Building a Super Bowl Winner (Part I)

Building a Super Bowl Winner (Part I)
Building a Super Bowl Winner (Part I)
Photo: USA Today Sports Images

by Scott Kacsmar

While the NFL would love to promote every team as having an equal opportunity of winning the Super Bowl, we know the odds are rarely ever 1-in-32. Recent history is a solid predictor of which teams have a much higher or much lower chance of going the distance this season.

The Oakland Raiders are a trendy dark horse in 2016, but should you really trust a team that went 7-9 in 2015 and has not had a winning season since 2002? In the DVOA era (since 1989), 39 of the 54 Super Bowl participants (72.2 percent) made the playoffs in the previous season. Only six teams (11.1 percent) followed a losing season with a Super Bowl appearance. This sounded good a day ago for the Minnesota Vikings, another trendy 2016 pick, but the terrible news about Teddy Bridgewater's torn ACL puts a real damper on their season before it even starts.

But even if Bridgewater was healthy, are the Vikings really great enough on either side of the ball to be a top contender? Last season, Minnesota finished 16th in offensive DVOA and 14th on defense. Strong special teams (ranked fourth) aside, that is average at worst and balanced at best.

Is balance really the best model for building a championship team? A look at all Super Bowl winners or the recent DVOA dynasty in Seattle would suggest an obvious answer. However, in the salary-cap era, building and sustaining a team with great players on both sides of the ball is tougher. Only 26 teams in the DVOA era have finished in the top five in both offense and defense. The first three teams to do so (1989 49ers, 1991 Redskins, and 1992 Cowboys) all won the Super Bowl, but the last dozen teams have compiled an 0-4 Super Bowl record.

While balance sounds great, the Broncos did just win Super Bowl 50 with the No. 25 offense and No. 1 defense. The only other champion since 1989 with that much imbalance also featured Peyton Manning at quarterback, but the 2006 Colts paired the No. 1 offense with the No. 25 defense.

So how should a team go about building a Super Bowl winner today? There may always be alternative methods, but over this three-part study we will identify the trends that were most significant in helping achieve the ultimate goal of every NFL team.

The Process: Growing from Good to Great

Last week, I created a Twitter poll about this topic. The character limit made things difficult, but the question was which type of team you would prefer if your goal was to win a Super Bowl. The first choice was a balanced team with a top-10 offense and defense. The second choice was an imbalanced team that was top two on only one side of the ball (voter's preference). That could mean the most imbalanced team in the league, but I generally was thinking the weaker unit would rank in the 11 to 25 range. Again, no team since 1989 has been able to win a Super Bowl with an offense or defense ranked worse than 25th in DVOA.

The results were 60 percent in favor of the balanced team, which sounds reasonable given the actual results in the DVOA era. There have been 12 Super Bowl winners that ranked in the top 10 (top eight actually) on both sides of the ball in DVOA. A total of 10 imbalanced teams have won the Super Bowl, so that is very close, and DVOA estimates for the 1988 49ers (balanced) and 1987 Redskins (offense-dependent) seem unlikely to give either side an advantage.

The following table shows where each Super Bowl winner and loser since 1989 ranked in the regular season in DVOA (total, offense, and defense). Top-three units have their rankings in bold, while those ranking outside of the top 12 (the playoff field, essentially) are in red. We also included a key note on how the team was performing coming into the season in question. True "surprise" teams are highlighted in red while teams that were only in their second season of relevancy (read: winning) are highlighted in green.

DVOA Rankings for Super Bowl Teams, 1989-2015
Year SBW Total OFF DEF Note
Year SBL Total OFF DEF Note
1989 SF 1 1 5 7th-straight PO (repeat) 1989 DEN 4 13 4 3rd SB loss in 4 years
1990 NYG 1 7 4 3rd-straight 10-win year 1990 BUF 6 1 14 3rd-straight PO
1991 WAS 1 1 3 8-of-9 years w/10+ wins 1991 BUF 4 3 16 2nd-straight SB loss
1992 DAL 1 2 5 1991 NFC-DIV loss 1992 BUF 7 6 10 3rd-straight SB loss
1993 DAL 1 2 18 3rd-straight PO (repeat) 1993 BUF 12 13 8 4th-straight SB loss
1994 SF 3 1 7 12th-straight 10-win year 1994 SD 8 6 18 3rd-straight .500+ year
1995 DAL 2 1 13 5th-straight PO 1995 PIT 4 12 4 4th-straight PO
1996 GB 1 3 1 4th-straight PO 1996 NE 8 14 13 6-10 in 1995; PO in 1994
1997 DEN 2 2 8 1996 AFC-DIV loss 1997 GB 1 4 3 5th-straight PO
1998 DEN 1 1 20 3rd-straight PO (repeat) 1998 ATL 7 10 5 Last PO year: 1995
1999 STL 1 4 3 Last winning year: 1989 1999 TEN 5 3 20 8-8 in 1996-98
2000 BAL 3 22 2 8-8 in 1999 2000 NYG 11 8 12 Last PO year: 1997
2001 NE 11 11 13 Last PO year: 1998 2001 STL 2 2 5 3rd-straight PO
2002 TB 1 20 1 4th-straight PO 2002 OAK 2 2 7 3rd-straight PO
2003 NE 4 14 2 2nd SB in 3-year span 2003 CAR 16 18 10 Last winning year: 1996
Year SBW Total OFF DEF Note Year SBW Total OFF DEF Note
2004 NE 2 3 7 3rd SB in 4-year span 2004 PHI 6 9 16 5th-straight PO
2005 PIT 4 8 3 4th PO in 5 years 2005 SEA 3 1 16 3rd-straight PO
2006 IND 7 1 25 5th-straight PO 2006 CHI 5 20 2 2005 NFC-DIV loss
2007 NYG 14 18 13 3rd-straight PO 2007 NE 1 1 11 5th-straight PO
2008 PIT 4 21 1 4th PO in 5 years 2008 ARI 21 15 21 Last winning year: 1998
2009 NO 6 2 17 Last PO year: 2006 2009 IND 8 6 16 8th-straight PO
2010 GB 4 7 2 3rd PO in 4 years 2010 PIT 2 5 1 .500+ for 7th year in a row
2011 NYG 12 7 19 .500+ for 7th year in a row 2011 NE 3 3 30 3rd-straight PO
2012 BAL 8 13 19 5th-straight PO 2012 SF 4 5 3 2011 NFC-CG loss
2013 SEA 1 7 1 3rd PO in 4 years 2013 DEN 2 1 15 3rd-straight PO
2014 NE 4 6 12 14th-straight winning year 2014 SEA 1 5 1 3rd-straight PO
2015 DEN 8 25 1 5th-straight PO 2015 CAR 4 8 2 3rd-straight PO
AVG - 4.0 7.8 8.3
AVG - 5.8 7.2 10.5

We will get into the most fascinating results shortly, but for a general takeaway, it is hard to reach the Super Bowl without being a top-eight overall team in DVOA. Only three winners and four losers of the Super Bowl ranked outside of the top eight in DVOA. Not surprisingly, the winners were a little stronger overall than the losers, thanks mostly in part to an edge on defense.

As mentioned before, the 27 winners included 12 balanced teams, 10 imbalanced teams, and a handful of teams that did not fit either description. But when looking at the 27 losers, nine were balanced, four were imbalanced, and 14 would fall under an "other" category. Only one Super Bowl (XLI: Colts vs. Bears) matched two imbalanced teams. The only two matchups between "other" teams were both Giants-Patriots matchups, because DVOA was not impressed by those New England defenses, which twice allowed go-ahead touchdowns in the final minutes of what were generally low-scoring upsets for Eli Manning and company. The most common outcome was an "other" team losing six times each to balanced and imbalanced teams. Balanced teams were 3-3 against imbalanced teams.

Build a Dominant Unit

We have gotten used to the best overall team failing to win many championships in this era. Since 2000, only the 2002 Buccaneers and 2013 Seahawks won the Super Bowl after ranking No. 1 in total DVOA. But we have seen more instances of the No. 1 offense (six) or the No. 1 defense (five) coming through for a total of 11 rings. Ten more units ranked in the top three also finished the season on top, meaning 21 of the 27 Super Bowl winners fielded at least one top-three unit. For the losers, it was 15 teams, but that is still more than half.

In all six cases where a team won with a unit ranked in the 20s, the other side of the ball was always a top-two dominant unit. Only five of the 54 Super Bowl teams ranked outside of the top 10 on both sides of the ball. Technically, a team ranking 12th on both sides of the ball would be above average and balanced, but that type of team does not have much playoff success. Since 1989, 29 teams ranked in the 11-to-16 range on both offense and defense. Only 13 made the playoffs, compiling a 12-12 record, with the Patriots of 1996 and 2001 being the only teams to reach the Super Bowl. (This has been part of the problem for Marvin Lewis in compiling a historically bad 0-7 playoff record in Cincinnati. The rankings of his average playoff teams have been 13.6 on offense and 13.0 on defense. The three top-five units he did bring to the playoffs failed to show up on game day in the postseason.)

That stronger top-10 balance sure feels ideal, but only two of the last 10 champions fit that definition (2010 Packers and 2013 Seahawks). Before the turn of the century, eight balanced teams won from 1989 to 1999. Having confidence that at least one of your units will show up in grand fashion is very appealing. Even though the imbalanced winners were split among offense and defense, most fans would probably still pick the great defense. "Defense wins championships" carries validity, and 2015 only served as a reminder.

Rome Wasn't Built in a Day Either

Most of the dominant units that carried teams to championships were not built in one season. It can take years to put that type of talent together. There is also the idea of "learning to win" in the playoffs, where the intensity and importance of each moment is intensified. There are often-cited examples of this in other sports. Wayne Gretzky and the Edmonton Oilers needed five seasons to win their first title, advancing a step further each year and tasting defeat in the Stanley Cup Finals before finally winning the first of five titles in 1984. Michael Jordan and the Chicago Bulls needed seven postseasons to rise to the top of the Eastern Conference, finally vanquish the Detroit Pistons and go on to win six NBA titles in the 1990s.

Due to the NFL's shorter seasons and one-and-done playoff system, upsets and randomness are a lot more common. But there are still examples of teams that needed to overcome and learn from some heart-breaking defeats before realizing what it took to finish the job. The Ravens (2008-2012) and Broncos (2011-2015) just recently completed five-year journeys to rings. Were the 2012 Ravens and 2015 Broncos the strongest teams among those runs? No -- in fact, you could argue that the 2012 Ravens were John Harbaugh's weakest playoff team, and that the 2015 Broncos were the weakest Denver club since Manning arrived in 2012. But those were the years where things all worked out for both teams. Joe Flacco had a spotty playoff history early in his career, but led a very productive offense in those four games on his way to the richest contract in NFL history at the time. Denver learned a lot from the Super Bowl XLVIII ass-kicking to Seattle that a stronger defense was the way to go, and John Elway improved the roster accordingly over the next two seasons. Health at playoff time finally worked out for Denver and the rest is history. Much like Baltimore in 2013, the Broncos may fall off this year to 8-8 territory at best, but this is about winning one Super Bowl, not building a dynasty.

So it makes sense that most of these Super Bowl teams were relevant in previous seasons as well. Twenty-one of the winners and 18 of the losers were in the playoffs in the previous season. Most of the other teams were in that 7-9 or 8-8 range, and some even won nine or ten games but missed the playoffs. That's the NFL sometimes. The Giants missed the playoffs at 10-6 in 2010, but won the NFC East in 2011 with a 9-7 record. Only three teams reached the Super Bowl after losing 10-plus games the previous year: the 1996 Patriots, 1999 Rams, and 2001 Patriots.

Bill Belichick was on the coaching staff for the 1996 and 2001 New England teams, two of the 10 "surprise" Super Bowl teams in this study. The 1996 team, with Bill Parcells at head coach, was actually not that big of a surprise. New England made the playoffs at 10-6 in 1994 behind a young Drew Bledsoe, then added key players such as Ty Law, Curtis Martin, Ted Johnson, Terry Glenn, Lawyer Milloy, and Tedy Bruschi in the 1995 and 1996 drafts. Kicker Adam Vinatieri was an undrafted free agent in 1996. While the team disappointed with a 6-10 record in 1995, this was a young, talented roster with some of the best coaching in the league. Pete Carroll went 27-21 with New England from 1997-99 before Belichick took over the job in 2000, getting off to a slow 5-11 start. Everyone knows what happened in 2001 with Bledsoe and Tom Brady, and the rest is history. But a lot of the core of that team, including Willie McGinest, Troy Brown, and Otis Smith, was from the 1996 Super Bowl team.

Of course, the Patriots turned into a dynasty, but some of the role reversals on the above table are very interesting. The Rams were Cinderella in 1999, then Goliath in 2001 when losing to New England's David. Six years later the Evil Empire Patriots were the huge favorites, but the Little Giants took care of them, then did it again four years later. You almost want to thank Bill Belichick, the Patriots, and those resilient Giants for playing so many close Super Bowls since 1990.

[ad placeholder 3]

The other person to thank is Kurt Warner. Out of the 54 teams on that table, two should really qualify as batshit crazy for being there: the 1999 Rams and 2008 Cardinals. This is a huge part of Warner's Hall of Fame argument. The Rams had nine consecutive losing seasons from 1990 to 1998. After Trent Green was lost in the preseason, an unknown Warner stepped into the starting role and led one of the greatest offensive attacks in NFL history over the next three seasons, playing at his very best in 1999. After some injuries derailed him and teams tried to replace him with younger quarterbacks, Warner took the starting job back from Matt Leinart in Arizona during the 2007 season. In 2008, he did the unthinkable and led the Cardinals to the Super Bowl, taking a last-minute lead before Pittsburgh's Santonio Holmes made an incredible touchdown catch. Arizona had not had a winning season from 1999 to 2007, but Warner led all three of his teams to the Super Bowl in the seasons where he started all 16 games. The 2008 Cardinals (21st) are the only team to rank worse than 16th in DVOA to reach the Super Bowl, but that just adds to Warner's incredible legacy.

While Warner and Larry Fitzgerald almost brought the Cardinals a ring, we cannot forget the 2003 Panthers, with Jake Delhomme and Steve Smith put on similar heroics. At 16th in DVOA, the 2003 Panthers are the second-lowest ranked team in DVOA to reach the Super Bowl, and they too lost a heart-breaking Super Bowl. John Fox was hired in 2002 to turn around a 1-15 mess, and after a 7-9 finish, he benched Rodney Peete in the 2003 opener for Delhomme, who promptly led a comeback win. It was the first of eight game-winning drives for Delhomme that season, which is still tied for the NFL record.

When predicting teams for Super Bowl LI, it is best to focus on recent success and stable coach/quarterback situations. That means New England, Pittsburgh, Cincinnati, and Kansas City in the AFC, and Seattle, Arizona, Carolina, and Green Bay in the NFC. Can the other 24 teams crash the party? Sure, but that would be betting on things like Ben McAdoo fixing a defense better than Tom Coughlin could, Chuck Pagano getting a clue, or that Dak Prescott and Trevor Siemian are the next Warner and Brady. Stick with known quantities.

Once is an accident, twice is a coincidence, but three times is a pattern. Nearly three-quarters of these Super Bowl teams were relevant for three-plus seasons. While some teams may be able to catch lightning in a bottle, getting to the Super Bowl is usually just a great achievement for a team that has already done good things in past seasons.

Breaking Points and the Value of DVOA

Now comes the part where I aggravate several fan bases (and depress several more) by pointing out just how lucky they are to have these rings. But this was really one of the most interesting findings from the data, because it shows the benefits of using DVOA as a predictive tool.

Among the 27 Super Bowl winners, we noted that 21 had a top-three dominant unit. The six winners without a dominant unit -- the 1990 Giants, 2001 Patriots, 2007 Giants, 2011 Giants, 2012 Ravens, and 2014 Patriots -- are quite arguably the six most fortunate champions since 1989, winning by razor-thin margins in the postseason. I introduced the concept of "breaking points" in solving the Peyton Manning playoff puzzle earlier this year. A breaking point in a team context would be that one play late in the game where, had things gone the other way, the outcome would almost certainly reverse as well.

These six teams were pushed hard to the brink of elimination. Consider these facts:

  • Fifteen of the 27 champions needed a fourth-quarter comeback in the playoffs. Only the 1990 Giants, 2007 Giants, 2011 Giants, and 2014 Patriots needed two of them.
  • Seven champions trailed in the final two minutes of a playoff game: the 2006 Colts, the 2008 Steelers, and five members of our group of six, the 2014 Patriots excluded. And those 2006 Colts were already at the 11-yard line in a 34-31 game after the two-minute warning.
  • Eight champions trailed by at least four points in the fourth quarter of a playoff game: this group of six, the 2006 Colts (down 28-21 to New England, tied game with 13:24 left), and the 2013 Seahawks (who took the lead for good against San Francisco with 13:44 left).
  • Only two champions trailed by at least four points in the final minute: the 2007 Giants (who took the lead with 35 seconds left) and 2012 Ravens (who tied the game against Denver with 31 seconds left).
  • The 2001 Patriots and 2014 Patriots are the only champions to trail by two scores in the fourth quarter of a playoff game. New England trailed Oakland 13-3 in 2001 and trailed Seattle 24-14 in Super Bowl XLIX.

Let's take a closer look at each of these six teams:

1990 Giants (No. 7 offense, No. 4 defense): New York actually finished No. 1 overall in DVOA with a boost from its second-ranked special teams. In the NFC Championship Game in San Francisco, the 49ers had a first down at the New York 40, leading 13-12. Roger Craig fumbled and the Giants recovered with 2:36 left. Matt Bahr's 42-yard field goal with no time left gave the Giants a 15-13 win. In Super Bowl XXV, the Giants kicked a short field goal to take a 20-19 lead over Buffalo with 7:20 left. Scott Norwood infamously missed the 47-yard field goal wide right with four seconds left, the only do-or-die missed field goal in Super Bowl history. A Buffalo win would have given another No. 1 offense a championship, but the Bills never got closer than this play.

2001 Patriots (No. 11 offense, No. 13 defense): In the first playoff game of the Belichick era, Tom Brady's apparent season-ending fumble against Oakland was overturned by the now-discarded Tuck Rule, but the Patriots still needed Adam Vinatieri to deliver the greatest field goal in NFL history. He did, from 45 yards away in the snow, to force overtime, then won it with another field goal on the opening possession, which is another thing that would no longer be possible today. The Patriots then blew a 14-point fourth-quarter lead in Super Bowl XXXVI against the Rams before winning with a walk-off 48-yard field goal by Vinatieri. The 2001 Patriots still have the smallest postseason margin of victory (4.3 points per game) among the 50 Super Bowl winners. The 2001 Rams would have made for an elite champion, with two top-five units.

2007 Giants (No. 18 offense, No. 13 defense): The second-smallest margin of victory for a Super Bowl winner belongs to the 2007 Giants (5.0), the only team to register three game-winning drives in one postseason. After picking off Tony Romo in the end zone in Dallas, the Giants survived the cold and overtime at Lambeau Field, and authored a massive upset by beating the 18-0 Patriots in Super Bowl XLII. Eli Manning's broken sack and throw to David Tyree should go down as the greatest play in the first 50 Super Bowls. It is hard to see the Giants winning without that one. The 2007 Patriots were another No. 1 offense that flopped in the big game.

2011 Giants (No. 7 offense, No. 19 defense): These two New York wins may confuse people for decades, but it's not that hard to figure out. The Giants were generally a competitive team under Tom Coughlin, and often finished close games well with Manning at quarterback while featuring a strong pass rush. They were just inconsistent in the second halves of seasons, but put it all together for two title runs, even if they failed to win any other playoff games. Both Giants teams drew fairly weak opponents in the wild-card round (2007 Buccaneers, 2011 Falcons), then took down some juggernauts that they challenged well in the regular season. Yes, all six of those games were rematches from the regular season, with the Cowboys, Packers, and Patriots in 2007and the Packers, 49ers, and Patriots in 2011. That had to help immensely in gameplanning.

The playoff breaking points were not overly strong for the 2011 Giants, but they may have been the most fortunate team on the list just to make the playoffs. They are the only Super Bowl champion in 50 years to be outscored during the regular season. Manning had to lead six game-winning drives in the regular season alone to get to 9-7, including an epic 12-point comeback in Dallas in Week 14. Remember Miles Austin losing the ball in the lights and Dan Bailey, the most accurate kicker in history, having his 47-yard kick blocked at the end? Those Giants lived on the edge all year.

In the 2011 NFC Championship Game in San Francisco, it was déjà vu with more big 49ers fumbles. New York punted on 10 of its 12 second-half possessions, but scored twice on drives that covered 40 total yards and were set up by a pair of Kyle Williams fumbles. Williams' muffed punt led to a Giants touchdown, and his fumble on a punt return in overtime basically gifted the Giants a game-winning field goal. In the Super Bowl rematch with New England, Brady and Wes Welker could not connect on a big pass, but Manning and Mario Manningham did to set up New York's eighth game-winning drive of the season. Rob Gronkowski could not get to a Hail Mary in the end zone as time expired. The 2011 Patriots would have made some history with the 30th-ranked defense winning the Super Bowl, but the reality is that a unit that bad should have surrendered back-to-back game-winning drives in the final minutes. Our next team would certainly agree.

2012 Ravens (No. 13 offense, No. 19 defense): One annoying piece of revisionist history about this team is that people pretend the Week 12 San Diego game was a breaking point for Baltimore making the playoffs. Recall a dump pass by Joe Flacco to Ray Rice going for 30 yards on fourth-and-29, a stab in the heart to ALEX. Even with a loss, though, Baltimore still would have been 8-3 and in good shape. But the Ravens did slump a little before getting hot at the right time. Of course, the huge breaking point in the playoffs came in Denver in the AFC divisional round with the Ravens trailing 35-28 in the final minute. Flacco stepped up on a third down to deliver a bomb and Rahim Moore took his place in history with one of the worst misjudgments on a pass you'll ever see. A 70-yard touchdown to Jacoby Jones was the result, and the game was tied with 31 seconds left. Flacco had an interception dropped in overtime, but the Ravens held onto Manning's pick to set up Justin Tucker for a tough 47-yard field goal. This was redemption for the previous year when Lee Evans and Billy Cundiff failed the Ravens in New England at the end of the 2011 AFC Championship Game. Jones and Tucker came through, and the Ravens passed their toughest challenge that postseason with an unforgettable play. In Super Bowl XLVII, San Francisco had top-five balance, but needed a better throw to Michael Crabtree in the end zone. (They also needed one in 2013).

2014 Patriots (No. 6 offense, No. 12 defense): This era has featured the greatest comeback attempts in Super Bowl history, but most still went for naught as the other team usually won at the end. New England negated comebacks by both St. Louis (XXXVI) and Carolina (XXXVIII). Now on the other side of the coin, the Patriots had to come back from a 24-14 deficit against Seattle in Super Bowl XLIX. They did, but left 2:02 on the clock, an eternity for Russell Wilson. A ridiculous tipped catch by Jermaine Kearse down to the 5-yard line may have been Seattle's undoing, as it led to a second timeout taken and affected the endgame strategy. I stand by what I said the night of the game. If I'm Pete Carroll, once the ball got to the 1-yard line, I am running Marshawn Lynch three times in a row if I have to for that touchdown. The decision to throw a quick pass in congestion to Ricardo Lockette is unforgivable. The result was the costliest interception in NFL history as Malcolm Butler etched his name in NFL lore. In the beginning I mentioned 26 teams since 1989 with a top-five unit on both offense and defense. That includes three Seattle teams, but not the 2013 version, which was only seventh in offense but great at closing games on defense. Go figure that the 2013 team is the only Seahawks club with a ring to this point.

Most of these teams were not No. 1 seeds, so it was not surprising to see them struggle against some really strong teams. But I was just floored by how many of the closest calls were highlighted by the DVOA deficiencies. (Again, these DVOA ranks only account for regular-season games.) Some of those non-dominant units obviously made huge contributions in the playoff run, but the margin of victory was still so thin with these teams.

When I flipped things around and looked at which Super Bowl losers deserved better fates, I really did not find anything different from the examples already mentioned. Focusing on team strength, some of the stronger losers had really poor Super Bowl performances, including the 1992 Bills, 1998 Falcons and 2002 Raiders. Of course, I thought about the 32-point comeback Buffalo needed from backup Frank Reich in 1992 to keep the Super Bowl streak alive. I still get mad over Gary Anderson missing that field goal against Atlanta to deny the 1998 super matchup between Denver and Minnesota. And I still marvel at the 2003 Panthers and 2008 Cardinals, the weakest Super Bowl teams on paper, playing as well as they did in the postseason. The 2004 Eagles are a bit of a special case, ranking ninth on offense and 16th on defense for the season, but this was because they sat most of their starters in both Week 16 and Week 17. Going into those games, the Eagles ranked second in overall DVOA (fourth on offense and 12th on defense) before putting up DVOA ratings of -52.9% and -77.8% in big losses to the Rams and Bengals, both of whom finished 8-8 that year.

(Ed. Note: The 2004 Eagles are the toughest example of judging a team that sat starters, I have written multiple times in the past about why we can't simply drop these games from regular-season DVOA. First of all, we've actually found that DVOA is more predictive of future performance if we incorporate these "sit starters" games. Second, it's difficult to define what exactly counts as "sitting starters." Just the quarterback? Half the starters? What if the starters only play the first half? What if they come out after three quarters? What if they are playing but don't try very hard, which some people feel was the case with the Arizona Cardinals in Week 17 of last season? Where do we draw the lines? -- Aaron Schatz)

1996-2003: The NFL's Twilight Zone

Lastly, I wanted to draw attention to the cluster of weird results from the period of 1996 to 2003. Eight of the 10 surprise Super Bowl teams happened in this most unusual era of NFL history. These years may have given a false sense of parity gone wild in a new salary-capped NFL.

[ad placeholder 4]

Prior to this period, the NFC had the best teams and dominated the Super Bowl matchups, while the AFC had many of the best quarterbacks, albeit on flawed teams. After the salary cap started in 1994, there was more significant player movement around the league and rosters could be reconstructed more quickly. The older Hall of Fame quarterbacks also eventually started to decline and/or retire, including Joe Montana, Jim Kelly, John Elway, Steve Young, Dan Marino, Warren Moon, and Troy Aikman. This is why we started to see quarterbacks coming out of grocery stores, NFL Europe, the XFL, Canada, and the later rounds of the draft to fill those spots. The quarterback market is nothing like that today.

The 1995 season also brought expansion for the first time in two decades, and the Jaguars and Panthers both shocked the world by reaching their Conference Championship Games in 1996. We said goodbye to the Browns, hello to the Ravens, and hello again to the Browns in 1999. That expansion team was a nightmare, as were the early years of football's return to Houston with the Texans in 2002. Things soon normalized with the current setup of 32 teams spread out across eight divisions.

This wild era did not continue once the passing game took over again starting in 2004. It was that year when the NFL reinforced illegal contact and there was a great draft class with Eli Manning, Philip Rivers and Ben Roethlisberger. Drew Brees broke out in San Diego while Peyton Manning, Tom Brady, Donovan McNabb, and Daunte Culpepper all had the best seasons of their careers up to that point. The only "surprise" Super Bowl teams since then had Kurt Warner and Drew Brees at quarterback, so in hindsight they shouldn't really have been surprises. New Orleans quickly rebuilt in 2006 a year after Hurricane Katrina by bringing Sean Payton, Brees, Reggie Bush, Marques Colston, Roman Harper, Jahri Evans, etc. to town.

While the last decade has produced huge Super Bowl upsets and unprecedented title runs by sixth seeds, the types of teams getting to the big game look more in line with those of the pre-cap era. They are talented and usually have a dominant unit and stability at quarterback, and likely got close in a previous tournament.

Winning the Super Bowl is a process, and most teams will not succeed by skipping steps. Signing a bunch of free agents is not likely to turn your 6-10 finish into hoisting the Lombardi Trophy the next year -- though it might in a couple of years, if you continue to build in the draft and make some other good choices. When I start to write a team's essay in Football Outsiders Almanac, I first consider what kind of team we are looking at. If it's a Jeff Fisher special that is always stuck near 8-8, then I focus on what the team needs to do to make the playoffs. Don't worry about the Super Bowl. Advance one tier and make the playoffs. If I was writing about Marvin Lewis, who can get Cincinnati to the playoffs but never win there, then I would focus on finishing games in January against the better competition. Take one step at a time. Respect the process.

So much has to go right for a team to win one Super Bowl, including health and some good luck or fortuitous bounces of the football. But the teams that win have already put in a lot of work to get to the point where they have a shot at the championship.

In Part II of this series, later this week, we will look at the offensive units of Super Bowl teams. How long does it take to become a great offense, and how do you best build one? Can an offense significantly improve in one year despite few changes to personnel or coaching? We will then conclude with a similar study on defenses in Part III next week.


68 comments, Last at 06 Sep 2016, 1:53pm

#1 by TimK // Aug 31, 2016 - 12:45pm

Thanks Scott, interesting read.

Look forward to reading the remaining two parts.

Points: 0

#2 by Aaron Brooks G… // Aug 31, 2016 - 12:52pm

Why was the Vinatieri's kick against Oakland the greatest FG in NFL history?

It's not Matt Bryant's bomb to deny one of the great 4th quarter comebacks in NFL history?

It's not Nick Folk's and Adam Vinatieri's dueling last minute do-or-die FGs that cemented Manning's loser legacy and established the Sanchize as a clutch playoff QB immune to any future legacy changes?

Points: 0

#3 by Scott Kacsmar // Aug 31, 2016 - 1:33pm

I'm welcome to challengers for that title, but these wouldn't have been too high on the list for me. A big part of the greatness for Vinatieri's FG is that the ball just disappears once he gets it up. The visibility was that bad from the snow. Sure, maybe you can blame the lack of HD broadcasting at the time for the picture quality, but that kick was incredible. 45 yards still isn't a gimme, and it was even lower percentage back in 2001. That kick saved a season in the divisional round and we know what it led to. If they go one-and-done at home in a 13-10 game, do we see Drew Bledsoe starting in Week 1 of the 2002 season? I can't think of a successful kick with more significant consequences, and the kick itself was really impressive.

Points: 0

#5 by Travis // Aug 31, 2016 - 2:50pm

Some contenders:

1958, Giants vs. Browns: Pat Summerall's 49-yarder through the snow in the season-ender gives the Giants a 13-10 win and forces a tie for the Eastern Division title. The Giants won the playoff rematch the next week (there were no tiebreakers or wild cards), then lost the NFL Championship to the Colts (the so-called Greatest Game Ever Played, the NFL's breakout game, surely helped by talking place in New York). A plus if this in any way led to the Summerall-Madden broadcast team.

1970, Saints vs. Lions: Tom Dempsey hits a last-second 63-yarder. Insignificant game (the Saints were 1-5-1 at the time and finished 2-11-1), but 7 yards longer than the previous record, on grass, and at sea level. Did lead to rules changes on the construction of kicking shoes.

1990, Giants vs. 49ers, NFC Championship Game: Matt Bahr's last-second 42-yarder ends the 49ers' bid for a threepeat. Unknown future consequences with a miss (among them: Does the Giants coaching staff still leave en masse? Does Steve Young, filling in for the injured Joe Montana, have a breakout game in the Super Bowl? Does Buffalo win?)

Points: 0

#65 by LionInAZ // Sep 03, 2016 - 5:51pm

Dempsey's FG had so many question marks it should not be counted as remarkable. There were questions about whether his special boot was legal, and the Lions were so astonished at the attempt that they didn't even attempt to block it. Some have speculated that Earl McCullough might have been able to block it at the crossbar, but I don't anyone believed it would even go that far.

Points: 0

#7 by SandyRiver // Aug 31, 2016 - 4:24pm

Vinatieri's kick was actually mis-hit a bit, resulting in a line-drive duck that should've been eminently blockable. That may have helped in the cameras not tracking it well. Pats' line must have done a good job, or maybe the slippery conditions kept the Raiders from jumping well. And it was a hurry-up play with almost no time to clear a spot for the ball. Perhaps Vinatieri's view of the ball was partially obscured by a snow pile, and thus he contacted the ball a bit higher than desired. (Unlike his much shorter game winner, when Oakland's dumb "icing" TO gave the Pats time to clear a patio-sized area for him to work in.)

I can still recall the pic of Summerall's kick just inside the left upright, and the SL headline, "49 yards and one foot." The Giants' mastery of Jim Brown in those two games was epic. He took it 65 yards to the house on the 1st play from scrimmage in the 13-10 game, then he and the Browns scored only a figgie over then next 1:19:30. He had something like 12 yards on 10 carries in the 10-0 playoff.

Points: 0

#14 by Aaron Brooks G… // Sep 01, 2016 - 8:35am

I just want clarity on what the criteria are? Greatest leverage? Because it wouldn't be that one -- that just tied the game in a division round game. Vinatieri might have the greatest leverage FG kick -- he broke two SB ties late in games -- but it's not the Tuck Game kicks. Hell, it's probably not the greatest FG of those playoffs.

Is it relative difficulty? You seem to imply that. I'm not sure I hold the league-wide improvement since 2001 against Bryant, though. Was Jordan too good of a player for the clincher against the Cavs to be iconic? Has 3-pt shooting gotten too good for Curry to be special?

Points: 0

#4 by theslothook // Aug 31, 2016 - 2:15pm

Nice work Scott. Truly lots went into this. You pretty much covered most of my thoughts on this subject.

I do think Seattle to me is the quickest bad to insta contender I can remember. They were 6-10 and within a season, became 10+ winning juggernaught - mostly due to really hitting the draft all at once and getting some real quality free agents at a bargain.

Maybe the 90s cowboys are a similar example of a dramatic all at once ascension.

Points: 0

#20 by pitt1980 // Sep 01, 2016 - 12:34pm

Rams went 4-12 in 1998

signed Trent Green to be the new QB in the offseason, drafted Torry Holt #6 overall, and traded a 2nd and 5th round pick for Marshall Faulk

a preseason injury to their starting QB later...

13-3 and Super Bowl champs


which raises an interesting question,

where should paradigm shifting coaching strategy changes fit into your 'how to win Super Bowl theory'?

Should the Greatest Show on Turf count as one?

How about the Legion of Boom defense? Is what Seattle came up with different than what teams before them came up with? (seems like they saw some value in big corners that other teams didn't)

how about whatever iteration of whatever Belicheck is up to?

(or maybe we retrofit those narratives on after the fact, idk)

Points: 0

#21 by Will Allen // Sep 01, 2016 - 12:49pm

There wasn't anything revolutionary about the Martz offense. It really was just sub-type of the Air Coryell attack, differing from the Gibbs sub-type in that it didn't put the same premium on protecting the qb. It helped to have Orlando Pace and other good linemen, and a great qb, and great receivers. When Martz tried the same approach with lesser personnel, it's drawbacks became rather more apparent.

Points: 0

#6 by sharky19 // Aug 31, 2016 - 3:12pm

I feel like the 2000-2001 AFC Championship is one of the least talked about "what if" games. If Siragusa doesn't commit felony assault against Gannon, the Raiders have a punchers chance in that game, and most likely wipe the floor with the Giants. Add on the Tuck Rule, and playing against TB with a coach who knows all your plays, and 2000-2002 for Raiders has to be one of the most heartbreaking stretches for a team in (semi) recent memory, not even considering the 15 years of hell afterwards.

Points: 0

#25 by ramirez // Sep 01, 2016 - 3:50pm

The Tuck Rule call was made correctly. Continuing to use it as an excuse is just sour grapes. The 2000, 2001, and 2002 Raiders were all good teams, but each of them was beaten by a superior team in the playoffs. You can come up with what if? scenarios for any team that loses in the playoffs.

Points: 0

#26 by Will Allen // Sep 01, 2016 - 4:02pm

When it gets to the point that you are using a miniscule hand movement to define "superior team", well, that's just silly.

Points: 0

#28 by theslothook // Sep 01, 2016 - 4:21pm

Which then tells you - if the result of winning and losing can be altered by one or two plays, how certain can you be that that the superior team won?

Were the giants of 2007 the best team in the nfl? How about the broncos this year?

Frankly, if you're going to be on this site, you really should get acquainted with randomness and how much it can alter common perceptions.

Points: 0

#29 by ramirez // Sep 01, 2016 - 4:28pm

First, the Tuck Rule was called correctly. That's not my opinion, its a fact. There's a replay which you can find on Youtube which shows very clearly that Brady's arm was moving forward when he was hit by Woodson.

I said that the Raiders were beaten by 3 better teams, and statistically, this is true. I never claimed that the 2007 Giants were a better team than the 2007 Patriots. A manufactured controversy over a correct officials call doesn't alter the fact that statistically, the 2001 Patriots were a better team than the 2001 Raiders.

Points: 0

#35 by ramirez // Sep 01, 2016 - 5:43pm

Right, I'm aware that DVOA rates Oakland higher. But if you go by points scored and points allowed, either per game or per drive, New England had a better combination of offense and defense. Amd since I'm supposed to believe that Vinatieri pretty much won that Super Bowl title single-handedly, I think we can probably give the Patriots the edge on special teams, too.

If you look closely at the DVOA ratings, you'll see that it rates Oakland's offense as much better than New England's, but it also says that the two defenses were of equal strength. But that's a purely stat-based analysis, that misses a key point about that Patriots team. NE was a much better team at the end of the 2001 season than at the beginning, so a season-wide analysis probably underrates them. And one of the reasons why the Oakland offense was better is because unlike Gannon, Brady wasn't throwing to Jerry Rice and Tim Brown.

Even if you ignore the point differential, and insist that DVOA makes Oakland the better team, it doesn't alter the point from my original post. Since the Tuck Rule was called correctly, using it as an excuse for Oakland's failure to win the SB amounts to sour grapes.

Points: 0

#37 by theslothook // Sep 01, 2016 - 5:52pm

I was arguing the notion of superior teams. DVOA is a much better statistic for measuring overall quality than point differential. Drive stats are in the right direction too, but then both miss the context of opponent adjustments and field position.

Lots of analysis has been done to show full season ratings are a much better predictor than half season ratings. The more you truncate the season, the more you open yourself up to small sample bias. Sure, you may believe NE was a much better team by the end of the year(I'm sure you're not being persuaded after the fact given that they won the SB); but thats a subjective opinion. I can easily say I thought the raiders were the greatest team of all time and that's just an opinion as well.

Also I don't understand the relevance in bringing up Gannon's receivers. This is a discussion about which team was better, not what you think of the quarterbacks.

Points: 0

#39 by ramirez // Sep 01, 2016 - 6:19pm

The point about the WRs is a side issue. But if you're saying that we can't use points to measure offense and defense, then I'd assume you would agree with me that the New England defense was below average that year. That's what DVOA says. Which is funny, because I thought Brady won those early Super Bowls because of his amazing defenses.

The point about the Patriots improving is something I believe, because I watched that team very closely. It's supported by what the Rams coaches said after NE played St Louis very tough in the regular season. Behind closed doors, the Rams coaches said they wouldn't be surprised if they saw New England again in the Super Bowl. Trust me, the 2001 Patriots looked nothing like that at the start of the season. Its obviously possible for a team to improve significantly over the course of the season. What I'm saying is that circa January 2002, the Patriots had a better team, top to bottom, than the Raiders. But if you're going to insist on looking exclusively at season-long DVOA rankings, I guess we'll just have to disagree. I think that's an incomplete analysis in the case of the 2001 Patriots, for reasons that seem pretty obvious to me.

Points: 0

#41 by theslothook // Sep 01, 2016 - 6:55pm

I'm not sure why I am arguing with you about this given that I'm not a Raider's fan, but here goes.

I think you are misunderstanding what I'm saying.

First of all - a defense can be awful most of the year, but if plays well in the playoffs - then it plays well in the playoffs. The converse is true as well. As a result, it is still logically consistent to say a team had a poor defense overall, but it played well in the playoffs and carried the team. Those two things are not mutually exclusive. I realize this is going to start another discussion about how great brady was or wasn't in the 2001 postseason, so let me stop that discussion before it happens. I have no interest in debating that.

Now to your second point. I can buy the argument that you think the pats improved in the second half; including all the testimonials, but the fact that second half splits do a worse job of predicting things, tells me that such improvements can sometimes be mirages or random variance more so than actual quality of play improvement. You may find that answer unsatisfying / unconvincing, but since its what the data shows; I'm inclined to believe it.

Now to NE. Looking at the splits, there was definitely an improvement in Total DVOA between the first half and the second half of the season, but probably not in the way you are thinking. Total dvoa jumped from -2% to 19%

The defense remained unchanged between both periods. The pass offense remained unchanged. The big jump came from the running game and special teams.

So basically, the whole NE turnaround wasn't tom brady getting better or the defense hitting harder, it was better special teams and a better run game. I wonder if that's what you remembered in the Pats second half turnaround.

Points: 0

#44 by ramirez // Sep 02, 2016 - 12:05am

Its worth remembering that this whole discussion is a follow-up to my assertion that the Raiders lost to 3 superior teams in the playoffs. In order for that to be accurate, all I have to do is show that when the Raiders and Patriots faced off, New England was the better team. I believe that's true. It's because I don't really care if the data shows Oakland was a better teams in September, in fact I agree that they were. But I don't understand your argument that the first half of the 2001 season does a better job of predicting playoff outcomes than the 2nd half. You have conceded that New England's running attack and special teams improved, and the other units stayed pretty much the same. That appears to confirm what I'm saying, that New England, as an overall team, improved over the course of the season, and by playoffs was a better squad than Oakland.

That claim is backed up by the raw numbers, which show that NE had a much better point differential in the last 8 games than in the first 8. They finished the season on an 8-1 stretch, with the only loss coming to the Rams. By contrast, Oakland started 6-1, but finished on a 2-4 stretch, although I recognize that those losses were all by small margins. Then they beat an unremarkable Jets team at home in the wild card round, a team the Patriots had also beaten at home. I don't see why that should dissuade me from believing that the Patriots were a better team by the end of the season.

I get that you don't like the Patriots, but pushing a narrative that presents the 2001 Patriots as a bunch of lucky frauds is disingenuous and wrong.

Points: 0

#45 by theslothook // Sep 02, 2016 - 12:24am

As I mentioned in the beginning, I am not a Raiders fan so I have no axe to grind here. I simply pointed out - full season numbers suggested NE was not the superior team, not first half numbers like you suggested. Second - you keep skating past the issue I keep bringing up. That second half numbers do NOT predict future performance. As in, when we look at second half splits like you are doing for NE, they are less informative than if we looked over the entire season This is not altogether surprising given the reasons I said above. For every 2001 NE, there are lots of other teams that improved in the second half that amounted to nothing. Its fun to pretend NE is the lone exception to the rule because you happened to think so, but statistics don't really care about such value judgements.

For the record, as a practicing statistician, that's why we have statistics. If we believe something is true; fine, then it should show up in the data. If it does not, you either rethink your theory or redo the design of the study. The worst outcome is to say, numbers be damned, I am right and I don't care what anyone else says.

"I get that you don't like the Patriots, but pushing a narrative that presents the 2001 Patriots as a bunch of lucky frauds is disingenuous and wrong."

This statement sort of epitomizes the whole lack of understanding on your part. Where anywhere did I say that I thought they were lucky frauds??? If other patriot fans believe that's my agenda, feel free to second Ramirez's point.

Points: 0

#46 by ramirez // Sep 02, 2016 - 12:40am

I have to say, I'm baffled by some of the opinions I see on here. I understand your argument, but I don't agree with it. You're leaning on the idea that the 16 game sample for the 2001 Patriots is a better predictor than the last 8 games of the regular season. If there wasn't a clear pattern showing the improvement of the Patriots over those 16 games, your argument would make sense. But the Patriots are CLEARLY an example of a team that improved between the start and end of the season.

I never said that the Patriots are the lone exception to the rule. There are other teams that reached the Super Bowl that weren't highly rated in the preseason. I've already used numbers to make my argument, that make your stubborn refusal to admit that the Patriots improved amusing to me. You're arguing that there are no exceptions to the sample size rule, and I find that baffling. How can that be the case? How can there be zero examples of teams that demonstrably improve, and for whom the last 8 games is a better indicator than the first 8?

Your case that Oakland was the better team leans entirely on the fact that their season long DVOA numbers are superior, and completely ignores context. It just seems narrow-minded and self defeating. If you refuse to accept the idea that the quality of the two teams evolved over time, I guess your views on this question will never change.

Points: 0

#47 by theslothook // Sep 02, 2016 - 12:50am

Claim: People have looked at second half numbers and tried to regress that on future performance. They found a weaker correlation than using full season numbers; suggesting full season numbers are a better predictor of future performance than half a season's worth. In other words, a full season's worth of data is a much better representation of a team's quality than just looking at the second half.

If you feel that statement is wrong, I encourage you to do some research yourself and post the results.

Points: 0

#48 by ramirez // Sep 02, 2016 - 1:05am

"They found a weaker correlation than using full season numbers; suggesting full season numbers are a better predictor of future performance than half a season's worth. In other words, a full season's worth of data is a much better representation of a team's quality than just looking at the second half."

I agree with this. As a general rule, most NFL teams will be cases where a 16 game sample is more reliable than an 8 game sample. But you're saying that there are NO exceptions to this principle. I believe there are exceptions, and that the 2001 Patriots are an example of such an exception. I don't see what is controversial about that.

Imagine a team that won it's first 14 games in dominant fashion, outscoring opponents by 30 points a game, but suffers a rash of injuries to all their best players with 2 weeks to go. Then they lose two in a row, and head into the playoffs with backups starting in place of the injured superstars. theslothook would argue that that team's season-long DVOA numbers, which would be exceptionally good, would be a better predictor of their playoff fortunes than the last 2 weeks of the season. I have to say, I find that position baffling.

Points: 0

#49 by theslothook // Sep 02, 2016 - 1:19am

Actually, my position is more nuanced. My position is, its hard to distinguish improved second half play vs random noise.

Consider, you flip a coin ten times. The first five are roughly 50 50, the next 5 are like 4 tails and 1 head. Do you automatically say therefore that the first 5 flips were a fair coin and the second 5 flips were from a rigged coin?

In any case, I doubt this argument will convince you so I'm going to conclude that we just agree to disagree

Points: 0

#50 by ramirez // Sep 02, 2016 - 1:38am

But the coin flip example doesn't work, because that's an activity that over the long haul, produces 50-50 outcomes. But the fortunes of football teams don't work like that. They're not just the result of a series of coin flips.

This reminds me of an argument I once had with a guy who claimed that Peyton Manning outplayed Brady in the 2004 playoffs. His reasoning was that Manning produced a significantly better anypa figure in his 2 games than Brady did in his 3 games. And in purely statistical terms, that's true. But what the numbers don't tell you is that Manning was fantastic one week, and then subpar the next. Brady had lower highs and higher lows, but was much more consistent, and produced significant value in all 3 games. Its obvious to me which performance is more valuable to a team in a one-and-done playoff format, but I was unable to convince the guy to agree with me. Stats are very useful, but if they're not put into proper context, it can lead to faulty conclusions.

Points: 0

#53 by RobotBoy // Sep 02, 2016 - 6:47am

I don't disagree with you that the Patriots improved over the course of the season or that larger statistical analyses have trouble measuring such improvements, at least in terms of first half/second half performance against full season. (Although, there are ways to look at how teams improve over the course of the season, and examine if that is relevant to postseason outcomes - record, changes in point differential, etc. We hear a lot about how 'teams get hot' or 'go on a run.' Must be a way to clear out the noise and see if there's anything there besides anecdotes).
That said, how can Brady in the post-season have had 'higher highs and lower lows' than Manning, AND been more consistent? Those seem to be mutually exclusive claims. Perhaps you didn't fully develop your point there.

Points: 0

#56 by ramirez // Sep 02, 2016 - 9:29am

Robotboy, you misquoted what I wrote. If you go back and read it again, you'll see it makes sense. Slothook's problem is that he relies exclusively on DVOA numbers, and then declares that Oakland's better ranks, which is really just a better offensive rank, makes them a better team overall. And I think that's an incomplete answer in cases like this one. The 2001 Raiders just weren't that great of a team.

Points: 0

#60 by theslothook // Sep 02, 2016 - 12:59pm

The 2015 Texans "improved" in the second half, but their postseason meltdown was more indicative of their full season statistics than their second half splits. This is the biggest problem I have with these arguments. Until you have a consistent logical way of distinguishing real second half improvements from random noise, you have nothing. And you can be certain of nothing.

And while you are it, please also explain why the improvements were relegated to the running game and special teams, yet the pass offense and the defense remained unchanged.

This whole argument feels cooked up way after the fact. If the tuck rule gets assessed by today's standards, NE loses that game and we're never even discussing this perceived second half improvement that made them a better team than the raiders.

Points: 0

#62 by ramirez // Sep 02, 2016 - 3:33pm

I haven't looked at the numbers for the 2015 Texans. But let's assume that like the 01 Pats, they had better point diff numbers in the 2nd half. That would make me feel better about their chances against KC than otherwise, but it wouldn't necessarily lead me to believe they were the favorites, unless I was convinced they were a better team than KC. We also both know that it's virtually impossible to accurately predict the NFL playoffs. My claim that NE improved in 2001 isn't based on their playoff results. It's based on their reg season numbers.

Now, you reject that position because it's not based on DVOA. My problem with that is that I don't think DVOA is completely reliable, either. DVOA says that last season, Tyrod Taylor, Matt Stafford, and Jay Cutler all had better years than Cam Newton. The hell they did. I believe DVOA significantly undervalues Newton's 2015 season, and at least one of the writers at FO agrees with me. DVOA was designed to be a tool to help us understand the game. It wasn't designed to be the Holy Grail of football stats, and treating it as such is a mistake. Like with any individual stat, there are things that will show up on game film that DVOA will miss. That's why a purely statistical analysis is never complete.

Points: 0

#63 by theslothook // Sep 02, 2016 - 3:51pm

DVOA measures team wide statistics. I agree, using it to judge players is pretty wrong. With that said, the bills, lions, and bears all had worse passing dvoas than the carolina panthers.

Finally - dvoa and numberfire are the only places where the statistics adjust for the context you are talking about. If you have another stat that is better reflective of team performance(besides you're subjective opinion) - please pass that to me and explain why its better.

Points: 0

#64 by ramirez // Sep 02, 2016 - 5:32pm

I thought I would follow up what I wrote about this by including something I found in an article written by Aaron Schatz about the 2001 DVOA ratings.

"In Weeks 1-7 of 2001, the Patriots had a defensive DVOA of +10.1%. In Weeks 8-17, they had a defensive DVOA of -14.3%. In the playoffs, they had a defensive DVOA of -23.5%."

For anyone who doesn't know, a negative DVOA rating means a good defense. But you're right slothook, Brady's passing stats actually went down in the 2nd half of the 2001 season. His passing stats weren't that different, but he took a huge number of sacks which hurts his anypa figure. He was better in the playoffs than in the last 8 games of the 2001 reg season by anypa, and of course his individual numbers have improved massively in the years since 2001.

Points: 0

#54 by winslow82 // Sep 02, 2016 - 7:17am

Ramirez, while I agree with you that not every single time can we go with year long stats to asses a playoff teams current strength, you have to be careful with Brady vs Manning comparisons in the 2004 season. Brady had an incredible edge by having a video camera stealing opposing defensive signals and Ernie Adams and his photographic mind breaking it all down at halftime or a previous matchup.

In case you don't recall, at some point late in the playoff run of 2004, Brady was asked how he could beat such great defenses like Pittsburgh when few others could. Brady responded to the on field reporter by saying "when you have the answers to the test before you take it, it's easy.

Also , following the Eagles loss to the pats in 2004, an Eagles corner said "everytime we called a blitz , the patriots called a screen pass. Every time,!"

Points: 0

#55 by ramirez // Sep 02, 2016 - 9:22am

If you want to believe in conspiracy theories while providing nothing but anecdotal evidence, go over to Infowars and please stop wasting my time.

Points: 0

#43 by Grendel13G // Sep 01, 2016 - 8:38pm

Wow, somehow I've never really put it together and realized how crushing these three years plus aftermath were for the Raiders.

Points: 0

#52 by Bright Blue Shorts // Sep 02, 2016 - 2:44am

Or you could look at it the other way round that those three years were an oasis in the desert. Raiders hadn't been relevant for 5-10 years.

Points: 0

#9 by eagle97a // Aug 31, 2016 - 10:23pm

I might be in the minority here but I stand by Pete Carrolls' decision to pass the ball during that fateful play. It was a defensible call and they had some success running that pick play on other occasions.

Points: 0

#11 by eagle97a // Sep 01, 2016 - 3:16am

No one talks about the pick but everybody will be talking about SB god Russell Wilson throwing a TD on that play and leading that improbable drive. The throw could have been a bit more to the left but under the circumstances it was a good tight throw and Browner deserves more of a mention for a good jam that allowed Butler to drive thru to make the pick.

Points: 0

#12 by panthersnbraves // Sep 01, 2016 - 7:13am

It's a bit fuzzy, but I seem to recall seeing a video where the Patriots talk about having watched film and seen a similar play, so Butler knew exactly where to be. Great coaching and execution.

Points: 0

#17 by Will Allen // Sep 01, 2016 - 10:18am

That's the problem with the playcall. Browner is not a great player, but he is extremely well suited to jam a receiver at the goal line. It's probably what he is physically best able to do. You don't call plays to attack the opponent's strengths, you do so to attack the opponent's weaknesses.

Frankly, that entire drive by the Seahawks, as successful as it was until the int, showcased the fact that Belichik had simply done a better job of preparation than Carroll. Disorganization and failure to think through scenarios well ahead of time caused Carroll to waste precious time. Not getting the right physical matchup on the play that led to the int was just the last of the coaching screwups that Carroll displayed on the drive.

Points: 0

#23 by Vincent Verhei // Sep 01, 2016 - 2:40pm

I agree 100 percent with everything Will wrote here. And to go further with this paragraph:

"That's the problem with the playcall. Browner is not a great player, but he is extremely well suited to jam a receiver at the goal line. It's probably what he is physically best able to do. You don't call plays to attack the opponent's strengths, you do so to attack the opponent's weaknesses."

I did a whole thing on this in FOA2015, but Seattle was second in the league in short-yardage rushing offense that season, and even had success on plays when they went three-wide and opponents stuck with goal line personnel. New England, meanwhile, had been the league's worst short-yardage rushing defense. With a minute to go, the Seahawks still had tons of time to do what they did best (power rushing) against what New England did worst (stopping power rushing) THREE TIMES, but instead opted to throw a pass to "maximize the number of plays" or whatever their excuse was.

Points: 0

#24 by Scott Kacsmar // Sep 01, 2016 - 3:43pm

Just wanted to add an old article for more on this situation -

Offenses with a first down inside the 5-yard line in those situations get the TD 75% of the time.

Points: 0

#27 by theslothook // Sep 01, 2016 - 4:18pm

The problem I have with this argument is two fold. First - while statistical splits might show the Seahawks were better off running than passing because of their strengths vs ne's weaknesses, such analysis might lead you into thinking Seattle should only do power runs the entire game.

In fact, game plans are adaptive and game theory is very much at work here. if NE is crowding the line of scrimmage expecting a run, then maybe the optimal play IS to throw the ball.

Secondly - its also unfair to blame the play because it didn't absolutely maximize the given mismatch. But the same argument could be made that throwing to anyone but Gronkowski is not "maximizing" the strategic advantages. After all, everyone including BB calls plays that aren't the Most optimal. It just happened to be this one play where we can discuss it over and over.

Frankly - throws on the goal line should be low and away. Even if Butler reads the play correctly, a low throw in that situation where the pick happened is nearly unguardable and at worst is an incomplete pass. It really was a combination of not a very good throw and a good defensive play. The play call to me is a secondary issue.

Points: 0

#31 by Will Allen // Sep 01, 2016 - 5:20pm

The problem was not that is didn't maximize what Seattle does best, it was that it maximized Bronwer's ability to do what he does best. Yes, there are times when being counterintuitive can make sense, but I assure you, Bill Belichik is not going to the Hall of Fame by playing to what his opponent does best.

Points: 0

#33 by Vincent Verhei // Sep 01, 2016 - 5:33pm

When I say "power runs" I am speaking situationally, not tactically. By power runs, I mean anything with 1 or 2 yards to go.

Now, as for this:

"if NE is crowding the line of scrimmage expecting a run, then maybe the optimal play IS to throw the ball."

This is factually untrue, and I have the numbers to prove it.

The Seahawks had 79 short-yardage plays in 2014. On 56 runs, they had a 75 percent success rate and 28.2% DVOA. On 23 passes, they had a 43.5% success rate and -16.4% DVOA.

What about when Seattle spread the field with 3 or more receivers in short-yardage? On 33 runs, they had a 72.7 percent success rate and a 32.6% DVOA. On 15 passes, they had a 46.7% success rate and a -10.8% DVOA.

And what about when Seattle spread the field in short-yardage and opponents stuffed the box, creating one-on-one matchups outside? In 18 runs, they had a 72.2% success rate and a 50.2% DVOA. On seven passes (admittedly a teeny-tiny sample size) they had a 42.9 percent success rate and a -28.2% DVOA.

In any situation, those Seahawks were an unholy terror in short-yardage rushing, and a pretty crappy unit in short-yardage passing. Now, remember that New England was also THE WORST SHORT-YARDAGE RUN DEFENSE IN THE LEAGUE. Instead of exploiting that mismatch, the Seahawks asked Jermaine Kearse to beat the jam of the biggest cornerback in the league, to clear space for their fifth wide receiver to make or break their season. To say that the pass was a good play call is to say that it was as likely or more likely to succeed than a run, which is statistically completely inaccurate.

I just realized that the Seattle chapter was available as a sample last year, so here you go if you want to read the whole thing.

Points: 0

#38 by theslothook // Sep 01, 2016 - 5:56pm

Fair enough Vince, you clearly thought a lot more about this issue than I did. I tend to think of league wide generalities, but sure - in this specific matchup, maybe its best to run right into the teeth of a defense no matter how many defenders they have stuffed at the line.

Points: 0

#34 by Vincent Verhei // Sep 01, 2016 - 5:38pm

On another note, it's not fair to put all the blame for the loss on that one play. The Seahawks had the ball with a ten-point lead in the fourth quarter -- twice -- and couldn't kill the clock or get insurance points, and their defense couldn't keep New England out of the end zone. That's the real reason they lost, not the results of one goal-line play.

In a similar note, I get annoyed when Scott pins Minnesota's playoff losses on the missed field goals of Gary Anderson and Blair Walsh, as if nothing else happened that day. Give credit to Atlanta's defense for limiting Minnesota's record-setting offense to one touchdown in the second half, plus two drives in overtime. And even after Anderson's miss, the Falcons had to drive 80 yards for a touchdown, and then score the winning overtime points. And last year, let's remember that if the Vikings had scored even one touchdown all day, then Walsh never even needs to kick at the end.

There are about 150-ish plays in an NFL game. Let's not pretend that the last one is the only one that matters.

Points: 0

#36 by Will Allen // Sep 01, 2016 - 5:45pm

I thought the biggest element of the Seahawks loss to the Patriots was the injuries the Seahawks suffered in the game. In game injuries are too often overlooked with regard to the outcomes of close games, in my view. I thought Neal's injury when the Pats were upset by the Giants the first time was critical.

Points: 0

#40 by dmstorm22 // Sep 01, 2016 - 6:47pm

The irony that two of the injuries came on both of Brady's interceptions (Lane and Bennett) is just beautifully awful.

Points: 0

#42 by theslothook // Sep 01, 2016 - 7:00pm

Avril got hurt too.

But honestly - I spent that game watching edleman and vareen constantly getting open for first downs. And they weren't in especially clever ways. They would literally fan out into a void between the linebackers and third corners. Vareen in particular basically ran flare outs from the backfield and kept getting third down conversions.

I suspect Seattle was content with forcing NE to march down the field than it was to change up their game plan mid game

Points: 0

#57 by Aaron Brooks G… // Sep 02, 2016 - 10:41am

SEA had run out of healthy DBs. Those problems are what happens when you're on your 4th-string DB.

Points: 0

#59 by dmstorm22 // Sep 02, 2016 - 10:50am

Ah, maybe it was Avril, not Bennett. Can't stand either team in that game enough to rewatch it so was going off of memory.

They also had injuries to Sherman & Thomas to where they were not 100%. That was a really injured Seattle team on defense.

I think Green Bay would have had a decent shot of winning the Super Bowl had they not absolutely collapsed against Seattle, but that's neither here nor there.

Points: 0

#15 by Aaron Brooks G… // Sep 01, 2016 - 8:37am

Passing was fine. Passing from shotgun was indefensible. He telegraphed it was a pass. You're at the 1. Don't snap the ball back to the 7.

Points: 0

#16 by Will Allen // Sep 01, 2016 - 10:07am

I disagree, because the play asks a Seattle receiver to defeat a jam at the line of scrimmage in which he he is physically outmatched by Browner. The point of play calling is to get the opponent in a physical matchup where he is at a decided disadvantage.

Points: 0

#18 by eagle97a // Sep 01, 2016 - 11:54am

Physical mismatch but the playcall was designed as a pick play with the receiver not needing a full block but just a chip to let the other receiver drive thru and make a play since if executed well he'll a have a small window to work with. It does show the preparation and coaching skill of NE but for me it doesn't make Pete a bad coach. I tend to agree with BBs' observation that Pete was being judged by people so much less qualified than him. PC does have the credentials of a top flight HC and they just had a bad outcome there considering for the most part he coached the game well and was ahead by 10 pts late in the 4th. Breaks of the game.

Points: 0

#19 by Will Allen // Sep 01, 2016 - 12:23pm

I didn't say Carroll was a bad coach. I said he was outcoached by Belichik on the Seahawks final drive, even though the Seahawks advanced to the goal line. This happens to be true. Belichik was better organized, and his decision-making was more rapid, in a situation where there was value to rapid decision-making. With regard to the int, the play called required a receiver to at least partially defeat a jam against an defender whose primary physical advantage in that setting was the ability to execute a jam. If your argument is that it is irrelevant to playcalling to ask your players to attack what the defender is best at doing, we just aren't going to agree. Now, sometimes, the defenders have so many physical advanatges over the offense, it doesn't much matter what you call, your guys are going to try to win a matchup in which they are at a significant physical disadvantage. You then get whipped 4 times out of 5, and them's the breaks. That wasn't the situation here.

Points: 0

#32 by eagle97a // Sep 01, 2016 - 5:31pm

I didn't say its irrelevant in playcalling to ask your players to attack what the defender is best at doing, of course that matters as well. What I'm trying to say is there has been a lot of second-guessing about the play from armchair head coaches. Yes the Patriots are the gold standard when it comes to preparation and coaching but in this case the Hawks were beaten by an almost fluke play after benefiting from another fluke play themselves. You can hear Pete saying the Pats were going goal line defense and he structured his playcall with that and the remaining time in mind while the Pats countered with 3 corners. There was a lot of things that needed to go right for that interception to happen and that is what I meant by them breaks. You can never prepare for all eventualities with a limited time to practice. Even the Pats learned this when they didn't practice and prepare well enough for David Tyree, BB was on record that they took Tyree for granted.

Points: 0

#58 by Aaron Brooks G… // Sep 02, 2016 - 10:43am

"I said he was outcoached by Belichik on the Seahawks final drive, even though the Seahawks advanced to the goal line."

No he wasn't. Belichick wanted to call a TO. The DC called the play and the jam.

Points: 0

#61 by Bright Blue Shorts // Sep 02, 2016 - 1:01pm

DC asked "Do you want a timeout?"

Belichick said "No, just play goal-line".

Points: 0

#66 by gomer_rs // Sep 05, 2016 - 11:29am

I think passing the ball was the right call, because it gave Seattle three plays to score a TD rather than two plays to score a TD. (There was only 24 seconds on the clock at the snap of the ball and Seattle only had 1 Timeout).

However, that they ran THAT play has always angered me. They should have run the safest play possible, and with RW as QB and Pete Carroll being an old West Coast original San Francisco Guy, why they heck they didn't roll RW out of the pocket to the wide side of the field with a receiver dragging at the goalline and a TE at the back line I will never know. You then get one low percentage play to add to two high percentage play, with little chance at a turnover.

I can't remember the name of the play, but it's the most famous West Coast Play in history, every West Coast team uses it, Seattle used a version of it 5-6 times a game that year, and "The Catch" to Clark was on this play.

I remember when they were the Sea-chickens.

Points: 0

#68 by Billy Everyteen // Sep 06, 2016 - 1:53pm

Sprint Right Option.

And I agree with you wholeheartedly. Possibly could also have used the good old Spider 2 Y Banana, which is similar.

And now I will go back to never ever thinking about that play.

Points: 0

#13 by panthersnbraves // Sep 01, 2016 - 7:21am

As I was reading through, and then saw the chart, I wondered about the 2003 "Cardiac Cats" and then you went on to explain how unusual their run was. I also was thinking about last year and (not glossing over Von Miller's incredible day) in the end, it just seemed that Denver had a huge advantage of "having been there" while the Panthers couldn't seem to get out of their own way some times. The huge punt return jumps out as the kind of thing that inexperienced teams do on the "big stage." Hopefully they get another chance at it.

Points: 0

#22 by pitt1980 // Sep 01, 2016 - 1:02pm

"Lastly, I wanted to draw attention to the cluster of weird results from the period of 1996 to 2003. Eight of the 10 surprise Super Bowl teams happened in this most unusual era of NFL history. These years may have given a false sense of parity gone wild in a new salary-capped NFL"


you're sort of hinting at it, but how much of that was due to GMs not understanding how to properly manage the cap, and getting themselves into cap trouble?

49ers and Cowboys were both dominate teams at the start of that stretch, they might have both spent themselves out of contention by the end of it

did teams get better at that since that period?

interesting test of this might be to take Chase Stuarts AV, and see how much of it was available in FA for that stretch, especially in contrast to now

at teams better at keeping talented guys off the market now?


" But there are still examples of teams that needed to overcome and learn from some heart-breaking defeats before realizing what it took to finish the job. The Ravens (2008-2012) and Broncos (2011-2015) just recently completed five-year journeys to rings. Were the 2012 Ravens and 2015 Broncos the strongest teams among those runs? No -- in fact, you could argue that the 2012 Ravens were John Harbaugh's weakest playoff team, and that the 2015 Broncos were the weakest Denver club since Manning arrived in 2012. But those were the years where things all worked out for both teams."

idk about learning from defeat

an interesting question is whether its best just to approach the playoffs as totally random

so what you want to do, it expand your window by getting into it as many years as possible,

in contrast to shortening your window by trying to sell out to win in any particular year

example: I can't find the link to this, but I have a memory of Bill Simmons complaining about Belicheck always trading back in the draft, and that he was wasting Tom Brady's prime doing that, (while it seems sort of obvious to as a rule not doubt Belicheck, especially in favor of Simmons), it seems to fit the method of buy more tickets to the lottery that is the playoffs, don't sell out the future for any particular year

Points: 0

#67 by LionInAZ // Sep 05, 2016 - 10:09pm

The difference between Bill Simmons and Bill Belichick is that Belichick has kept the Patriots in the playoff hunt for 15 years, while Simmons gets to hang out with idiot celebrities like Ben Affleck. Big difference.

Meanwhile, the Packers pursue a similar strategy, and guess what? They're in the playoff picture year after year.
Just not as successfully. But then, they're not on Bill Simmons' radar because anything that happens outside of the Northeast or LA is irrelevant to him.

Points: 0

Save 10%
& Support Scott
Support Football Outsiders' independent media and . Use promo code WRITERS to save 10% on any FO+ membership and give half the cost of your membership to tip Scott.