Writers of Pro Football Prospectus 2008

Most Recent FO Features

GurleyTod16.jpg

» OFI: SEC Surprises

In an opening week where even the elite teams in college football looked mortal, the SEC had two big surprises in Texas A&M and Georgia defeating their South Carolinian opponents by big scores.

02 Dec 2003

Week 13 Team Efficiency Ratings

by Aaron Schatz

Here are the team efficiency ratings after Week 13, measured by our proprietary Value Over Average (VOA) system that breaks down every single play and compares a team's performance to the league averaged based on situation in order to determine value over average.  (Explained further here.)

Moving up: MIA (from #20 to #12), BAL (from #17 to #11), BUF (form #22 to #18)
Moving down: CLE (from #16 to #23), TAM (from #2 to #6), NYG (from $19 to #24)

Two weeks ago, Ian and Al had to promise to stop talking about the Tampa Bay Buccaneers.  After today's ratings commentary, I'm going to have to make the same promise.  This weekend was a real setback if I want people to take VOA seriously, because Tampa Bay and San Francisco -- the two teams I had ranked in the top ten despite losing records -- both spit the bit like the bit has never been spit before.  Tampa lost to a team with one of the league's worst records, Jacksonville, a game they should have easily won.  San Francisco lost to a team on the rise, Baltimore, but by a ridiculous 38 points.  It was spankings like this that made many states ban corporal punishment in the schools.

But please, bear with me as I indulge my discussion of Tampa one more time.  This is where I get to repeat, as I have said many times, that past performance does not guarantee future returns.  Right now, what we are measuring with the statistics on this site is how teams have performed so far; if they suddenly switch and play like a completely different team, we don't know of any way to predict that (unless the team you are trying to predict is playing a home game on Thanksgiving against a quarterback who for some unknown reason historically struggles in domed stadiums).

In the comments section on last week's ratings, a reader took me to task for constantly referring to Tampa Bay as underrated.  Part of the problem is that people just don't want to believe that a team can constantly lose close games despite being one of the best teams in the league.  Part of the problem is that we want the VOA ratings to be predictive even though they aren't.  But despite those things I must admit that I am still embarassed that the Bucs (and San Francisco and Oakland) keep coming out so high despite loss after loss.

Before looking at stats, I wanted to bring something up about coaches.  Eric McErlain of Off Wing Opinion has a post about this weekend's 49ers-Ravens debacle that "49ers head coach Dennis Erickson deserves to be fired right now. There isn't any reason why this 49ers team shouldn't be in the thick of the playoff race, and the only difference between this year's team and last year's is the man at the controls."

For the most part, I agree with McErlain.  I thought the Erickson hiring was a really strange move, given the talented coordinators (Charlie Weis, Romeo Crennel) and free agent coaches (Dennis Green) who were passed over.  But just like there isn't much difference between the 2002 and 2003 49ers personnel, there really isn't much difference between the 2002 and 2003 Buccaneers personnel.  Sure, there are some injuries, particularly to Brian Kelly and John Lynch, but plenty of teams overcome injuries (cough, Patriots, cough) and a couple injuries are the difference between the Super Bowl and the wild card, not the Super Bowl and 5-7.  Obviously, Jon Gruden has a far better track record than Dennis Erickson, and did even before he won a Super Bowl.  For why do most NFL fans (and I'm including myself here) tend to blame Erickson for the 49ers lost season without blaming Gruden for Tampa's problems?

So that's my subjective question for the day; now back to the objective stats.  I went back and looked at Tampa's record over the season, and I believe I've found a couple of reasons why their ranking on this site differs from public perception, not to mention the actual record.  The VOA ratings given here represent the entire season's play, with no single game -- or even single play -- treated differently than any other.  (The exception is the WEIGHTED DVOA number, a very rough attempt to make recent games count more than early season games.)  A team that played better earlier in the season will rate higher than you would expect; so would a team which has mixed dominant wins with close losses.  Tampa has both these problems.  Here is Tampa's rating, broken down by each game of the season.  These are the opponent-adjusted numbers, including offense, defense, and special teams.  I've also split the season in half.


Week 1 (W) 2 (L) 3 (W) 5 (L) 6 (W) 7 (L) Games 8 (W) 9 (L) 10 (L) 11 (L) 12 (W) 13 (L) Games
Opp. @PHI CAR @ATL IND @WAS @SFO 1-6 DAL NOR @CAR GNB NYG @JAC 7-12
DVOA +51% +25% +79% +20% +95% -79% +32% +101% -27% +6% -41% +27% -21% +9%

Tampa has seven losses, but in three of those losses the VOA system says that they played better than average (after adjusting for strength of opponent).  In two of these losses, both to Carolina, they straight-out outplayed their opponent.  In the third, the amazing Indianapolis comeback, they were outplayed but still rate positive because of the high quality of the Colt offense.  Those "tough-luck losses" boost Tampa's DVOA rating -- and so does the fact that the Bucs have had four dominant wins and only one big loss (to San Francisco).  So why do I use total DVOA and not the average of a team's games over the season?  It turns out that the average doesn't correlate any closer with won-loss record, and it correlates worse with points scored and allowed.  Total DVOA may be a better measure of a team's true quality.  But a team that wins five games by 20 and loses seven games by a field goal is still 5-7.

Now notice how Tampa's performance has declined over the course of the year.  Based on Tampa Bay's play-by-play performance, I've been touting the Bucs as underrated since very early in the season, but then again they were a lot better early in the season.  The Bucs are as inconsistent as everyone thinks, with their DVOA going up and then down and then up and then down all season long.  But the "ups" earlier in the season were higher than the recent "ups," and the "downs" earlier in the season were higher than the recent "downs" (heck, two of them were positive).  The split between the two halves of Tampa's season will look even worse at the end of the year, when you're splitting the season after eight games and that season-best performance against the Cowboys moves from the second half group into the first half group.

It turns out that a number of rankings on this site that differ from public perception do so because those teams have improved (NWE, TEN, PHI) or declined (TAM, SFO, MIN).  Take a look at the top teams ranked by DVOA for Weeks 1-7.  These ratings will differ from those here because I'm now using the full-strength opponent adjustments.  I'll also include the DVOA for these teams for Weeks 8-13, with rank, and you can see how things have changed:


Team DVOA
Weeks 1-7
Rank DVOA
Weeks 8-13
Rank
MIN 44.0% 1 -31.2% 31
SEA 37.0% 2 13.8% 8
IND 36.1% 3 11.1% 11
DAL 34.1% 4 -19.9% 27
TAM 31.5% 5 7.3% 13
KAN 29.3% 6 36.0% 2
SFO 25.4% 7 -12.0% 22
STL 23.5% 8 21.8% 5
DEN 16.4% 9 16.3% 7
BUF 13.7% 10 -19.5% 26

Well, that certainly demonstrates why some teams are ranked higher here than you would expect.  Minnesota, as we all know, has completely gone in the tank.  Did you realize that Dallas had collapsed as well?  That doesn't bode well for their matchup with Philadelphia this week.  Where is Philly, anyway?  And if Kansas City is #2 since Week 8, who is #1?  Here are some teams not included in the above table:


Team DVOA
Weeks 1-7
Rank DVOA
Weeks 8-13
Rank
TEN 4.3% 16 37.2% 1
NWE 4.1% 17 26.2% 3
PHI -24.8% 28 25.9% 4
BAL -14.0% 24 21.0% 6
JAC -18.3% 27 12.9% 9
CIN -15.4% 25 12.2% 10
CAR -9.5% 21 5.8% 14

Oh, there are the teams that people are talking about as Super Bowl contenders.  The top five teams over the past six weeks are the teams considered the favorites to win it all: TEN, KAN, NWE, PHI, STL.  Which is more surprising over the past six weeks -- that Baltimore is #6, or that Jacksonville is #9?  The difference in team performance from the beginning of the season to the end of the season gives me two things to work on in the offseason.  First, developing a system for weighting VOA ratings that I feel more comfortable with, so that we can have these rankings more in line with how teams are playing as of the current week.  Second, figuring out over the past couple years if teams that play better in the second half of the year are likely to improve the following year, as I believe is the case in baseball.  If so, Jacksonville could be the sleeper team of 2004.

Oh, and I included Carolina on this second table to demonstrate that they are the one team that VOA rates far lower than their record that hasn't been much better in recent games than they were early in the season.  They have played a little bit better, yes, but they've gone from slightly below average to slightly above average.  I still believe they are a one-year wonder.  And yet, they will probably grab a first-round bye in the NFC playoffs.  Right now, they are one game behind St. Louis and Philly, but the schedule for Carolina's last four games is astonishingly weak.  Their last four opponents' average DVOA is -32%.  The next-easiest remaining schedule belongs to Jacksonville, and their last four opponents' average DVOA is much higher, -14%.  Carolina finishes up with Atlanta, Arizona, Detroit, and the Giants.  Yikes.  At least that first game should be a lot harder than it looks, now that Michael Vick has returned.  St. Louis still needs to beat Seattle and the surging Bengals, while Philly has a series of unpredictable opponents: Dallas, San Francisco, and Miami.

Speaking of Dallas and Miami, how strange was Thanksgiving's result?  Not that Miami won the game, but that they won with such a strong offensive performance.  Prior to Week 13, Dallas was our #3 defense, while Miami was our #30 offense.  Looking at each team's week-by-week record for 2003, the Thanksgiving game stands out like Michael Jackson at a Nation of Islam rally.  According to DVOA, Miami's previous best offensive performance this year was Week 2 against the Jets, +10%.  Their offense against Dallas, adjusted for Dallas' previous defensive performance, rated +59% DVOA.  On the flipside, the Cowboy defense had only allowed positive offense in two of eleven games prior to Thanksgiving -- +27% by Atlanta in a Week 1 loss, and +19% by Washington in a Week 9 game that the Cowboys managed to win anyway.  But Thanksgiving day, the Dallas defense, adjusted for Miami's previous offensive performance, rated +49% DVOA.

Does this game signal that Miami's offense is better than we thought, and that Dallas' defense isn't quite as good as before?  Or was it just a fluke?  We'll have to see over the last four weeks, but I will note that according to the Football Outsiders quarterback ratings, Jay Fiedler has played far better than Brian Griese this year -- and that's before those ratings have been updated to include the Dallas game.

Surprising to see the Titans lose to the Jets. So strange to root for Gang Green, but they really did a solid for the Patriots last night.  Maybe we weren't that nuts when we rated Chad Pennington as the top quarterback of 2002.  By the way, not that I'm complaining as a Patriots fan, but that Samari Rolle non-interception in the first quarter last night got me thinking.  First, isn't it a bit absurd to enforce a taunting penalty on a celebration for a play that technically never happened?  Second, why is the throat-slashing gesture so bad anyway?  Who does it hurt?  Does anyone remember why this was banned other than the fact that the NFL doesn't want anyone basically doing anything?

  • Opponent adjustments are now at full strength and will be for the rest of the season.
  • As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.
  • All numbers are adjusted for opponent quality except for NON-ADJ TOTAL VOA.

Here are the ratings through Week 13:
 


TEAM
TOTAL
DVOA
LAST
WEEK
NON-ADJ
TOTAL VOA
W-L
OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
VOA
S.T.
RANK
1 KAN 33.5% 1 39.3% 11-1 22.6% 1 -6.1% 9 4.8% 2
2 SEA 25.8% 5 32.8% 8-4 22.0% 3 -3.4% 14 0.5% 15
3 IND 23.7% 3 21.8% 9-3 22.3% 2 -2.0% 16 -0.6% 23
4 STL 22.6% 7 27.8% 9-3 0.8% 11 -22.7% 2 -1.0% 26
5 TEN 18.6% 4 21.8% 9-3 14.3% 5 -4.1% 13 0.2% 17
6 TAM 18.6% 2 20.3% 5-7 -0.3% 14 -21.1% 3 -2.2% 30
7 DEN 16.4% 10 21.3% 7-5 -1.2% 15 -16.9% 5 0.7% 12
8 NWE 14.0% 9 13.2% 10-2 -1.5% 16 -14.1% 7 1.4% 9
9 SFO 9.0% 6 7.8% 5-7 6.5% 8 -4.4% 12 -1.9% 28
10 DAL 7.0% 8 14.5% 8-4 -7.8% 24 -15.1% 6 -0.2% 20
11 BAL 6.9% 17 3.7% 7-5 -22.9% 31 -24.1% 1 5.7% 1
12 MIA 5.1% 20 5.8% 8-4 -13.9% 27 -19.7% 4 -0.8% 24
13 OAK 4.9% 13 -2.2% 3-9 0.7% 12 -2.1% 15 2.1% 6
14 MIN 4.7% 11 12.3% 7-5 18.1% 4 12.2% 28 -1.2% 27
15 GNB 2.9% 12 0.7% 6-6 5.0% 9 2.7% 20 0.6% 13
16 PHI 0.7% 15 4.4% 9-3 4.6% 10 5.6% 22 1.7% 7
TEAM TOTAL
DVOA
LAST
WEEK
NON-ADJ
TOTAL VOA
W-L OFFENSE
DVOA
OFF.
RANK
DEFENSE
DVOA
DEF.
RANK
SPECIAL
VOA
S.T.
RANK
17 PIT -0.1% 14 -4.5% 4-8 -1.6% 17 -0.7% 17 0.8% 11
18 BUF -0.6% 22 -4.8% 5-7 -11.7% 25 -11.0% 8 0.1% 18
19 CAR -1.0% 18 -3.6% 8-4 -1.8% 18 1.8% 18 2.6% 3
20 NYJ -2.0% 21 -3.7% 5-7 13.8% 6 17.4% 31 1.6% 8
21 CIN -2.7% 23 2.9% 7-5 11.0% 7 10.5% 25 -3.2% 31
22 JAC -4.7% 25 -8.2% 3-9 -6.1% 21 -5.6% 10 -4.2% 32
23 CLE -7.9% 16 -13.9% 4-8 -12.5% 26 -4.8% 11 -0.3% 21
24 NYG -9.2% 19 -10.5% 4-8 -6.8% 23 1.9% 19 -0.6% 22
25 WAS -10.8% 24 -10.5% 4-8 -2.0% 19 9.9% 24 1.1% 10
26 SDG -12.9% 26 -23.6% 2-10 0.7% 13 12.7% 29 -0.9% 25
27 NOR -17.0% 28 -14.6% 6-6 -5.5% 20 11.5% 27 0.0% 19
28 HOU -17.6% 27 -21.9% 5-7 -6.7% 22 13.3% 30 2.4% 4
29 CHI -18.5% 30 -15.1% 5-7 -14.3% 28 6.4% 23 2.2% 5
30 ATL -27.5% 29 -28.8% 2-10 -17.3% 29 10.8% 26 0.5% 14
31 DET -27.7% 31 -28.9% 4-8 -24.7% 32 3.4% 21 0.3% 16
32 ARI -44.6% 32 -45.9% 3-9 -20.3% 30 22.2% 32 -2.0% 29

  • PAST SCHEDULE lists average DVOA of past opponents, while FUTURE SCHEDULE lists average DVOA of upcoming opponents.  Teams are ranked from hardest schedule (#1, most positive) to easiest schedule (#32, most negative).
  • The PAST SCHEDULE number will differ from the difference between DVOA and (non-adjusted) VOA because schedule strength is based on the opponent's total efficiency rating, while opponent adjustments to VOA take into account the situations faced within each specific game.
  • WEIGHTED DVOA combines the team's DVOA performance from each game.  The past four weeks are each weighted at 100%, while each week before that is weighted 5% lower, beginning with Week 1 at 60%.
  • VARIANCE measures the statistical variance of the team's weekly DVOA performance.  Teams are ranked from least consistent (#1, highest variance) to most consistent (#32, smallest variance).

TEAM
TOTAL
DVOA
W-L
PAST
SCHEDULE
RANK FUTURE
SCHEDULE
RANK WEIGHTED
DVOA
RANK VARIANCE RANK
1 KAN 33.5% 11-1 -3.8% 27 -7.0% 26 38.7% 1 20.4% 20
2 SEA 25.8% 8-4 -8.3% 31 -2.9% 18 27.7% 2 20.6% 19
3 IND 23.7% 9-3 0.4% 16 -2.9% 17 26.6% 4 20.3% 21
4 STL 22.6% 9-3 -8.6% 32 -4.3% 20 27.0% 3 31.6% 7
5 TEN 18.6% 9-3 -3.3% 25 5.9% 11 23.8% 5 24.3% 14
6 TAM 18.6% 5-7 -2.5% 24 -12.6% 30 20.2% 6 31.0% 8
7 DEN 16.4% 7-5 -2.0% 23 14.6% 2 19.0% 7 24.7% 13
8 NWE 14.0% 10-2 2.1% 12 -0.5% 15 18.5% 8 18.5% 23
9 SFO 9.0% 5-7 0.5% 15 -6.2% 25 9.0% 10 39.6% 3
10 DAL 7.0% 8-4 -8.0% 30 -10.9% 29 8.6% 11 28.5% 9
11 BAL 6.9% 7-5 3.7% 9 -4.8% 21 9.3% 9 22.0% 16
12 MIA 5.1% 8-4 1.3% 14 3.9% 12 7.3% 12 31.8% 6
13 OAK 4.9% 3-9 5.2% 4 -0.4% 14 -6.0% 22 11.3% 31
14 MIN 4.7% 7-5 -7.1% 29 -1.0% 16 3.7% 14 28.1% 10
15 GNB 2.9% 6-6 -0.2% 18 -5.1% 22 3.9% 13 20.7% 18
16 PHI 0.7% 9-3 -3.4% 26 2.8% 13 1.2% 16 13.6% 29
TEAM
TOTAL
DVOA
W-L
PAST
SCHEDULE
RANK FUTURE
SCHEDULE
RANK WEIGHTED
DVOA
RANK VARIANCE RANK
17 PIT -0.1% 4-8 6.4% 3 -3.5% 19 -0.1% 17 21.0% 17
18 BUF -0.6% 5-7 3.7% 8 11.7% 5 -1.5% 18 47.6% 2
19 CAR -1.0% 8-4 -1.0% 20 -32.0% 32 1.8% 15 9.9% 32
20 NYJ -2.0% 5-7 2.2% 11 6.0% 10 -2.6% 20 13.4% 30
21 CIN -2.7% 7-5 -1.0% 19 8.6% 8 -2.4% 19 15.2% 28
22 JAC -4.7% 3-9 8.2% 1 -13.9% 31 -5.3% 21 15.5% 27
23 CLE -7.9% 4-8 4.6% 5 13.2% 3 -10.8% 23 37.2% 4
24 NYG -9.2% 4-8 3.1% 10 -5.7% 24 -11.6% 24 32.0% 5
25 WAS -10.8% 4-8 1.4% 13 -5.1% 23 -13.6% 25 25.1% 12
26 SDG -12.9% 2-10 7.7% 2 -8.5% 28 -14.8% 26 22.9% 15
27 NOR -17.0% 6-6 -1.4% 22 -7.1% 27 -19.7% 28 20.2% 22
28 HOU -17.6% 5-7 4.4% 6 16.3% 1 -21.6% 29 16.6% 26
29 CHI -18.5% 5-7 -5.4% 28 8.2% 9 -18.7% 27 18.4% 24
30 ATL -27.5% 2-10 0.0% 17 10.8% 6 -32.8% 31 27.6% 11
31 DET -27.7% 4-8 -1.3% 21 13.2% 4 -31.8% 30 18.0% 25
32 ARI -44.6% 3-9 4.1% 7 10.5% 7 -51.5% 32 47.8% 1

PREVIOUS WEEKS:

Posted by: Aaron Schatz on 02 Dec 2003

comments