Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2018 Free Agency Cost-Benefit Analysis

Is Kirk Cousins the best free-agent quarterback in recent memory? Should Trumaine Johnson or Malcolm Butler have gotten the larger contract? And what makes a free-agent contract good or bad, anyway?

21 Sep 2004

Week 2 VOA Ratings

by Aaron Schatz

In 2003 the Miami Dolphins traded for an extra draft pick to take left tackle Wade Smith in the third round of the NFL draft.  To get this draft choice, they traded their 2004 second round pick to the New England Patriots.  The Patriots then turned around and traded this draft pick to Cincinnati for running back Corey Dillon.  This is the difference in the NFL between a winning franchise and a losing franchise.  In effect, the Dolphins gave up a running back with 244 yards over the first two games of this season in order to get an offensive lineman who allows more penetration than Jenna Jameson.

In Week 1, almost everything I had predicted came true -- the Packers looked great, the Steelers looked like they were back, the Panthers looked overrated, the Ravens looked terrible.  But the second week of the 2004 season gave us the strange, topsy-turvy mixture of results we've come to expect from the NFL, tossing all conventional wisdom into a blender, mixing it on "puree," and then attaching the blender to a stick of dynamite and blowing it to bits.  At least I'm still correct about the Jets.

After two weeks it is hard to tell which unexpected results demonstrate that teams are much better or worse than we thought, and which results are just due to the fact that in the NFL, any team really can win on any day.  After Week 1, you have to resist the urge to jump to conclusions, and after Week 2, you have to resist that urge even more.  The fact that we can't really do opponent adjustments to our statistics at this point makes this even clearer.  I mean, who honestly believes that Seattle suddenly has a below average offense but the best defense in the NFL?

Of all my projections for this season, however, there is one that I am ready to admit is wrong after only two weeks.  I wrote that it would be historic for a team like Tampa Bay to get worse this season when they had so many statistical indicators pointing to a rebound.  Well, Jon Gruden apparently decided to make history.  As Russell points out in his column this week, Tampa's offense is just appalling.  Over the first two games, the Tampa defense is playing at the level of the 2002 champions, but this team isn't going anywhere unless the offense suddenly puts it together.  With Keenan McCardell still holding out and everyone old looking, well, old, that doesn't look like it will happen.  I think this defense makes 8-8 more likely than 4-12, but there will not be double digit wins for Tampa Bay this season.

One prediction I'm not ready to admit was wrong is Jacksonville.  You may remember that I made a last-minute escape from the Jacksonville bandwagon because their offense was so horrid in the preseason.  Well, the Jags are 2-0, but of all the strange results of Week 2, the strangest by far was Jacksonville's 7-6 victory over Denver.  According to VOA (explained further here), Denver clearly outplayed Jacksonville.  If you read my comments on ESPN Page 2 yesterday, you know that Jake Plummer had the fourth-best game of any quarterback in Week 2 according to our ratings.  This game was decided by two issues.  First, Jason Elam missed a field goal from 51 yards, a field goal that the average NFL kicker hits about half the time.  Second, each team had one fumble from a running back. Remember, which team hops on the fumble is essentially random.  Jacksonville managed to hop on both of them, the second one with Denver driving and in winning field goal range.  Game over.

Look below at the ratings after Week 2 and you can see that it is pretty clear which team gets the "worst 2-0 team in the NFL" award.  Seven teams have won both games this season.  Six of them are ranked #1 through #6 in VOA.  Jacksonville is #26.  Of course, this is what I said about Carolina last year, so you never know what may happen.  That's why we watch the games.

As for the Lions, beware of the easy 2-0 start.  After two games in 2002, the top team in VOA was San Diego, which had beaten Cincinnati 34-6 and Houston 24-3.  They actually won two more games to start the season -- but finished 8-8.

I have some comments about running backs, including the big day from DeShaun Foster, but first let's run those team numbers...

* * * * *

Here are the Football Outsiders team efficiency ratings for the first two weeks of 2004, 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.)

Since the season is still young, there are fewer statistics listed than usual.  Most importantly, these ratings will not be adjusted for strength of schedule until Week 5.  OFFENSE and DEFENSE VOA are still adjusted to consider all fumbles, kept or lost, as equal value.  SPECIAL DVOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season.  NON-ADJ TOTAL VOA does not include these adjustments.

As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.

Before I give this week's numbers, I have to thank a few people who have made this process of breaking down play-by-play and turning it into VOA much, much easier.  This season, instead of cutting and pasting text files and running Excel formulas to parse text, I have a data parsing program that accurately breaks down about 98% of a week's plays in 30 seconds.  Programming on the data parser was started by a friend who I unfortunately can't name on the site -- let's call him "One-Time Omegalon" -- and then completed by Dennis Doughty, who is also taking care of ironing out the last few bugs in the parser and adding some fun new things (like tracking tackles, which may some day lead to better defensive stats).

Once I've pasted the data into my spreadsheets, I used to have to sort Pivot Table after Pivot Table to get the results that are given on the JUST THE STATS pages above.  Now, thanks to John Argentiero, I have Excel macros that take care it.  John also programmed the macros that compile each week's stats for the Loser League.  You'll find the Loser League results each Tuesday morning, at the Loser League link on the ARCHIVES menu above.  The Loser League itself was programmed by Patrick Laverty.

The last step in this automation will be to turn the tables from John's macros into the stats pages automatically.  If anybody knows of a utility which can turn Excel tables into PHP pages or HTML pages -- without including all the garbage codes that Microsoft throws in when you do a "save as HTML" -- please email me at aaron@footballoutsiders.com.  It would seriously make my life easier.

Finally, I still haven't spent enough time thanking Bob Sawyer and Benjy Rose for their work on the new site redesign.  If you like the new look of the site, and you need work done on a website, please consider their companies, B:COMPLEX Creative and Built for the Future

Team stats pages are updated now.  Player stats pages will be updated later today.

1 PHI 59.1% 49.6% 2-0 56.3% 1 7.2% 23 9.9% 2
2 NE 50.7% 54.3% 2-0 32.3% 4 -17.7% 7 0.7% 14
3 ATL 41.2% 58.8% 2-0 31.6% 5 -5.3% 15 4.3% 7
4 DET 28.9% 34.2% 2-0 4.7% 10 -21.3% 5 2.8% 8
5 NYJ 28.7% 42.7% 2-0 44.6% 2 17.7% 27 1.8% 10
6 SEA 24.2% 30.4% 2-0 -13.1% 22 -36.0% 1 1.4% 11
7 DAL 19.6% 5.8% 1-1 18.9% 7 0.3% 20 1.1% 12
8 TEN 6.9% 5.9% 1-1 1.6% 16 -5.6% 13 -0.3% 21
9 NYG 3.8% 14.1% 1-1 4.6% 12 -10.1% 10 -11.0% 31
10 DEN 3.0% -1.2% 1-1 3.2% 13 0.1% 19 -0.1% 17
11 IND 1.3% -5.5% 1-1 30.5% 6 35.4% 30 6.2% 6
12 KC -0.8% -8.0% 0-2 0.3% 17 -4.0% 17 -5.1% 27
13 CAR -1.4% -4.9% 1-1 2.8% 14 2.8% 21 -1.4% 22
14 PIT -3.8% -14.3% 1-1 -7.1% 19 -2.8% 18 0.5% 15
15 SF -5.3% -24.5% 0-2 -7.8% 20 -8.9% 12 -6.4% 29
16 BUF -5.7% -7.1% 0-2 -22.0% 26 -16.5% 8 -0.2% 19
17 OAK -6.6% 1.0% 1-1 -4.0% 18 -5.2% 16 -7.8% 30
18 CHI -6.6% -6.3% 1-1 -14.0% 23 -5.3% 14 2.1% 9
19 BAL -6.6% -0.1% 1-1 -16.8% 25 -13.9% 9 -3.7% 25
20 GB -8.0% -1.2% 1-1 6.7% 9 15.1% 26 0.3% 16
21 SD -8.7% 3.1% 1-1 4.6% 11 13.1% 25 -0.2% 20
22 CIN -8.9% -18.9% 1-1 2.7% 15 9.9% 24 -1.7% 23
23 NO -10.9% -5.5% 1-1 -16.0% 24 3.8% 22 8.9% 3
24 CLE -12.3% -3.5% 1-1 -42.7% 29 -18.6% 6 11.8% 1
25 WAS -18.9% -29.0% 1-1 -46.4% 30 -27.6% 4 -0.2% 18
26 JAC -19.0% -9.9% 2-0 -35.6% 28 -9.2% 11 7.4% 5
27 MIN -20.4% -5.2% 1-1 33.0% 3 42.3% 32 -11.1% 32
28 TB -21.2% -21.5% 0-2 -50.8% 31 -32.1% 3 -2.5% 24
29 MIA -23.5% -22.6% 0-2 -57.5% 32 -33.1% 2 1.0% 13
30 STL -28.5% -38.4% 1-1 16.4% 8 39.8% 31 -5.1% 26
31 HOU -37.6% -51.6% 0-2 -8.3% 21 23.7% 28 -5.6% 28
32 ARI -43.8% -29.3% 0-2 -27.1% 27 24.3% 29 7.5% 4

* * * * *

OK, let's talk about running backs a little bit.  The big rushing stars of Week 1 were Curtis Martin, Priest Holmes (both on my fantasy team, which was fun), and Quentin Griffin, who ran for 156 yards against Kansas City and had every NFL analyst in existence proclaiming that Mike Shanahan had once again turned a random running back into a star with the Denver system.

Then came Week 2, and Quentin Griffin rushed for only 66 yards on 25 carries.  Three out of four times he ran on 3rd-and-1, he was stuffed.  He fumbled a handoff in field goal range and cost the Broncos a chance to win the game.  Now, the troubles in Week 2 might not be entirely Griffin's fault.  I watched the NFL Network On Demand highlight package for this game this morning, and although they showed only one of the four short yardage attempts, that one was clearly a blown block by left tackle Matt Lepsis.  And I'm not sure why Shanahan was running in that situation anyway, instead of just putting the ball in the center of the field and letting Elam kick it.  But 2.6 yards per carry is a pretty bad day no matter what blocks were missed.  What's the deal here?  Can the Shanahan system turn any running back into a star, or not?

Maybe instead of Shanahan's system being able to turn any running back into a star, it is Kansas City's defense that can turn any running back into a star.  Which brings us to my favorite whipping boy, DeShaun Foster.  At halftime this week, Foster had 14 carries for 43 yards.  Two of those carries went for 17 yards and 13 yards, so the rest of the time Foster was averaging less than a yard per carry.  As I noted on the Open Thread, "he still sucks."  Apparently, however, when a resistible force meets a moveable object, the resistible force wins big time.  A few posts later, I was stuck writing this: "Foster runs for 71 yards. I knew that the Chiefs run defense would have to show their true colors eventually. Their suck far outweighs his suck."

Foster ended up leading NFL running backs in Week 2 with 174 yards.  He was celebrated as the NFL's latest rushing star by all the same NFL analysts who the week before were celebrating Quentin Griffin.  The collective short term memory of the NFL commentariat is staggering.  So is the inability to consider the context in which performances take place.  I think ESPN.com has their front page already laid out for every Monday for the rest of the season with a big picture of whatever running back played the Chiefs the day before.

Now, here's an interesting similarity between Griffin's performance in Week 1 and Foster's in Week 2 -- neither actually scores that highly according to the VOA system.  Griffin's Week 1 performance was worth 3.0 PAR (Points Above Replacement, explained here), which was eighth among running backs.  Foster's Week 2 game was worth 2.8 PAR, which was eleventh.  That's before taking into account any kind of opponent adjustment that considers Kansas City's poor run defense.  Both running backs had the same problem, mixing in a few great long runs with a lot of losses and one yard stuffs.  You can see the issue with both running backs when you look at our new statistic, Running Back Success Rate

("Wait," you say, "don't you mean Running Back Batting Average?"  Well, we took a poll of readers as part of the Loser League signup, and the overwhelming majority of you didn't like the whole "make it look like batting average" aspect of the statistic.  So we're ditching the batting average motif, and the subtracted .200, and we're making it look like a football statistic instead -- a straight percentage, like a quarterback's completion rate.)

Despite all the yardage, Griffin was successful on only 39% of his runs in Week 1.  Foster in Week 2 was not much better, successful on only 41% of his runs.  Foster lost yardage on eight carries.  That's astonishing -- on one of every four carries, Foster went backwards.  Three more carries went for zero yards.  Contrast this with another running back who also had 32 carries in Week 2, and fewer yards:

Week 2 Carries, by
Yards Gained
DeShaun Foster Corey Dillon
Carries Yards Carries Yards
lost yards 8 -21 1 -1
0-3 yards 12* 19 12 16
4-7 yards 4 20 11 57
8-10 yards 3 27 5 42
11-20 yards 4 58 3 44
21+ yards 1 71 0 0
Total 32 174 32 158
*Does include one successful play, TD on 1st-and-goal.

The result of this is that DeShaun Foster was successful on 41% of his carries, and rates 2.8 PAR.  Corey Dillon was successful on 59% of his carries, and rates 6.7 PAR.  Foster was the big star, but Dillon was just better this week.  So were Thomas Jones, Chris Brown, Edgerrin James, and even T.J. Duckett (9 carries for 52 yards, 3.4 PAR) and Charlie Garner (13 carries for 75 yards, 4.1 PAR).

By the way, the Panthers seem to accept Foster's feast-or-famine performance.  An article on their website talks about how Foster is a "Barry Sanders style" running back and notes, "The multiple carries that lose yardage or gain very little must be tolerated because the big one is always a possibility with Foster."   This is going to come back to bite the Panthers.  Every time DeShaun Foster loses yardage, he makes things harder on the Panthers offense.  It takes a lot of those long carries to outweigh that damage.  On Sunday, he had enough of them to be worth positive value.  On most Sundays, when he's playing defenses above a high school level, he will not.  I haven't been able to go back to measure Barry Sanders yet with my formulas, but I have a feeling he had more positive carries to outweigh his losses than DeShaun Foster does.

Finally, this thought: Since Foster and Griffin both had games that mixed bad runs with great ones (Griffin had four runs for a loss, four for no yardage), perhaps we are all wrong about the Kansas City defense.  ESPN's Scouts Inc. (premium link) write this week that "the Chiefs are not only missing tackles, but when they do tackle it is on the second level because they are losing the battle at the line of scrimmage."  But when you stop the running back for a loss on one out of four plays, are you really losing the battle at the line of scrimmage?  Perhaps the Chiefs defensive line isn't that bad, but when they do let the running back through, the linebackers and secondary can't tackle or pursue in the slightest.  Certainly that seemed to be the issue with Foster's 71-yard run.  Foster slipped through a small hole to the left of guard Tutan Reyes, who was blocking Chiefs' middle linebacker Monty Beisel, and after that there were hardly any Chiefs near him except for someone from the Chiefs secondary (Dexter McCleon, I think, it is hard to tell from the NFL On Demand highlights) who dove and missed a tackle.  This wasn't 71 yards of cutting and avoiding tacklers, this was bad defense.  In fact, Beisel -- normally a special teams player, and filling in for the injured Kawika Mitchell -- is going to be a serious problem for the Chiefs.  The Panthers' offensive linemen and fullback Brad Hoover were manhandling him all day.

Posted by: Aaron Schatz on 21 Sep 2004