Writers of Pro Football Prospectus 2008

Most Recent FO Features

HarrisChr11.jpg

» Film Room: Chris Harris

Is Harris one of the league's top cover corners, or a product of the system in which he plays? Cian Fahey says the answer lies somewhere in the middle.

03 Jul 2014

Let Me Check My Schedule

by Kenneth Arthur

"Who is the best among them, given an equal playing field?"

This is one of the basic questions that "sports" tries to answer each year when determining a champion, but it’s easier to do in some competitions than it is in others. Individual sports like golf and cycling simply pit everyone against each other on the same field and see who has a better tournament, but football simply can’t have every team play every other team once. You could try putting every team on the same field at one time, but I’m pretty sure that’s what the XFL tried and that didn’t last very long.

Inevitably the schedules will vary so much that some teams will have it easier than others, and most likely those that have it hardest will have a more difficult time winning the Super Bowl that season. At its base level, a team's strength of schedule is mostly random. With the exception of the NFL's method of matching a team against their "division equal" for two games a year, the other 14 games are based entirely upon factors that have nothing to do with how good or bad your opponent is.

You'd prefer that no one team had a decided advantage over another based on something that's entirely outside of their control, but that’s just never going to happen. I’d prefer to have abs like Channing Tatum, but I’d be cast in a movie about Magic Shell ice cream topping before I’d get a role in Magic Mike.

Last season's biggest surprise, the Kansas City Chiefs, faced the easiest schedule in the NFL by DVOA. The AFC's representative in the Super Bowl, the Denver Broncos, had the next easiest. Chip Kelly's quick turnaround of the Philadelphia Eagles was certainly somewhat aided by the 29th-ranked schedule in the league. Meanwhile, two of the biggest disappointments in the NFL last year -- the Tampa Bay Buccaneers and the Atlanta Falcons -- were first and second in SOS DVOA.

Only a fool would blame a 4-12 record on strength of schedule alone. (The Cardinals had the fourth-hardest schedule and went 10-6, while the Saints were fifth and the Panthers were sixth.) But is it enough of a difference to change a "playoff team" into a January afterthought? Is it better to take the easy road or get put through the ringer so that you're "battle tested" for the playoffs?

After tracking the last 25 seasons by strength of schedule, here are some of the interesting facts that we can use to start to understand more about what your DVOA SOS means about your chances to W-I-N.

First or worst?

Every year, some team will inevitably have to play the most difficult schedule and another will get to face the easiest. One could imagine that the difference between facing the hardest schedule and the easiest schedule is significant, but when examining the evidence, it would appear that the difference is even more monumental than you may think.

The teams that have had the misfortune of facing the hardest schedule every season since 1989 have averaged 5.84 wins per year and a 20.84 seasonal ranking in DVOA. Only one team, the 1997 Steelers squad that went 11-5, has managed double-digit wins with the league’s hardest schedule. The difference in facing the league’s easiest schedule is slightly less than double the same total of wins; the 25 teams that had the league's easiest schedule have averaged 10.68 wins.

In theory, the team with the hardest schedule should have an average DVOA ranking of 16.5, but what we’ve seen through 25 years is a mark more than four spots lower than that. Whether or not that’s random, since a sample size of 25 teams is still relatively small, is unknown, but it is interesting nonetheless.

Easiest schedules, by year:
Team Sched Rank Year Wins DVOA Finish
WAS -5.5% 28 1989 10 10
SF -6.4% 28 1990 14 7
BUF -14.3% 28 1991 13 4
SD -9.0% 28 1992 11 10
GB -5.4% 28 1993 9 9
PIT -3.8% 28 1994 12 2
NYJ -7.0% 30 1995 3 28
CAR -6.4% 30 1996 12 5
SF -6.7% 30 1997 13 4
ARI -10.8% 30 1998 9 25
STL* -15.7% 31 1999 13 1
PHI -8.7% 31 2000 11 8
PIT -6.2% 31 2001 13 7
PHI -6.9% 32 2002 12 5
SEA -7.2% 32 2003 10 8
SEA -9.5% 32 2004 9 15
SEA -10.3% 32 2005 13 3
SEA -8.4% 32 2006 9 24
SEA -10.6% 32 2007 10 9
NYJ -6.9% 32 2008 9 17
ARI -9.7% 32 2009 10 12
STL -11.0% 32 2010 7 27
SF -5.4% 32 2011 13 6
IND -7.4% 32 2012 11 25
KC -7.0% 32 2013 11 6
Easiest schedules, all time by DVOA:
Team Sched Rank Year Wins DVOA Finish
STL* -15.7% 31 1999 13 1
BUF -14.3% 28 1991 13 4
STL -11.0% 32 2010 7 27
ARI -10.8% 30 1998 9 25
MIA -10.6% 27 1991 8 18
SEA -10.6% 32 2007 10 9
SEA -10.3% 32 2005 13 3
JAC -10.2% 30 1999 14 2
NYJ -10.0% 26 1991 8 20
DAL -10.0% 29 1998 10 9
ARI -9.7% 32 2009 10 12
SEA -9.5% 32 2004 9 15
GB -9.2% 31 2009 11 2
SD -9.0% 28 1992 11 10
ARI -9.0% 31 2007 8 23
ARI -9.0% 31 2010 5 32
CAR -8.9% 29 1999 8 19
TB -8.9% 30 2007 9 8
PHI -8.7% 31 2000 11 8
CAR -8.7% 31 2005 11 10
TB -8.6% 30 2005 11 14
SEA -8.4% 32 2006 9 24
TB -8.3% 31 2004 5 16
SEA -8.3% 30 2010 7 30
MIN -8.2% 25 1991 8 16
CHI -8.2% 31 2006 13 5
DEN -7.9% 30 2000 11 9
STL -7.9% 30 2006 8 25
KC -7.9% 29 2010 10 17
SF -7.8% 28 2010 6 24

The easier schedule has clearly led to more wins for those teams, but most of those regular season wins haven’t translated into postseason success. Only one team in the last 25 years had the easiest schedule in the league and went on to win the Super Bowl: The 1999 Rams. They were also the only team among those 25 to finish first in DVOA. The easy schedule so undermines the wins that it can lead to a lot of those "disappointing" teams that may not have been as good as they looked.

The 1990 49ers had the easiest schedule and won 14 games but were only seventh in DVOA. They lost to the Giants in the conference championship, a team that had 13 wins on the 10th-hardest schedule.

Last year, the Chiefs had the easiest schedule and blew it against the Colts in the playoffs. The Broncos had the second-easiest and were dismantled in the Super Bowl like they were an atomic bomb and the Seahawks were U2.

Meanwhile, five of the easiest schedules of all-time by DVOA came in 2010. That season the Packers slid under the radar with a team that was fourth in DVOA on the 15th-hardest schedule, then won the Super Bowl as the No. 6 seed in the NFC.

No team has benefited from an easy schedule more than the Seahawks of the mid-ought’s. Seattle had the easiest schedule in the NFL for five years running while playing in the abysmal NFC West. They parlayed that opportunity into five straight playoff appearances and a Super Bowl trip in 2005. That season they had the sixth-easiest mark in DVOA history at a -10.3% strength of schedule. But it doesn’t always work out that way for everybody.

In classic Jets fashion, New York had the easiest schedule in 1995 and still went 3-13.

What of the toughest schedules?

Hardest schedules, by year:
Team Sched Rank Year Wins DVOA Finish
SEA 6.5% 1 1989 7 25
NE 7.1% 1 1990 1 28
ARI 12.2% 1 1991 4 25
NYJ 7.9% 1 1992 4 19
TB 4.4% 1 1993 5 27
DET 3.6% 1 1994 9 12
ARI 5.6% 1 1995 4 30
SEA 5.9% 1 1996 7 25
PIT 5.4% 1 1997 11 5
NE 9.6% 1 1998 9 11
DEN 9.2% 1 1999 6 12
NYJ 11.0% 1 2000 9 12
KC 4.5% 1 2001 6 16
DEN 7.4% 1 2002 9 8
HOU 8.4% 1 2003 5 30
CLE 13.7% 1 2004 4 28
SD 10.7% 1 2005 9 6
TEN 9.7% 1 2006 8 22
BUF 7.8% 1 2007 7 21
CLE 9.3% 1 2008 4 26
TB 9.1% 1 2009 3 27
BUF 11.4% 1 2010 4 29
STL 5.4% 1 2011 2 32
ARI 10.8% 1 2012 5 26
TB 9.5% 1 2013 4 19
Hardest schedules, all time:
Team Sched Rank Year Wins DVOA Finish
CLE 13.7% 1 2004 4 28
ARI 12.2% 1 1991 4 25
BUF 11.4% 1 2010 4 29
NYJ 11.0% 1 2000 9 12
ARI 10.8% 1 2012 5 26
BAL 10.7% 2 2004 9 8
SD 10.7% 1 2005 9 6
CIN 10.0% 2 2010 4 19
OAK 9.7% 3 2004 5 24
TEN 9.7% 1 2006 8 22
NE 9.6% 1 1998 9 11
STL 9.6% 2 2012 7 15
CIN 9.5% 4 2004 8 11
TB 9.5% 1 2013 4 19
MIA 9.3% 5 2004 4 22
CLE 9.3% 1 2008 4 26
DEN 9.2% 1 1999 6 12
MIA 9.2% 3 2010 7 15
TB 9.1% 1 2009 3 27
STL 8.9% 2 1991 3 22
CIN 8.8% 2 2008 4 27
CLE 8.5% 4 2010 5 20
HOU 8.4% 1 2003 5 30
OAK 8.4% 2 2005 4 20
IND 8.2% 2 1998 3 26
NYJ 7.9% 1 1992 4 19
BUF 7.8% 1 2007 7 21
DEN 7.4% 1 2002 9 8
ATL 7.3% 2 2013 4 25
ATL 7.2% 2 1992 6 21

Of the 30 most-difficult schedules measured by DVOA, not a single one of those teams won more than nine games. Only the 1998 Patriots, 2000 Jets, 2004 Ravens, 2005 Chargers, and 2007 Broncos managed to get over .500. Of those teams, the most interesting might be Pete Carroll’s New England Patriots that made the playoffs as a wild card and featured talent like Ty Law, Lawyer Milloy, Drew Bledsoe, Tedy Bruschi, and Willie McGinest, all of whom would go on to win a Super Bowl three years later. Without Carroll, of course.

The path to the Super Bowl might be right down the middle

I think every study should be looked at from at least two sides, and in this case we can look at what schedules tell us about success and what success tells us about the schedule. So how hard or easy were the slate of games for the last 25 Super Bowl winners?

By virtue of facing the 16th-toughest schedule by DVOA last year, Seattle was middle-of-the-pack in those rankings, and their -0.4% mark puts them at slightly easier-than-average. The 2012 Ravens also finished 16th in SOS, but some Super Bowl winners have faced much tougher and easier schedules than these two.

Overall, the 25 Super Bowl winners since 1989 average a -1.08% strength of schedule by DVOA, and rank 17.68 in their respective seasons.

Super Bowl winners:
Team Sched Rank Year Wins DVOA Finish
STL -15.7% 31 1999 13 1
DEN -6.8% 28 1998 14 1
BAL -4.8% 24 2000 12 3
NE -4.2% 29 2001 11 11
DAL -3.6% 22 1992 13 1
TB -3.2% 26 2002 12 1
DAL -3.1% 23 1993 12 1
DEN -2.8% 24 1997 12 2
NO -2.5% 23 2009 13 6
SF -2.5% 25 1994 13 3
SF -2.0% 22 1989 14 1
GB -1.8% 23 1996 13 1
PIT -1.3% 17 2005 11 4
BAL -1.0% 16 2012 10 8
SEA -0.4% 16 2013 13 1
WAS 0.1% 17 1991 14 1
NYG 1.4% 12 2007 10 14
NYG 1.6% 10 1990 13 1
GB 1.9% 15 2010 10 4
NE 2.2% 10 2003 14 4
DAL 2.7% 7 1995 12 2
IND 3.1% 9 2006 12 7
NYG 3.6% 4 2011 9 12
NE 5.6% 6 2004 14 2
PIT 6.3% 3 2008 12 4
  -1.088% 17.68   12.24 3.84

The 2008 Steelers had the toughest schedule since 1989 for a team that went on to win the Super Bowl. They had the third-hardest SOS. Though Pittsburgh got to play Cleveland and Cincinnati twice, they only faced one other team with a losing record that whole season: the 5-11 Jaguars. The Steelers had been put through the ringer that year, but does being "battle tested" really matter at all? Or would you rather get the easiest road to the Super Bowl as possible?

I think it’s always better to go easy on yourself in the regular season, because the playoffs are hard enough. Don’t believe me? Look at the team that the Steelers almost lost to that year.

The Cardinals went 9-7 on the 27th-hardest schedule, in probably the worst division in football. A series of fortunate events helped Arizona get two home playoff games against both NFC wild card teams, while avoiding the No. 1 seed Giants. The Cardinals were only ranked 21st in DVOA that season, but came very close to being the Super Bowl champions, and they were helped immensely by their easy schedule and weak division. Nobody would really remember or care that Arizona wasn’t all that great that year if they had won the Super Bowl, just like most people forget that the Giants weren’t that great in their two most recent Super Bowl championships either.

Or that New York’s 2007 and 2011 season are such anomalies in nearly every category you can imagine that I’d much rather just wipe them clean off the record books.

In fact, the NFL’s most recent "dynasty" may have never become a dynasty if not for an easy schedule. The 2001 Patriots had the third-easiest schedule in the league, were only 11th in DVOA, won the AFC East over the Dolphins by virtue of a tiebreaker, and subsequently got the No. 2 seed thanks to a weak conference. They may have never upset the Rams in Super Bowl XXXVI if they had the third-hardest schedule that year instead.

The 2001 Pats got through the regular season on an easy schedule, which in turn helped build them into a dynasty, but it wasn't the easiest schedule ever had by a Super Bowl winner. Not even close.

'99 problems but a schedule ain't one

The 1999 St. Louis Rams stole America's heart that season, the story of a "grocery baggin' son of a gun" and his band of scrappy rascals that turned around a 4-12 season. "A Really Good Show On Astroturf" as they were called back then, I think, the Rams went 13-3 that season and averaged +17.3 points per game (an NFL record until the Pats broke it in 2007). But I think there are two aspects of that team that most people forget or were not aware of:

  • St. Louis was ranked higher on DVOA for defense (third) than they were on offense (fourth)
  • The Rams had the easiest schedule not just for that season, but for any season by any team since 1989.

Out of 765 team seasons, St. Louis’ SOS mark of -15.7% is lower than any team you can name since 1989. They played 13 games against teams ranked 19th or lower in DVOA and six games against the four worst teams in the NFL.

They went 1-2 against teams ranked higher than 19th in DVOA, with their only such win coming in Week 1 against the Ravens (11th in DVOA). They lost to the Titans (fifth) in Week 8, the best team that they played all year. Still, the easy schedule helped the Rams get to the Super Bowl, where they’d beat Tennessee by a nose.

Which goes back to the same point made earlier about the 2008 Cardinals: the most important thing you can do to win the Super Bowl is to get to the Super Bowl. However you can.

Had the Rams played even an average schedule, would they have won 13 games? Maybe. Were they the best team in the NFL that season? Probably. But we also know that the best teams don’t always win the championship. In fact, the 2013 Seahawks are the first team to rank first in DVOA in the same season that they won the Super Bowl since the 2002 Buccaneers. Had the 1999 Rams had a difficult schedule instead of an unbelievably easy one, they might not have had the edge they needed over the Bucs (11-5, eighth in DVOA) to get homefield advantage throughout the NFC playoffs. And considering that Tampa actually held a one-point lead over St. Louis late in the NFC title game that year, homefield could have played a crucial difference in the outcome.

It’s almost certainly better to have the kind of schedule that allows you to get every advantage you need headed into the playoffs, but is there anything we can do to use that information to possibly predict the successful teams of next season?

Last season is so 2013

Given the way that the scheduling system is setup, the difficulty of the NFC West as the most talented division, and the NFC as the stronger conference, the Seahawks could have a difficult time repeating as Super Bowl champions. This is possibly a factor that has kept teams from repeating championships in the 2000s like we had been accustomed to seeing in previous decades.

Seattle is set to play nine of their 16 games against teams that were in the top 10 in DVOA last year.

However, that only matters if you use 2013 DVOA data for 2014 teams, and that would be a waste of time. If you just look at last year's rankings, you'll see that the Chiefs jumped from 32 to 6, the Eagles from 28 to 8, the Cardinals from 26 to 10, and the Chargers from 22 to 12.

Meanwhile, the Texans dropped from 11 to 30, the Redskins from 9 to 29, the Giants from 7 to 27, and the Falcons from 10 to 25.

And those falls are a lot more "predictable" in hindsight than they were headed into the 2013 season. It wouldn't be fair to say we have no idea which teams will improve or fall flat on their faces next season, but certainly some things are going to happen that are unpredictable or won't make any sense. The Packers fell from 5 to 20 last year due to the injury to Aaron Rodgers, and will likely climb back into the top 10 if he plays a full season in 2014. The Texans placed almost their entire team on IR last year (the relevant players at least) and will get many of those guys back. Plus they added Jadeveon Clowney.

If you could accurately predict how the SOS will look for each team next season, you could also come up with some surprising predictions when looking for surprises in 2014. Instead, the best we can do is a ballpark figure that is likely going to be just as accurate as a ballpark guess on how teams will finish in the standings and in DVOA, which is always going to have some major misses.

Strength of schedule clearly matters. It has helped build both dynasties and dams. We’ll never know why the Patriots got that "random" fortune in the early 2000's that the Chargers did not in the mid-2000's, but I think it’s safe to assume that somehow the answer is "Norv Turner." It also matters that you're actually good enough to win the Super Bowl, and as a Seahawks fan myself, I can personally attest to how little it helps to win the division five years in a row if you’re simply the best of the worst. Like the New England dynasty, you still have to be really, really good. Not having a really difficult schedule just helps you get to where you need to be healthier, and quicker.

And then again, sometimes it’s just good enough to be the New York Giants; no stats or studies in the world can seem to explain how Eli Manning has two rings.

Kenneth Arthur is a writer for Field Gulls, SB Nation, and Fake Teams, and once had a blog about the movie The Room.

Posted by: Guest on 03 Jul 2014

25 comments, Last at 01 Sep 2014, 4:52pm by kicorse

Comments

1
by theslothook :: Thu, 07/03/2014 - 2:55pm

I think I've seen enough recent evidence to suggest that any team can win once it gets into the tourney. If that means an easy schedule then so be it. I don't really believe there's a mystical quality that the winning sb team had that others didn't. It's easy to say Seattle was the best team(they were), but if we were able to simulate these same playoffs again from the beginning, I'm pretty sure Seattle's probability wouldn't go above 30 percent(even if it is the highest of the 12 teams).

That's actually a better way to think about this. The eagles had a very successful run in the 2000s, making the playoffs 8 out of 10 times and winning their division 5 in five of them. By contrast, the giants made the playoffs 6 out of 10 times but won their division only 3 times. The steelers also made the playoffs 6 out of 10 times and won their division 5 times.

The eagles had the best chance of these three teams to win at least one superbowl. Instead, it was the latter two that ended up with multiple titles. The point is, it was just how that random draw turned out. I like to think in another universe, Peyton Manning ends up with a jordanesque 6 for 6 and Bellichick gets the Mart Schottenheimer label.

7
by t.d. :: Fri, 07/04/2014 - 5:45am

I don't quite buy the 'just get into the tournament and you've got a good shot at it' argument, in that the teams that have shocked the analytics guys have all had quarterbacks that were more popular with scouts than with statisticians (Flacco and Manning, in particular). Much like Maddux and Glavine were better regular season pitchers than Randy Johnson or Schilling, but it didn't carry over to postseason, I do think there's something to the laser-rocket arm versus the crafty efficiency guys that is hard to quantify, but is real (maybe it's the crappy weather in the postseason, or the better typical opponents, or something else, hard to say)

8
by theslothook :: Fri, 07/04/2014 - 1:05pm

I think the narrative of both players was framed after they won the sb. Eli in particular was still viewed as a bust right up until that 07 sb.

9
by Karl Cuba :: Fri, 07/04/2014 - 4:05pm

For me it's more of a case of you have a chance if you can get to the playoffs in good form and healthy.

14
by Duff Soviet Union :: Sat, 07/05/2014 - 9:39pm

Except the 2007 Giants weren't in good form. They went 4-4 in their last 8 and none of those wins were especially convincing. There was a hindsight claim that narrowly losing to the 16-0 Patriots in week 17 showed how good they were, but again that's complete hindsight.

The 2011 Giants won 3 out of 4 going in, including a week 17 elimination game against the Cowboys, but they'd lost 4 straight before that. I don't think they were in any better form than about 10 other teams going into the playoffs. They were getting healthier though.

15
by Karl Cuba :: Sat, 07/05/2014 - 10:13pm

No but I think the 2007 Giants got healthy, returning a bunch of dB's they needed in the playoffs. As you've said the other team of Giant rat-bastards also got much healthier.

18
by stevo :: Tue, 07/08/2014 - 2:06am

You should add the 70's and mid to late 90's Vikings to the Eagles comparison. All those teams had as good a chance anyone in getting ring based on the number of opportunities and division titles. But alas, they never pulled all the way through.

I do agree that the chance of winning in the playoffs decreases just by the very nature that teams are playing the best of the best. Even more chance is added (versus basketball, for example) since it is one and done and there are so many players on the field at one time that one fluke play could end it for any team (2012 Broncos anyone?).

However, I would propose that the best team does most often win. Best team however, cannot be based solely on regular season success (as the article shows, good teams can get a raw schedule deal and miss the playoffs or bad teams can be inflated by an easy SOS). Add to that the season long ebb and flow of strategies, injuries and team dynamics, and "best" can change multiple times in a season. Quantifying the best team come playoff time has to be based on a mixture of stats like DVOA, trend analysis and the context of injured players. Last year, for example, Seattle and San Francisco were the most obvious "best" teams in the league come playoff time if you combine those concepts together and the real Superbowl happened in the NFC championship game.

Oh yeah, and it helps to have a coach who can make the right tough calls (thank you Dennis Green...)

19
by Jerry :: Tue, 07/08/2014 - 2:26am

For those of you who don't remember, Doug Drinen simulated 10,000 seasons on the old pro-football-reference blog, and found the best team won about 24% of the time:

http://www.pro-football-reference.com/blog/?p=57

Of course, a championship is a championship, regardless of how good the champ actually is. And, as others have suggested, we tend to retrofit the narrative to explain things that are often just luck.

21
by stevo :: Wed, 07/09/2014 - 12:46am

So this would give evidence that the best team does win most often. Even though 24% doesn't seem that high it's the highest percentage among all the other teams. And if all teams that made the playoffs had an equal chance, from a field of 12, the best team would have only won 8% of the time.

20
by Perfundle :: Tue, 07/08/2014 - 4:42am

I think what you're actually proposing is a tautology, i.e. the best team does most often win because the teams that win are the best teams.

2
by Scott de B. :: Thu, 07/03/2014 - 3:09pm

What I'd like to see is a discussion of the variance in SOS. From the tables above, it seems there can be a major difference in how hard the 'hardest' schedule is from year to year, and equally a difference in how easy the easiest schedule is.

3
by Perfundle :: Thu, 07/03/2014 - 4:00pm

DVOA doesn't extend beyond 1989, but Pro Football Reference does, and they have their own SOS that matches up pretty well with FO's SOS. Based on that, St. Louis' SOS was the second-lowest since the merger. The only team to have a lower SOS was the 1970 Miami Dolphins, in the first year of the merger.

4
by dryheat :: Thu, 07/03/2014 - 4:21pm

I always thought one was put through the "wringer". Interesting bit of research here, thanks.

Mmmmmm..............Magic Shell..........

6
by dbostedo :: Thu, 07/03/2014 - 9:51pm

"I always thought one was put through the 'wringer' "

That's correct. I think it refers to wringer washing machines. We had one when I was little, and I remember my arm going through the wringer, literally.

5
by Bjorn Nittmo :: Thu, 07/03/2014 - 5:51pm

It's not a coincidence that the worst teams each year have, on average, more difficult schedules -- they don't get to play themselves, while their superior opponents do (sometimes twice). So there's some bias in your finding that the teams facing the hardest schedule have many fewer wins than ones facing the easiest. Even for 2 teams within the same division, who share 12 common opponents each season, those 2 games played against each other can probably swing SOS for each team enough for the average SOS to overstate the true difference in SOS.

10
by Perfundle :: Fri, 07/04/2014 - 7:22pm

You would think that this is true, but it's not. I compared DVOA rating against SOS for every team in the last 12 years, and the R-squared value is 0.0004. The average SOS for teams with a DVOA rating of at least 30% is actually positive, and the average SOS for teams with a DVOA rating of at most -30% is actually negative. Remember that DVOA is adjusted for SOS, so that an average team playing a very hard schedule still would have a DVOA of 0% despite their expected poor record.

17
by hoegher :: Mon, 07/07/2014 - 1:27pm

Yeah, but this isn't looking across all teams, it's just looking at the single most difficult/easy schedule.

The most difficult schedule in any given year would have to have a couple strong teams in it, which logically cannot be the team in question. Vice versa for the easy schedule.

11
by Gladiator of th... :: Sat, 07/05/2014 - 1:50pm

I think you alluded to this towards the end of the article, but I'd like to see a chart that illustrates the difference between the perceived SOS at the beginning of a season, and the actual SOS in retrospect following the season. (I guess that might as well simply be a comparison of the most recent season and its prior season... but it's still interesting and can sometimes provide some sort of perspective when your team got trampled all season long and you realize they had one of the toughest schedules rather than the cakewalk it appeared to be in the preseason.)

12
by BigDerf :: Sat, 07/05/2014 - 6:26pm

What is with the two basically random sentences about the Giants? Felt you had to mention them in passing and pot shot Eli while totally ignoring them for the rest of the study/article?

Also, Just because things are inexplicable with stats you are currently using doesn't make them inexplicable all together. Random poking of Norv Turner was random when it could have easily been used as a piece to pull together the "eye-test" with the stats or a point in the article where you reference that SoS doesn't tell the whole story. Instead it's basically just "HAHA NORV"

13
by Jetspete :: Sat, 07/05/2014 - 9:19pm

Good article, but i think you run into problems trying to equate regular season SOS with postseason success. In reality, I imagine postseason SOS correlates better with postseason success than reg season SOS. Case in point, the 2008 steelers you reference above played 8-8 SD in the second round followed by 6th seeded Baltimore and 9-7 Arizona. The Steelers got lucky that Baltimore won in Round 2, or else they were going to have to travel to Tennessee, a team that dominated them in week 16.

I'd also be interested in the variance in SOS and if it was different pre/post 02 when only 2/16 games were determined from prior year schedule.

16
by andrewmilne :: Sun, 07/06/2014 - 9:58pm

I think there are a lot more questions to be asked here, about causative factors, for instance. I emailed back and forth with Mike Tanier on this a few years ago, but he never had a chance to research it (or never cared to - he does have more important thing to do than the things that I care about!)

The thing that interests me most about strength of schedule is "What does it do to a team?' So does playing harder teams make a team stronger (as coaches like to say) or does it just discourage them and make them worse? Does playing easier teams make them soft, or does it give them confidence (as coaches like to say)?

It seems to me that there is a really interesting study to be done where you look at difference in DVOA between the teams that play (perhaps for sequences of opponents - the last five, or something) and see how it correlates with future DVOA.

22
by Mr Luan :: Wed, 07/23/2014 - 9:25pm

I think you alluded to this towards the end of the article, but I'd like to see a chart that illustrates the difference between the perceived SOS at the beginning of a season, and the actual SOS in retrospect following the season. (I guess that might as well simply be a comparison of the most recent season and its prior season...
http://www.jvzooreview.us/2014/07/wp-traffic-tagger-review-bonus.html

23
by RickD :: Fri, 08/15/2014 - 6:21pm

"In theory, the team with the hardest schedule should have an average DVOA ranking of 16.5, "

No, in theory the team with the hardest schedule should have a below-average DVOA.

Look at this division-by-division. Invariably, the team with the toughest in-division schedule is the weakest team, while the team with the easiest in-division schedule is the strongest team.

Next, each team in a division plays identical schedules against one division from its own conference, and one division from the other conference.

And that leaves the only part of the schedule that favors the weaker teams: the last-place team plays two other last-place teams from its own conference, while the first-place team plays two first-place teams from its own conference.

So, for example, while the Pats get two games against the Bills (compared to the Bills facing the Pats twice), the Pats also have to face Cincy and Indy, while the Bills get Cleveland and Houston. You might think these things would balance out, but the issue here is that "last place" and "weakest team" don't necessarily coincide. Is Houston really an easier mark than Jacksonville? DVOA doesn't think so.

Most of what determines "schedule strength" is the aggregate strength of each division. That's why last season the teams from the AFC West had relatively easy schedules, playing the NFC East and the AFC South, while the NFC East had relatively hard schedules, playing the AFC West and the NFC North. Washington in particular was nailed, playing a 1st place schedule with a QB recovering from injury. So was Houston, minus the QB injury.

It shouldn't be surprising that the average DVOA rating of the team with the hardest schedule is below 16.5. Let's suppose DVOA ratings and schedules were randomly assigned. And to keep the math simple, lets assume a uniform distribution to DVOA ratings. Then the average team on the average schedule for the highest rated DVOA team would be rated 17th in DVOA (the average of 2 through 32), which the average team on the the average schedule of the weakest team by DVOA would be 16th in DVOA (the average of 1 through 31). And the average strength of the kth best team will be (33*16 - k)/31. [This number is found by summing 1 through 32 and subtracting k, then dividing by 31.] So these teams would be lining up, in terms of SOS, with the best team having the easiest schedule down to the worst team having the hardest schedule.

The unrealistic assumption here is that DVOA's are uniformly distributed, but you can see from how the argument develops that using the actual distribution of DVOA's shouldn't change the expected ordinal result. We should expect the hardest schedules to be showing up on weakest teams, simply because they cannot play themselves. I'm not surprised that the observed value is more than 4 spots lower than 16.5, I'd have thought it would be even lower than that. Presumably the reason it isn't is because of the churning of DVOA values, and also because said values are better described as a normal distribution than by a uniform distribution.

25
by kicorse :: Mon, 09/01/2014 - 4:52pm

That's absolutely right. I was going to comment something similar, but instead I'll summarise the logic:

1) The team with the easiest full schedule, by DVOA, in the league will probably have the easiest intra-divisional schedule, by DVOA, in their division (and if not, they'll be close). This is especially true since 2002.

2) The team with the easiest intra-divisional schedule, by DVOA, in their division, will have the best DVOA in their division.

On the team-DVOA point:

3) The team with the best DVOA in their division will probably have a good DVOA.

On the number of wins point:

4) The team with the best DVOA in their division will probably win, or come close to winning, their division.

5) A team that wins, or comes close to winning, their division, will probably have a good record.

(Obviously, you can reverse all of the above for hardest schedule.) There are three "probably"s there, which explains things like the 3-13 Jets, but we shouldn't be surprised by these findings.

24
by ThatemW :: Tue, 08/26/2014 - 5:02pm

Would a team with an easier schedule, and performs better than they should have, actually hurt themselves come draft?
Assuming that the easier schedule will give you a record better than you actually are?
The interesting thing would be to see a team with a series of easier or harder schedules (thus worse or better picks), perform on the opposite schedule. Would that performance be a suprise all of a sudden?