Writers of Pro Football Prospectus 2008

Most Recent FO Features


» Futures: Nick Chubb & Sony Michel

The Georgia Bullddogs' dynamic duo should be on NFL rosters at some point in the next 72 hours. Which will be the better pro? That depends on what kind of running back you're looking for.

13 Aug 2004

Pittsburgh of Despair

by Aaron Schatz

Two weeks ago, when I announced that we had collaborated with Football Project on Brassey's Pro Football Forecast 2004, I listed all of the articles and statistics that we contributed to the book. Unfortunately, it looks like not everything we contributed to the book actually made it into the book. The biggest omission is the 2001 and 2002 DVOA and DPAR data, which is disappointing since I've been keeping the 2001 numbers off the site for months now so they would be a treat for people who bought the book. In addition, three essays written for the book did not make it in, and a fourth had a good amount of stuff cut for space. I feel a bit guilty because one of the essays that did not make it was the one about Green Bay's offense and defense, which I mentioned was written specifically to make reader "A Packer Owner" buy the book, and I hope that if he did buy the book he isn't disappointed (pieces of my essay are used in a different Green Bay essay, but the DVOA stuff is missing).

Despite the fact that a few of our contributions are missing, it is still the best preseason annual you can buy. You should click on that link over to the right and buy one now if you have not done so yet. Dan Lewis' long essay on fantasy football is a lot of fun and you get to play "spot the pop culture reference" in the subheadings on my big DVOA essay. Since we have this website, though, I also get to publish all the material that didn't make it in so that it can be enjoyed by Football Outsiders readers. The 2002 DVOA numbers are on the site already, of course (just click on the stats pages, and each lists 2003 data with a link to 2002) and I'll just have to move the 2001 numbers and commentary ahead of some other articles on my "to do" list. Over the next two weeks I'll republish the essays that didn't make it into the book, starting with this one on the Pittsburgh Steelers (slightly edited since it was written before the draft).

A Steelers article is also the perfect place to introduce our new staff member, Ryan Wilson, who will be blogging articles and also writing occasionally. Ryan writes the weblog focusing on the strangest combination of teams in existence, HeelsSoxSteelers, and also writes for Pro Football Digest. With Ryan in and Fritz and Bruce saying adios so they can concentrate on their own sites (who2.com and Boston Sports Media Watch), the staff count is now Patriots fans: 1, fans of other teams: 9.

The Pittsburgh Steelers went into 2003 as the favorites to win a third-straight AFC North title, and all seemed well after a 2-1 start. Five straight losses later, the season was completely shot to pieces. Pittsburgh stood at 2-6, three games behind the division-leading Ravens, and had pretty much given up. They stumbled home with four wins in their final eight games, but were only of interest to hardcore Steeler fans and fantasy football players scanning the waiver wire to see if Jerome Bettis might fill a hole on their team that week.

And yet, the Steelers were not that bad. DVOA ranked the Steelers #16 at midseason, when they were 2-6, and ranked them at #16 for the entire season as well. Pittsburgh declined in 2003, to be sure, but the decline was less from a good team to a bad team, and more from a good team to an average one. (What's DVOA? Learn more here about our innovative statistic that compares every play to league average based on situation.)

The 2003 Steelers, it turns out, are a textbook example of why strength of schedule has to be taken into account when looking at the NFL. The Steelers played by far the most difficult schedule of the first half of the season, and that schedule left them looking like a team that had crashed and burned instead of what they really were -- a team that was having some struggles but was no worse than the other three teams in their mediocre conference.

How much difference was there between the Steelers first eight games, and their last eight games? In Weeks 1-9, Pittsburgh's opponents averaged a 10-6 record, and included six different playoff teams. In Weeks 10-17, Pittsburgh's opponents averaged a 6-10 record, and included one playoff team, Baltimore. Some teams faced much better offenses in the first half of the year; others faced much better defenses. But Pittsburgh was the only team whose schedule was far more difficult against both opposing offenses and defenses over the first eight games (remember when you look at the table that defensive DVOA is better when it is more negative).

Harder First Half Schedules, by Opponent's Average DVOA
Played better defenses in first half Played better offenses in first half
Team Average Defense faced, Wk 1-9 Average Defense faced, Wk 10-17 Difference Team Average Offense faced, Wk 1-9 Average Offense faced, Wk 10-17 Difference
PIT -5.8% 4.1% -9.9% PIT 5.3% -4.4% 9.7%
KAN -4.1% 4.0% -8.2% CAR 1.8% -7.8% 9.6%
SDG -5.9% 0.5% -6.4% NYG 1.6% -5.1% 6.6%
JAC -5.2% 0.5% -5.7% GNB 3.7% -1.7% 5.4%
ARI -8.7% -4.0% -4.7% NOR -0.2% -5.1% 4.9%

Pittsburgh's first eight games weren't just harder when compared to their second eight games -- they were harder when compared to everyone else's first eight games. Only Houston faced a harder schedule of opposing offenses in the first half of the season, and only Arizona and San Diego faced a harder schedule of opposing defenses.

Hardest and Easiest Schedules, 2003, by Opponent's Average DVOA
Weeks 1-9 Weeks 10-17
Hardest Hardest
PIT 12.1% OAK 7.8%
ARI 7.5% DET 7.3%
HOU 7.1% NYJ 6.0%
NYG 5.7% CLE 5.5%
ATL 4.0% MIA 5.2%
Easiest Easiest
DAL -12.2% CAR -10.5%
BAL -5.3% PIT -9.5%
MIA -4.6% NOR -7.3%
PHI -4.1% KAN -7.2%
NYJ -3.8% STL -6.7%

There was lots of blame to go around, but the main scapegoat for Pittsburgh's early struggles was quarterback Tommy Maddox. The dominant meme among football journalists was that Maddox's comeback from purgatory (a.k.a. the XFL) had been a fluke, and that he was not really good enough to be a starting quarterback in the NFL. As Steve Grogan once said, the most popular player on a football team is the backup quarterback, and the name "Charlie Batch" was often heard in the bars and barbershops of Pittsburgh.

Was Maddox the main culprit in Pittsburgh's early season struggles? Maddox's year is difficult to analyze, for a number of reasons. First, the season was bookended by two extreme performances that both came against Baltimore, the league's best defense. In a Week 1 victory, Maddox went 21-of-29 for 260 yards and three touchdowns with no interceptions. It was one of the best quarterback performances of the year after adjusting for the opposition. Then, in a Week 17 loss, Maddox completed 14 of 27 passes for 108 yards. He was sacked five times, threw three interceptions and lost a fumble. It was one of the worst quarterback performances of the year -- even after adjusting for the opposition.

Tommy Maddox pass DVOA by week, 2003
(adjusted for opponent)
(not adjusted)
NFL Passer
Week 1 126.0% 88.4% 134.3
Weeks 2-8 -20.9% -25.4% 59.5
Weeks 9-16 19.0% 29.1% 84.5
Week 17 -123.8% -164.3% 22.4

Take out the Baltimore games and then split the season in half, using DVOA, and it becomes clear that Maddox was indeed struggling during the first half of the season. But his struggles looked even worse than they really were because of the strength of the defenses the Steelers were facing. When Maddox turned things around in the second half of the year, the improvement looked better than it really was because of the Steelers' easy second-half schedule.

On the other hand, a large part of the turnaround didn't really come across in conventional NFL statistics like Maddox's passer rating. That's because the main improvements in Maddox's game came in turnovers, which have a stronger impact on DVOA than on the NFL passer rating, and sacks, which don't count in the NFL passer rating at all. While Maddox's schedule got much easier starting in Week 10, his turnaround really began in Week 9 against Seattle, when he went 21-for-35 with no turnovers. In Weeks 2-8, Maddox threw 11 interceptions; in Weeks 9-16, until that last Baltimore game, he threw only three. In Weeks 2-8, Maddox was sacked on average 3.3 times per game; in Weeks 9-16, that was cut in half, to an average of 1.6 sacks per game.

Overall, for the year, Maddox was not as good a quarterback as he had been in 2002. He was about league average, which is about what people expected from him. The real collapse in Pittsburgh came in two other areas: the running game, and pass defense.

Most people already know about the trouble Pittsburgh had running the ball last season. Look way, way down at the bottom of our running back ratings and you'll find both Amos Zereoue and Jerome Bettis, ranked #48 and #49 in terms of per-play performance by DVOA. For those of you who prefer standard numbers, Zereoue and Bettis had matching 3.3 yards per carry averages, which is a bit like a newlywed couple filling their brand new two-car garage with a pair of matching Yugos. Used Yugos, without mufflers and with the car stereos torn out.

The problems with the pass defense have received less attention. The Steelers actually allowed fewer passing yards per game in 2003 than they did in 2002, but that's because their opponents were playing with the lead much more often. Yards per pass attempt by Steeler opponents rose from 6.04 in 2002 to 6.29 in 2003. The Steelers dropped from 19 interceptions and 50 sacks in 2002 to 14 interceptions and 35 sacks in 2003.

Pittsburgh's pass defense did improve in the second half of the year, but this is another place where strength of schedule comes into play. That easy schedule of opposing quarterbacks made the Steelers pass defense look like it had finally returned to the higher level of play exhibited in 2001 and 2002, when in reality the defense had merely improved from bad to pedestrian.

Impact of Schedule on Pittsburgh Pass Defense
(adjusted for opponent)
(not adjusted)
Weeks 1-9 12.0% 19.0%
Weeks 10-17 1.7% -10.1%

Did the Steelers take the steps necessary to fix the pass defense this offseason? It doesn't seem like it. Jason Gildon had declined over the past couple years, but he was still last season's sack leader, and he's gone now. People expect big things from Ricardo Colclough, but how big are the things you can expect right away from a rookie cornerback, let alone a rookie cornerback adjusting from NCAA Division II to the NFL? Their one free-agent pickup in the secondary was cornerback Willie Williams, a 12-year veteran who has started a grand total of one game in the past two seasons and isn't planning to start any games here either. There are a lot of encouraging stories coming out of Pittsburgh about how well the young defensive backs are playing in camp, but since when are there not a lot of encouraging stories coming out of every city in the NFL about how well young players are playing in camp, 99 percent of these stories turning out to be absolutely pointless by the morning of September 13?

The Steelers paid attention to their problems at running back, and signed Duce Staley to take over as the starter. They paid attention -- perhaps too much -- to the inconsistency of Maddox, drafting Ben Roethlisberger to serve as understudy and eventually replace him. As I pointed out in our AFC Over-Under preview, the Baltimore Ravens are more likely to decline this year than any other NFL team. The good news for Steelers fans is that their team may be able to stumble into the playoffs even if the pass defense doesn't improve. But can Pittsburgh fans really be satisfied with an 8-8 or 9-7 season followed by a total thrashing in the wild card round at the hands of some pass-happy wild card team like Kansas City, Indianapolis, or Tennessee?

Posted by: Aaron Schatz on 13 Aug 2004