Comparing the final Football Outsiders F/+ projections to the AP poll and conventional wisdom.
02 Dec 2005
by Aaron Schatz
(Note: This article originally appeared on FOXSports.com during the 2005 season. Since it has disappeared from the FOX archives, we've reproduced it here on Football Outsiders. The appendix and tables appeared on FO only.)
I've received a lot of e-mail since Football Outsiders took over the FOXSports.com power rankings this year. Many of the e-mails I get read a little something like this:
"Team X blows out a losing team and you move them up your power rankings. But when Team Y pulls off a clutch victory against a winning team, you don't move them up. A close win against a good team is a lot more important."
There's only one problem with this statement. It isn't true.
People want to believe that the teams that can win the close ones are championship teams. But as counter-intuitive as it sounds, championship teams are generally defined by their ability to easily win games over inferior teams.
Football games are often decided by just one or two plays -- a missed field goal, a fumble that bounces one way instead of the other, a fourth down where officials spot the ball two inches from the first down marker. One dropped pass short-circuits a last-minute comeback. A cornerback smothers his receiver all day, only to get beat once and give up the winning touchdown.
The team that comes out with the victory in a tight game is one step closer to the postseason. But has that team really proven that it is better than its opponent? There are many times where two teams are evenly matched, and if they played again the next week the result might just as well go the other way.
When a team blows out its opponent, however, one unlucky bounce or missed kick isn't going to change the result. A lopsided win provides pretty good proof that the winner is a better team than the loser. That's why the teams that meet on Super Bowl Sunday are usually the teams that won a lot of games by big margins during the regular season.
Don't believe me? Let's take a look at how Super Bowl teams have built their resumes over the past decade.
I went through every regular season game from 1995-2004 to see whether teams that built their records on close wins over good teams did better in the postseason than teams that padded their records with big wins over bad teams. I created the following two categories:
STOMP: A win by at least 14 points over a team that will finish the year below .500.
GUT: A win by 1-8 points over a team that will finish the year above .500.
If the mark of a great team is the ability to gut out close wins over quality opponents, the teams with more GUTS should make the Super Bowl while teams with more STOMPS should get knocked off by superior competition in the playoffs. But that's not the case:
Some of you at this point are probably thinking, "Well, isn't the best win a big victory over another good team?" Others are thinking, "Well, some great teams don't whip up on bad teams because they do just enough to win."
To test these ideas, I created two additional categories:
DOMINATION: A win by at least 14 points over a team that will finish the year above .500.
SKATE: A win by 1-8 points over a team that will finish the year below .500.
(I should note that we are still leaving out all wins over .500 teams and all wins of 9-13 points.)
Look at all four categories, and it turns out that teams that go far in the playoffs don't generally have a habit of narrowly beating their bad opponents. In fact, many times a team with more big wins over bad teams will win the championship over a team with more big wins over good teams.
When we compare all four categories, it becomes clear that winning blowouts is a far better indicator of championship quality than winning close games. I looked at all 30 Super Bowls and conference championship games over the past decade to see which team had the higher total in each of these four categories. Here are the records, which don't necessarily add to 30 games because of matchups where the teams were equal in a certain type of win:
|Category||Super Bowl||Super Bowl and Conf. Champs|
|More STOMPS (big wins vs. bad teams)||7-2||21-7|
|More DOMINATIONS (big wins vs. good teams)||5-4||19-6|
|More GUTS (close wins vs. good teams)||4-6||9-15|
|More SKATES (close wins vs. bad teams)||2-6||7-18|
So far I've been looking at total wins, but of course to stomp on a bad team you have to play some bad teams, and to dominate a good team you have to play some good teams. How about looking at both wins and losses against good teams?
Many fans believe that the best way to judge a team is its record against the other good teams on the schedule, and that the team with the best record against quality opponents is the team most likely to win the Super Bowl. This is a particular obsession of the New England-based website Cold Hard Football Facts, which keeps track of what it calls quality standings. As the site points out, the New England Patriots had the NFL's best record against winning teams in both 2003 and 2004, and won the Super Bowl both years.
But two seasons are not really enough to judge the accuracy of any statistic, so I went back a decade to test if Quality Winning Percentage truly is the best indicator of how teams will fare in the postseason. Unlike the measures of GUTS and DOMINATIONS above, this time I counted every game against a team above .500 no matter the margin, and counted ties as half a win just like the official NFL standings.
It turns out that in the past ten years, only three teams that led the league in Quality Winning Percentage have emerged as Super Bowl champions. It just so happens that two of those teams were the 2003 and 2004 New England Patriots. (The other was the 1995 Dallas Cowboys.)
Look at the record when two teams have met in a Super Bowl or conference championship game over the past decade, and one has a better Quality Winning Percentage than the other. It's actually slightly worse than the record when one team has a better record than the other against teams .500 or worse:
|Category||Super Bowl||Super Bowl and
|Better record vs. winning teams||5-4||14-10|
|Better record vs. 8-8 or losing teams||6-3||18-9|
It is important to look at strength of schedule in order to recognize that the best teams don't win a lot of close games against the worst teams, but you can't make strength of schedule so important that you ignore what it means for a team to blow out its opponent. It's good to be able to win a closely-fought contest, but the best teams clobber their opponents so the outcome is never in doubt.
This article was written during the 2005 season. A slightly updated version was published in Pro Football Prospectus 2006. Here are the numbers for the four teams to make the conference championships in 2005.
Pittsburgh: 1 Gut, 5 Stomps, 1 Dominate, 1 Skate
Denver: 4 Guts, 4 Stomps, 1 Dominate, 1 Skate
Seattle: 2 Guts, 6 Stomps, 1 Dominate, 3 Skates
Carolina: 0 Guts, 2 Stomps, 2 Dominates, 4 Skates
Here's a table of all the teams to make the conference championships over the past decade, along with their totals in each of the win categories. OTHER WINS represents wins over .500 teams or wins of 9-13 points. QUAL WINS and QUAL GAMES are wins against opponents with winning records, and total games against opponents with winning records (to show that teams like the 1999 Rams didn't beat good opponents because they never played them in the first place). OTHER WIN% is the winning percentage against teams .500 or below.
Here's a look at the top teams in "quality win percentage" each year for the past decade, along with total record and postseason results. If there are not four teams listed for a year, that means there were not four teams that year with a quality win percentage above .500.
|1995||DAL||12-4||5-2||.714||won Super Bowl|
|1995||GB||11-5||4-2||0.667||lost NFC Championship|
|1995||BUF||10-6||4-2||0.667||won wild card, lost in second round|
|1995||PIT||11-5||3-2||0.600||lost Super Bowl|
|1996||BUF||10-6||6-3||0.667||lost wild card game|
|1996||DEN||13-3||3-2||0.600||bye, lost in second round|
|1996||GB||13-3||4-3||0.571||won Super Bowl|
|1996||DAL||10-6||5-4||0.556||won wild card, lost in second round|
|1997||GB||13-3||7-1||0.875||lost Super Bowl|
|1997||PIT||11-5||4-2||0.667||lost AFC Championship|
|1997||TB||10-6||6-5||0.545||won wild card, lost in second round|
|1998||MIN||15-1||4-0||1.000||lost NFC Championship|
|1998||NYJ||12-4||7-1||0.875||lost AFC Championship|
|1998||DEN||14-2||3-1||0.750||won Super Bowl|
|1998||ATL||14-2||3-2||0.600||lost Super Bowl|
|1999||TEN||13-3||3-1||0.750||lost Super Bowl|
|1999||TB||11-5||3-1||0.750||lost NFC Championship|
|1999||SEA||9-7||3-1||0.750||lost wild card game|
|1999||IND||13-3||4-2||0.667||bye, lost in second round|
|1999||BUF||11-5||4-2||0.667||lost wild card game|
|2000||TEN||13-3||5-1||0.833||bye, lost in second round|
|2000||DEN||11-5||4-1||0.800||lost wild card game|
|2000||STL||10-6||4-2||0.667||lost wild card game|
|2000||GB||9-7||6-4||0.600||did not make playoffs|
|2001||STL||14-2||6-1||0.857||lost Super Bowl|
|2001||GB||12-4||4-1||0.800||won wild card, lost in second round|
|2001||PIT||13-3||3-1||0.750||lost AFC Championship|
|2002||OAK||11-5||6-2||0.750||lost Super Bowl|
|2002||TEN||11-5||6-2||0.750||lost AFC Championship|
|2002||NO||9-7||5-3||0.625||did not make playoffs|
|2002||MIA||9-7||5-3||0.625||did not make playoffs|
|2003||NE||14-2||7-0||1.000||won Super Bowl|
|2003||STL||12-4||4-1||0.800||bye, lost in second round|
|2003||PHI||12-4||4-2||0.667||lost NFC Championship|
|2003||MIN||9-7||4-2||0.667||did not make playoffs|
|2004||NE||14-2||7-1||0.875||won Super Bowl|
|2004||PIT||15-1||6-1||0.857||lost AFC Championship|
|2004||PHI||13-3||2-1||0.667||lost Super Bowl|
|2004||ATL||11-5||2-1||0.667||lost NFC Championship|
78 comments, Last at 03 Apr 2006, 9:33pm by Dan Babbitt