about the teams that made huge strides in one year? I gathered the 20 teams with the largest single-year improvements in offensive DVOA (among all teams, whether they played in the Super Bowl or not) and tracked if they made changes at quarterback, head coach (or offensive coordinator), No. 1 receiver, or top running back. Players with "RK" in parenthesis were rookies. "FT" means the player had been on the roster the year before, but in a backup or limited role, then became a full-time player in a championship season. Players with an "H" designation had been on the roster the previous year, but had their season seriously compromised by health, failing to play in at least 10 games. Their return to a prominent role helped.

Five of these offenses were still below average in the second year, so we are not really interested in those cases, like when the 2006 49ers made Alex Smith the full-time starter in Norv Turner's offense that featured Frank Gore (very good) and Antonio Bryant (not so good). But anything would have been better than the 2005 abomination.

Every team made at least two big changes except for the 2011-12 Broncos. Replacing Tim Tebow with Peyton Manning was enough to turn the Broncos from a prehistoric offense into a great one for the next three years. 2010 draft picks Demaryius Thomas and Eric Decker shined in 2012, Willis McGahee led the team in rushing again, and the only other real changes were Manning bringing former teammates Jacob Tamme and Brandon Stokley along with him. There was a legitimate concern that Manning would never be able to play effectively again after four neck surgeries, but John Elway's gamble paid off.

Out of 14 quarterback changes, Cam Newton was the only rookie that has really continued to have an impact for his drafting team, but we still have the 2011 Panthers as the top Carolina offense in the last five seasons. This is also a reminder of just how abysmal Jimmy Clausen was in 2010. Also, not every change here was by design. Matthew Stafford was hurt after three games in 2010, leading to Shaun Hill taking over for most of the season.

Only seven teams changed head coaches, but some provided instant improvement, including four in 2013: Mike McCoy, Chip Kelly, Andy Reid, and Bruce Arians. Of course, the latter two are coaching Super Bowl contenders this year while Kelly is out of Philadelphia and McCoy may soon be taking a long walk off a short pier in San Diego.

The 2016 Raiders will be looking towards the 1990-91 Cowboys for inspiration, as Dallas relied on health and improvement from its young trio of Troy Aikman, Emmitt Smith, and Michael Irvin. Norv Turner took over as offensive coordinator, which Aikman always cites as a huge reason for his success. The Raiders really did not add much to the offense this year outside of left guard Kelechi Osemele, but they'll hope Derek Carr, Amari Cooper, Michael Crabtree, Clive Walford and Latavius Murray grow together in Year Two of Bill Musgrave's offense.

As far as new leading receivers go, nothing can really top Randy Moss going to New England in 2007 and scoring 23 touchdowns. That record-setting offense also added Wes Welker and Donte' Stallworth in the same offseason, so that is an extreme example of an offensive overhaul. In Buffalo, Eric Moulds did not eclipse 300 receiving yards in either of his first two seasons, but exploded in 1998 with Doug Flutie at quarterback. In fact, Moulds even topped Moss' rookie season in DYAR and DVOA.

There were a dozen running back changes, but only a few were significant, such as Marshall Faulk in St. Louis, Herschel Walker pairing up with a healthy Randall Cunningham in Philadelphia, Justin Forsett in Kubiak's system in Baltimore, and DeAngelo Williams' breakout season in 2008.

Teams like Dallas, Cleveland and Houston are best suited to join this table should there be a huge improvement in those offenses this season. We touched on Dallas already, but Cleveland has Hue Jackson at coach, Robert Griffin III at quarterback, drafted Corey Coleman and is getting Josh Gordon back. Just getting a new coach and quarterback alone is worth watching, though admittedly the Browns do look like a major underdog to become a success in 2016. Houston, a playoff team with the 24th-ranked offense last year, has a decent shot with the additions of Brock Osweiler and Lamar Miller. But even with greatly improved offenses, none of those three teams really fit the mold of a Super Bowl winner.

We will conclude our study in Part III next week with a look at the sustainability of offense versus defense, as well as the traits of Super Bowl-winning defenses.