DVOA Analysis
Football Outsiders' revolutionary metrics that break down every single play of the NFL season

AFC Midseason Report

by Aaron Schatz

This week and next, I'm going to take a look at where each team stands at the midpoint of the season, including breakdowns of offense, defense, and special teams.  This week is the AFC.  Next week is the NFC.  The following week we'll do the Peter King homage contest, the sequel to our TMQ homage contest.

I've got a lot of numbers for those who like statistics and a lot of commentary for those who don't.  But before we get to the commentary, I have to give the requisite explanation of the numbers.  For those new to Football Outsiders, VOA takes every single play during the season and compares the result to the league average, normalizing for a number of variables including down, distance, location on field, and so on.  I'm using this rather than simply yards or first downs, because VOA rewards progress towards both.  VOA is further explained here.  Remember that since VOA represents efficiency at scoring points, defensive VOA is better when it is negative.  In total VOA, you add in the opposite of defense so that better defense leads to a higher total VOA.  Trust me, it makes sense.

Most offense and defense numbers (including pass/run splits) are given as DVOA, which is VOA adjusted for the quality of opponents faced.  To show how teams have been affected by their schedules so far, I also give VOA for total, offense, and defense.  I've colored these non-adjusted numbers red so they stand out.  (Note: The adjustments for rushing based on current game score aren't quite perfect, so you'll notice a few bad teams who face offenses running out the clock are higher rated in rush defense than they should be... I'm working on fixing that.)

Special Teams numbers represent actual points.  FG/XP gives the points the kickers have been worth compared to how often average kickers hit from the same distances.  The other four numbers give the amount of "expected points" that a team gains in field position compared to average special teams play.  You can read more on that here.  I had to shorten punt return and kick return to "Pu-RET" and "Ki-RET" to save space so forgive the ugly headings.  These numbers are then turned into a total DVOA number which can be added to offense and defense to get total team efficiency.  (Note: Since special teams performance is based on 2002 averages, and early 2003 performance has been much better than last year's performance, you'll notice that there are more positive special teams than negative.  I can't fix that until the offseason.)

Individual player discussions in the comments may discuss the statistic DPAR, which is defense-adjusted Points Above Replacement.  This represents the total number of points scored due to plays where this player passed/ran/caught the ball, compared to a replacement-level player in the same game situations.  DPAR is further explained here.

Strength of schedule numbers give the average DVOA for all teams that have been played (past) or will be played for the rest of the season (future).  The final number given, projected record, is an estimate of the team's final record.  For all remaining games, wins are rewarded to any team 40% higher than its opponent in total DVOA, all games pitting two teams within 10% of each other count for .5 wins to each team, and games where the gap between opponents falls between 10% and 40% are on a sliding scale of determining how likely each team is to win.  (WARNING: This is totally unscientific, a broad guess on my part, not based on analysis of how teams have done in the past based on mid-season DVOA.  To paraphrase the great Keith Woolner, "DVOA and DPAR aren't meant to be predictive, but are instead evaluating the value or quality of a player or team's performance in the past.") Numbers in parentheses are estimated playoff seeding.

One more note: you may remember our early article that introduced the Pythagorean Theorem to football.  Introduced by Bill James, it says that a baseball team's record can be estimated using runs scored and allowed.  It works for football as well, and when a team's Pythagorean record differs from its regular record, I've noted that.  This could be an indication of luck, and luck can turn around.

All included numbers are through Week 9.  All statistics (except for non-adjusted VOA) are given a rank from #1 (best in league) to #32 (worst in league).  Schedule ranks go from #1 (hardest, most positive) to #32 (easiest, most negative).  If the math confuses you, just ignore the stats and focus on the rank.  If your team is #28 in everything, they suck.  If they are in the top five in everything, you are probably living in Kansas City.

If you would like to see what we wrote about these teams prior to the season, you can check out these IN FOCUS articles: AFC East, AFC North, AFC South, AFC West.


BUF -3.2% 18 -4.7%
MIA 5.2% 14 7.8%
NWE 11.4% 10 14.1%
NYJ -6.4% 21 -5.6%
-14.1% 26 -15.1% -15% 25 -13% 24
-16.7% 28 -12.9% -13% 24 -20% 30
1.2% 14 2.4% 17% 8 -16% 28
8.1% 8 4.3% 21% 5 -8% 20
-5.9% 12 -5.4% -6% 13 -6% 12
-24.2% 2 -22.9% -25% 3 -24% 4
-9.1% 10 -10.6% -12% 10 -5% 13
18.2% 31 13.6% 20% 26 17% 32


  DVOA Rank XP/FG Rank PUNT Rank Pu-RET Rank KICK Rank Ki-RET Rank
BUF 0.9% 17 0.7 20 5.7 10 -4.7 23 3.2 20 -0.4 16
MIA -1.0% 26 -4.6 30 0.8 19 -6.8 27 6.1 9 -0.5 17
NWE 1.1% 15 -2.9 28 -1.2 21 -6.0 26 12.8 2 3.6 7
NYJ 3.6% 3 1.1 18 -2.4 25 10.6 2 3.9 18 5.4 4
SCHEDULE Projected
Past Rank Future Rank
-0.5% 23 9.3% 1 7-9
0.6% 17 0.3% 18 9-7 (6)
0.6% 18 2.4% 14 11-5 (3)
1.4% 13 4.6% 11 5-11

NEW ENGLAND PATRIOTS (7-2, Pythagorean 5-4)

Admit it, you have no idea how they are doing this.  Everyone is injured.  They signed the biggest free agent of the year and lost him for the season.  They have no running game.  They lost their first game 31-0 (which explains the difference between the regular and Pythagorean records).  The kicker, best in the league last year, is in some kind of strange field goal slump.  Actual problem or fluke of sample size?  Who knows.  And yet, this team is 7-2, and now players are coming back, and three of their last four games are at home.  They have solved their rushing defense problem from last year despite Ted Washington's injury, and the passing game has completely changed in the past three weeks.  Check out Tom Brady's Points Above Replacement:

Tom Brady DPAR per game
Week 1 -9.6
Weeks 2-6  1.4
Weeks 7-9  7.9

By the way, while I was writing this the Patriots were called for encroachment another 28 times.

MIAMI DOLPHINS (5-3, Pythagorean 6-2)

Last year, I wrestled with the fact that Ricky Williams came out as an average back according to my numbers.  This year, there's no wrestling.  Ricky Williams and the Miami running game have been horrid.  Forget about average, Ricky shows up with -1.2 DPAR -- that's 1.2 points below replacement.  A schmo off the bench with an offensive line of backups would run better than Ricky Williams and the Miami offensive line have this year.  I know that he occasionally breaks off some amazing runs in the open field, dodging tacklers.  I know that he can run again and again without tiring in the fourth quarter.  But when he's this bad to begin with, what's the point?  Last year, his mediocrity was at least worth something because he pounded at the defense and freed up the passing game.  This year, yikes.  And I don't blame him entirely, the offensive line is awful, especially Wade Smith, as I have already written.

Can we knock it off with the "Brian Griese will save this team" nonsense.  A good half against San Diego does not a Pro Bowler make.  Griese's DVOA is -0.3%.  Fiedler's DVOA is -2.4%.  They are the same guy.

This team is doing it entirely on defense, more than any other in the league.  More than Baltimore, more than Dallas, more than Tampa Bay.  This defense is amazing and this offense is pathetic.  And the special teams aren't that great either.  And we all know that no matter what any numbers say, they are going to crash in December, and if that happens you can give one or two of their projected wins to the other teams in this division and you can send Denver to the playoffs.

Just to make it clear, in Boston we live and die by our baseball team, which constantly chokes, and our football team, which has come close to leaving town about a hundred times, always wins in amazing fashion and has a Super Bowl title.  In Miami they they live and die by their football team, which constantly chokes, and their baseball team, which has come close to being contracted by the league and has almost no fan following, always wins in amazing fashion and now has two World Series titles.  This is The Almighty's cruel joke on both cities.


Who knows what this team is all about, I sure don't.  Their variance, a stat measuring lack of consistency from week to week, is far higher than any other team in the league.  Guess who's fault it is?  I'll give you three guesses, and I know Ian and Al will get it on the first one.  Some days Drew Bledsoe is Johnny Unitas and some days he's Johnny Rotten.  We lived with this in New England for years and now you get to enjoy it in Buffalo.  In the words of Nelson Muntz, "Ha ha!"  As random as Bledsoe has been, the running game has been just as stable and consistent.  They consistently suck.  I stand by my belief that Travis Henry is the most overrated player in football, a man who turns bad performance into good fantasy stats with a ridiculous number of opportunities.  If they still had analogies on the SAT, they could say "Travis Henry is to TDs as Joe Carter is to RBIs."

Enjoy this table of the four aspects of the Bills (special teams, which are far more inconsistent in general for all teams, are not included).  Here is the DVOA every week of the season along with variance.  Yes, DVOA, meaning the horrid Bledsoe performances against Miami and Kansas City were actually worse than this, and have been adjusted upwards.

  NWE 1 (W) JAC 2 (W) MIA (3) L PHI 4 (L) CIN 5 (W) NYJ 6 (L) WAS 7 (W) KAN 8 (L) VARIANCE
Pass Offense 62.8% 158.1% -111.2% 19.4% 13.0% -74.8% 85.9% -152.6% 111.6%
Rush Offense -27.6% -17.1% -7.5% -45.2% -23.0% -39.4% 8.6% 17.1% 4.8%
Pass Defense -127.4% 11.3% -60.6% 76.5% 2.5% 44.3% -13.5% 52.8% 44.0%
Rush Defense 30.1% 24.3% 12.1% -7.5% -49.8% 12.3% -86.2% 4.8% 16.2%

NEW YORK JETS (2-6, Pythagorean 4-4)

Well, I already wrote a ton about the Jets after Week 6.  Back then, the Jets were overrated; since then, they've played down to their record.  It's nice to get back the player we rated as the #1 quarterback in 2002, but he can't save that awful defense.  If they were going to make a big comeback, they needed to beat the Giants in a game that never should have been close.  Now, they are stuck needing to go 7-1 to make the playoffs and that ain't gonna happen.


BAL -3.6% 19 4.3%
CIN -13.3% 26 -7.4%
CLE -0.2% 15 -8.1%
PIT -1.6% 16 -10.3%
-18.6% 30 -15.6% -39% 30 -4% 18
-2.1% 17 -1.2% 10% 13 -18% 29
-6.8% 22 -10.5% -6% 20 -8% 19
-5.5% 20 -12.2% 10% 14 -25% 32
-11.7% 6 -16.6% -10% 11 -14% 5
6.7% 21 1.8% 11% 22 2% 25
-7.5% 11 -3.3% -14% 9 -1% 18
-4.2% 14 -2.1% 15% 24 -24% 3


  DVOA Rank XP/FG Rank PUNT Rank Pu-RET Rank KICK Rank Ki-RET Rank
BAL 3.5% 5 4.5 12 6.8 8 0.7 11 4.9 14 1.0 11
CIN -4.4% 31 4.5 13 -8.7 31 -8.5 28 3.2 21 -13.0 32
CLE -0.9% 24 0.4 23 0.9 18 -11.1 30 5.0 13 0.4 12
PIT -0.3% 20 0.7 21 1.8 16 -0.7 13 -3.1 24 -0.3 14
SCHEDULE Projected
Past Rank Future Rank
-4.9% 31 7.4% 4 8-8 (4)
-1.8% 25 6.5% 5 5-11
3.7% 8 5.4% 8 6-10
10.8% 1 -6.8% 28 6-10


Look at that first table.  Yes, DVOA says that Baltimore is actually the third-best team in this division.  Two columns over, VOA says that they are the best team in the division if you don't adjust for strength of schedule.  And that, ladies and gentlemen, is the reason why Baltimore is in first place.  It isn't that they are that good.  They aren't.  Their offense is pathetic.  Their defense isn't as good as it was a few years ago.  They are a mediocre team in a division filled with mediocrity.  I'm falling asleep just thinking about the AFC North.  Nonetheless, they have a two game lead on a team that also has a difficult upcoming schedule, Cleveland, and a three game lead on a team that is better but has probably given up at this point, Pittsburgh.  Tennessee and Indy fans can take some satisfaction in the fact that whoever doesn't win the AFC South will get to destroy the Ravens in the first round of the playoffs.  Remember that for your fantasy playoff draft.

By the way, check out the RB ratings.  Jamal Lewis has a DVOA of +13.6% and the entire Baltimore rush offense has a DVOA of -4.1%.  Kyle Boller is more immobile than the ratings for Joe Millionaire 2 and the Ravens should never, ever let him run with the football.  EVER.


You only have yourselves to blame.  Everyone saw you guys as the big second half team.  You had the chance to go to Arizona and make a statement.  And you did.  The statement you made was, "Never, ever, ever believe in the Bengals."  To quote the great Robert DeNiro, "I gave you a chance to be a cop -- AND YOU BLEW IT!!!"


Last year, Kelly Holcomb was clearly better than Tim Couch.  This year it looks like Kelly Holcomb and Tim Couch have been about the same.  Our numbers say that isn't so.  Thanks to games against Baltimore and Indy, Holcomb's defense-adjusted numbers are much better than Couch's.  He also gets a boost for playing San Francisco from behind -- that sounds odd but the 49er pass defense has been very good this year when the 49ers are winning except against Holcomb (and the 49er pass defense has been awful when the 49ers are losing).  The result looks like this:

Holcomb 8.0 -2.4 4.1% -18.3%
Couch -3.0 -0.9 -19.8% -15.2%

So that 6-10 projected schedule could be 7-9 or 8-8 if Holcomb's struggles really are only a function of schedule.  Though the schedule is hard going forward, it isn't as hard as Baltimore's.  They could make it a race.

PITTSBURGH STEELERS (2-6, Pythagorean 3-5)

This is a huge disappointment, but we're all blaming the wrong people.  Tommy Maddox has taken the lion's share of the blame for the Pittsburgh collapse.  But once you adjust for the very difficult early Pittsburgh schedule, he's the same guy as last year.  His DVOA was 4.4% last year, and it is 2.5% this year.  That's not really a significant difference.  No, let's blame the rush offense and the pass defense.

The rush offense has gone from -3.8%, a little below average, to -24.5%, worst in the league.  A lot of that is Bettis (-19.5%) and Zereoue (-8.5%), but you also have leadfoot Maddox (-45%), and constant WR end-arounds are clearly not working this year (combined -20% DVOA on 17 runs).

The pass defense has gone from -6.5% DVOA, #10 in the NFL, to +14.8% DVPA, #24 in the NFL.

There is good news, as Pittsburgh's schedule is about to get a lot easier.  But will it be enough to overtake the Ravens?  Football is an emotional sport, and while the numbers say that Pittsburgh has a chance to rebound, that's gonna take quite a coaching job from Cowher.  Teams have been able to win without running games (Hi, Titans!) but without improved pass defense the Steelers won't get to eight wins no matter how easy the schedule looks.



HOU -11.6% 24 -18.9%
IND 35.6% 2 30.8%
JAC -17.9% 28 -15.9%
TEN 9.8% 12 10.7%
-2.3% 18 -4.1% 5% 15 -9% 22
23.3% 2 20.0% 45% 2 -3% 14
-11.3% 25 -12.2% -8% 22 -15% 26
17.8% 4 18.0% 52% 1 -20% 31
12.1% 27 17.6% 24% 28 1% 21
-10.0% 9 -8.6% -23% 4 5% 27
2.3% 18 -0.5% 7% 17 -3% 14
7.4% 23 6.7% 11% 21 1% 23


  DVOA Rank XP/FG Rank PUNT Rank Pu-RET Rank KICK Rank Ki-RET Rank
HOU 2.8% 6 0.0 25 7.5 7 -1.6 15 8.5 6 -0.1 13
IND 2.3% 8 13.8 1 5.0 13 -10.4 29 0.9 22 2.5 8
JAC -4.5% 32 -0.6 26 -1.7 23 -2.1 17 -15.8 32 -2.8 25
TEN -0.6% 22 10.9 3 5.7 11 -4.1 21 -14.5 31 -0.9 18
SCHEDULE Projected
Past Rank Future Rank
6.0% 5 1.5% 16 6-10
-0.2% 20 -6.5% 27 14-2 (2)
1.2% 14 2.0% 15 4-12
-0.3% 22 -1.1% 21 11-5 (5)

INDIANAPOLIS COLTS (7-1, Pythagorean 6-2)

If you are one of those people who believes that passing statistics are far more important to winning than rushing statistics, this is your team.  The rush offense and defense are basically the same as last year.  The passing offense, which was +10% DVOA in an off-year last year, is back at the top of the league.  But the amazing change is the passing defense.  Last year the Indy passing defense was ranked #26, allowing +9.9% DVOA.  This year, they are #4, allowing -23.1% DVOA.  While the Colts are only one game ahead of the Titans, our numbers say they are a significantly better team and have an easier remaining schedule.  They are the clear favorite for the division title and a first-round bye.

TENNESSEE TITANS (6-2, Pythagorean 5-3)

Look at those Tennessee offensive numbers.  This team has NO RUNNING GAME.  Zip, zero, nada.  Eddie George was great once but he has been a stiff for years now.  So what happens?  Steve McNair has to make it happen on third down.  The Flaming Thumbtacks average 7.6 yards to go on third down, the fifth-highest number in the league (the top four are WAS, OAK, NWE, NYG). Then the Tennessee offense has a +50% DVOA on third downs, tied for the best in the league with Minnesota.  By the way, that's +75% DVOA on third down passing, and -23% DVOA on third down rushing.

Their defense is below average.  Their special teams are below average.  This team has nothing but an amazing quarterback who is constantly bailing out the horrid running game with amazing third down conversions.  If Steve McNair isn't the MVP, there is seriously something wrong with the voters in this league.

But I have some numbers run for an article I still haven't written about the importance of a balanced offense.  While passing statistics usually correlate much better with winning than rushing statistics, teams are more likely to win when they have a balance of the two.  And balance, well, the Titans are about as unbalanced as a breakfast consisting of 50 slices of bacon and a grape.  That's why I don't think they'll be winning the Super Bowl, or even the division, in 2003.  

HOUSTON TEXANS (3-5, Pythagorean 2-6)

Oh Stacey Mack, why do you hurt me so?  Why did I call my preseason article on underrated running backs Stacey Mack Daddy and not Moe Williams No Problems?  The difference between Mack (-15.7% DVOA) and Dominick Davis (+11.1% DVOA) has been real, not a product of situational bias or an easier schedule for Davis.  But as the Houston offense has improved as the season moves on, so the Houston defense has faltered.  The Texans are getting more and giving more.  Which is why they are still mediocre.

JACKSONVILLE JAGUARS (1-7, Pythagorean 2-6)

The team that defines "meh."  Raise your hand if you thought the Jaguar passing game would be higher rated than the Jaguar running game.  OK, you with your hands up are lying.  Last year, the Jacksonville offense was ranked #7 in DVOA.  This year, they are #25.  If you want a young quarterback to learn, you gotta suffer.


DEN 5.8% 13 5.6%
KAN 37.8% 1 42.9%
OAK -2.5% 17 -1.4%
SDG -17.8% 27 -19.6%
2.1% 13 1.7% 3% 16 1% 12
16.2% 5 19.9% 17% 7 15% 4
-1.5% 16 2.0% -7% 21 7% 8
-0.4% 15 -5.3% -21% 26 24% 1
-1.6% 15 -1.9% 4% 15 -9% 7
-16.3% 5 -17.7% -34% 2 7% 29
2.3% 17 4.7% 8% 18 -2% 16
17.8% 30 14.7% 33% 31 1% 22


  DVOA Rank XP/FG Rank PUNT Rank Pu-RET Rank KICK Rank Ki-RET Rank
DEN 2.1% 10 9.6 4 -10.3 32 -3.7 19 18.1 1 -1.8 21
KAN 8.9% 1 2.4 16 13.3 1 12.2 1 4.8 15 13.0 1
OAK 1.3% 14 8.1 6 9.1 5 -3.9 20 -5.2 27 -1.1 19
SDG -0.9% 25 0.8 19 9.2 4 -0.4 12 -6.5 29 -7.8 30
SCHEDULE Projected
Past Rank Future Rank
0.2% 19 5.2% 10 8-8
-3.0% 27 -8.0% 30 15-1 (1)
-4.5% 30 5.8% 6 5-11
-0.3% 21 3.5% 13 3-13

KANSAS CITY CHIEFS (8-0, Pythagorean 7-1)

How do they kick ass?  Let us count the ways.  Top five in offense, defense, and special teams.  The rush defense hasn't been good but a lot of that is the fact that the variables aren't yet perfect when teams are protecting huge leads.  The special teams are so far ahead of the rest of the league it hurts.  Come on, when is the last time a kick returner was a guest on the Letterman show?  The only member of the special teams who isn't dominating the league is 83 years old.  Oh, and they have the third-easiest schedule in the league going forward.  Egads, they still get to feast on Chicago, Detroit, and San Diego for a second time.  The one loss listed in the projected record consists of a slight chance of losing to Denver or Cleveland and a 40% chance of losing to Minnesota.  I'm guessing that the '72 Dolphins pop the cork Week 14 in Whatever They Call It Field at Mile High.  One warning: Read the Pythagoras Grades the Coaches article from the beginning of the season.  If this team beats the Pythagorean projection, it will be the first Dick Vermeil team to do so since 1978.  Eight straight Vermeil teams have lost more close games than they should have.  You still can't assure me this won't come back to bite the Chiefs if they play a close game in the playoffs, and that's the kind of situation tailor-made for the Patriots to make another unexpected trip to the big one.

DENVER BRONCOS (5-4, Pythagorean 6-3)

Denver is just an amazing franchise.  They are the leading exemplar of the idea that the system and the line makes the team work, not the skill players.  If you enjoyed our first movie, "Anyone can be a star running back," you'll enjoy the sequel, "Crappy quarterback becomes MVP candidate."  Until said quarterback gets injured, and the backups don't prove that successful:


Plummer 4 1/2 20.3 30%
Kanell 2 1/2 -0.7 -16%
Beuerlein 2 -11.8 -62%

Nevertheless, if Plummer comes back and plays like Plummer was playing like, and Miami does its usual December el foldo, and the Dolphin offensive line continues to play like the Sharon High School offensive line, the Broncos can sign up for a likely first round trip to Indianapolis or Foxboro.

OAKLAND RAIDERS (2-6, Pythagorean 3-5)

Somewhere in the deepest, darkest regions of the Oakland Coliseum, there was a room where Al Davis kept a secret stash of paintings.  Hanging from the wall were pictures of grizzled old men, withered veterans of football.  And while this veritable gallery of Dorian Gray stood silently, unknown, hidden in the cavernous halls of the Raider offices, men like Rich Gannon, Jerry Rice, Tim Brown, Trace Armstrong, and Rod Woodson prowled the field like the young men, full of vim and vigor and cutting through the AFC like a chainsaw through Texas co-eds.

And then, the week before the Super Bowl, Barret Robbins in a drunken stupor apparently burned the whole room to ashes, and now the Raiders are left with a team filled with past-their-prime veterans who are playing like past-their-prime veterans.  The Super Bowl was ugly, the season so far has been uglier, and it is about to get ugliest.  Two words: Rick Mirer.  One number: 6, the rank of the rest of Oakland's schedule.  And one hope: Could the Raiders have fallen so far that they can dress Eli Manning in the silver-and-black for 2004?

SAN DIEGO CHARGERS (1-7, Pythagorean 2-6)

Hey, here's a link to a really good NFL draft site.  I hear some defensive players will be available this year.

Next week, the NFC.


There are no comments yet.