Writers of Pro Football Prospectus 2008

Most Recent FO Features


» The Deep Ball Project

Guest columnist John Kinsley breaks down the tape of every deep pass in the NFL in 2017 and comes away with a shocking conclusion: even without Andrew Luck, the Colts had the best long-ball quarterback in the league.

31 Oct 2006

Any Given Sunday: Raiders over Steelers

by Ned Macey

The defending Super Bowl champion Pittsburgh Steelers are 2-5. They just lost to the hapless Oakland Raiders, who thanks to Sunday's game had the same 2-5 record as Pittsburgh. Other 2-5 teams include Buffalo, Houston, Tennessee, and San Francisco. Can a team that keeps that sort of company really be a good team?

Football Outsiders DVOA rankings think so, ranking Pittsburgh as the 11th best team in football. (For a full explanation of DVOA, click here.) The Steelers have suffered through some terrible play by Ben Roethlisberger, but bad luck has also pushed them out of the playoff race.

Oakland, meanwhile, now has two wins in a row thanks to a defense that is developing into a very solid unit. The names may not be known to many around the country, but if the offense ever gets it together, wins and name recognition will not be far behind.

The Oakland offense is far from getting it together at this point. It is an undeniable fact that the better team did not win on Sunday. The Steelers outgained the Raiders 360-98. The Raiders won simply because they were fortunate enough to return two interceptions for touchdowns, including one from 100 yards out. Interceptions are a credit to the defense, but returning them for touchdowns is often a result of luck, not a repeatable skill.

Pittsburgh was only in a position to lose on those plays because the team made a mistake in starting Ben Roethlisberger. Roethlisberger threw four interceptions one week after sustaining a concussion. All four were ill-advised throws, but the worst was a forced throw at the goal line into triple coverage that Chris Carr returned for a touchdown.

Whether it was lack of practice or lasting effects of the concussion, Roethlisberger was well below 100 percent. His successful plays consisted of hitting his first read or dumping off to a safety valve. If his first read was covered, he often stood in the pocket helplessly or forced an ill-advised throw. He was sacked five times by a team that had totaled nine sacks in the previous six games.

Pittsburgh should have won the game anyway, but it would be an injustice to ignore Oakland's fine defensive performance. The Oakland offense managed only 17 net passing yards -- and yet the Raiders would have been in this game even absent the touchdown returns. All four of Roethlisberger's interceptions came on third or fourth down. The Steelers would have only been in position for one additional field goal had those interceptions instead fallen harmlessly to the ground.

This performance was the best of the year for the Raiders defense, but it has quietly developed into a very solid unit. The Raiders now rank 10th in defensive DVOA. The defense has given up only 35 total points over the past three weeks.

This defensive improvement is a surprise to many, but it is based on consistent development and should not have been unexpected. The 2004 Raiders were a team with a terrible defense and an average offense. That off-season, they signed LaMont Jordan and traded for Randy Moss. All the preseason moves got their offense a great deal of attention, but for a variety of reasons, it never clicked. The defense quietly improved to mediocre, a fact lost during a 5-11 flameout.

Art Shell (or at least Al Davis) apparently noticed the improvement and kept Rob Ryan on staff as defensive coordinator. He may not be matching the success of his father's feared Bears units of the 80's, but his defense is becoming a quality unit. This mostly anonymous unit -- quick, name four Oakland defenders; I'll even spot you Warren Sapp and Derrick Burgess -- is extremely young. Sapp is the only starter over 28, and they start two rookies.

On Sunday, Burgess continued to be the best pass rusher in football. He added 2.5 sacks to his season total and now has 22.5 sacks in his past 20 games. He played some left defensive end and easily got around Max Starks on multiple occasions. Constant pressure up the middle from Tommy Kelly and, yes, even occasionally Sapp, made it a long day for Roethlisberger.

Sunday also highlighted the play of a pair of linebackers, Kirk Morrison and Thomas Howard. Both players often sold out on the run and helped keep Willie Parker in check. Howard also made the pass deflection that led to Roethlisberger's second interception and the Raiders' first touchdown.

Even the secondary, a mostly inexperienced group, held their own. Safeties Stuart Schweigert and Michael Huff were often playing near the line of scrimmage, but young cornerbacks Fabian Washington and Nnamdi Asomugha managed without the support. Both came away with interceptions, and Washington in particular seemed to control his half of the field.

The Raiders continue to struggle because the offense is ill-prepared to help the defense. Andrew Walter remains an unknown quantity because he has no time in the pocket and no consistent receivers. The acquisition of Moss has to be counted as a bust. Pro Football Prospectus 2005 showed that Moss had performed much better on turf than grass over his career, but nobody could have foreseen such average production. Moss has averaged fewer than four catches and 60 yards per game as a Raider. With the trade of Doug Gabriel and the continuing Jerry Porter situation, Moss is on his own, and the results are ugly. Porter returned to the field on Sunday but was a definite non-factor.

The core problem is the offensive line. The Steelers pass rush just overwhelmed it from all directions, totaling six sacks. The offensive line did get a little push in the running game for Justin Fargas on a field goal drive in the second half. The development of Fargas bears watching, but a handful of successful plays in an otherwise desultory performance is not particularly encouraging.

The inept offense leaves the defense on the field a great deal. As a result, the defense is likely to give up more points going forward. Nonetheless, while nobody should confuse the Raiders defense with the Ravens, this improvement is legitimate and likely to be sustained. Oakland fans have something to enjoy for the first time since 2002.

Steelers fans are in no mood to enjoy this season of lost opportunity. The defending Super Bowl Champions are poised to be the third team this decade to follow up a Super Bowl win by missing the playoffs, largely due to the poor play of Roethlisberger.

Roethlisberger's life the last four months is worthy of a bad soap opera. In June, he got in a major motorcycle accident. He recovered amazingly, but just before the start of the regular season he had to undergo an appendectomy. After struggling early, he finally started playing well only to suffer a concussion. Fans in Cincinnati probably think this is karmic justice, but I think we can all agree it is a terrible run of luck.

Roethlisberger has played through the injuries for the most part, missing only the season opener and the end of the Week 7 affair against Atlanta. His play has been dreadful at times. The resulting struggles of the Steelers are not surprising, but they highlight his importance to the team.

Pittsburgh's identity under Bill Cowher has always been defense and running the ball. Over the past two years, while this mindset has not changed, the Steelers have actually put points on the board through the air. Their strong commitment to the run, and the resulting ability to consistently gain positive yardage, forces other teams to commit extra players to defend it. When that happens, Roethlisberger has been destroying opposing defenses. In both 2004 and 2005, Roethlisberger ranked third among quarterbacks in DVOA, which measures success on a per-play basis.

This year, he ranks 19th despite most of the same offensive personnel as a year ago. He played incredibly well against Kansas City and Atlanta, two teams that are a combined 9-5. In his other four games, he has one touchdown, 11 interceptions, and 15 sacks. That level of play makes Steelers fans yearn for the days of Kordell Stewart and Tommy Maddox.

What makes the situation even more sad is that the Steelers have a more-than-competent backup in Charlie Batch. In Batch's very limited time this year (37 pass attempts) he has a higher DVOA than Peyton Manning. He roughly matched Roethlisberger's production during his opportunity a season ago. Batch is an inferior player to a fully-healthy Roethlisberger, but he should be playing when Roethlisberger is struggling with his injuries.

Roethlisberger has been bad, but Football Outsiders DVOA rankings still have Pittsburgh as the 11th best team in football. The reasons for this are manifold. The ranking is somewhat skewed by the Kansas City game. The great play there only earned them one win, but those plays count across the season for DVOA. A team capable of that kind of performance almost always is a good team.

Second, the 2-5 record is the result of bad luck on a game-to-game basis. They outplayed the Raiders last Sunday. Against Atlanta, they played the Falcons to a draw on the road but lost the coin-flip in overtime. Against Cincinnati, they outgained the Bengals by over 100 yards, but Cincinnati fell on five of the six fumbles in the game. Quite simply, the Steelers should be at least 4-3 despite Roethlisberger's struggles. (Also note that the Steelers are 0-4 on the road, but 2-1 at home.)

One more hidden reason for the bad record is poor luck on special teams. The Steelers special teams have struggled in general this year, but their already suspect units have not been helped by happenstance. A team cannot, absent blocked kicks, control an opposing team's field goal attempts and punting yards. No team can control the opposition's kickoff yards.

Sometimes this is obvious, such as Philadelphia's loss on an incredible field goal by Matt Bryant. But most of these meaningful yards end up hidden from view. At Football Outsiders, we convert special teams plays into points. Based solely on things outside of Pittsburgh's control, opposing special teams have cost the Steelers 11 points. That ranks third in the NFL.

For perspective, despite their record the Steelers have outscored their opponents on the season 157-145. If you increased this 12-point advantage to 23 (adding in the 11 points of special teams bad luck), they would have the 11th largest point differential in football -- the same ranking as DVOA. On Sunday, Shane Lechler averaged over 48 yards per punt. Sebastian Janikowski hit both his field goals and had two touchbacks among his four kickoffs. Those little details matter in the result of a given game but tell us nothing about the ability of the opponent.

None of this means the Steelers do not have problems that extend beyond Roethlisberger. The special teams that the Steelers do control are among the worst in football. They have not adequately replaced Jerome Bettis, who provided real value on the field a year ago. The offensive line has struggled at times and almost never been dominant.

This is still a good team, and if Roethlisberger gets everything working, they are a very good team, but the schedule is too daunting for them to return to the playoffs. Six of their nine remaining games are against teams with winning records. For now, Cowher needs to evaluate which quarterback gives them the best chance to win each week. If the right quarterback is playing, they will quietly climb toward a respectable record. Playing spoiler is not a satisfactory role a year after being champions, but letting a banged-up Roethlisberger cost them several games leaves them no other choice.

The Raiders are likely basking in the glow of having played spoiler. Two straight wins are encouraging heading into a Monday Night affair next week. Take the chance in an otherwise ho-hum affair between the Raiders and Seahawks to get a peek at one of the game's up-and-coming defenses. Just don't forget to cover your eyes when the Raiders offense is on the field.

Best and Worst Luck in "Hidden" Special Teams
Team Kick
Kick Net
Punt Net
FG vs. Spec Tms
Spec Tms
STL -7.5 -12.5 3.3 -4.1 -0.6 3.2 -9.9 -13.0
GB -6.1 -9.4 3.5 1.8 -6.5 -3.0 -14.6 -11.6
PIT 0.5 -4.0 -7.0 -11.0 -2.5 -12.4 -23.4 -11.0
DEN -0.4 -6.9 -2.8 -4.4 -2.9 4.9 -6.1 -11.0
MIA -2.0 -5.3 2.0 -2.6 -2.2 -3.1 -13.3 -10.1
HOU -4.5 -8.0 1.2 -0.3 -5.1 0.7 -9.4 -10.1
NO -2.9 -6.5 6.1 11.7 0.9 14.8 17.6 2.9
KC -2.0 -4.9 5.5 8.4 4.2 8.4 12.7 4.3
IND -2.5 -6.3 5.7 6.2 8.0 0.2 4.9 4.7
CHI -1.3 -1.9 14.1 16.9 4.8 37.6 44.6 7.0
TEN -2.3 -5.7 1.6 4.3 8.1 8.0 15.3 7.3
These numbers are not weather-adjusted.
Each team's own field goals, punts, and kickoffs are not listed, but are included in totals

Each Tuesday in Any Given Sunday, Ned Macey looks at the most surprising result of the previous weekend. The NFL sells itself on the idea that any team can win any given game, but we use these surprises as a tool to explore what trends and subtle aspects of each team are revealed in a single game.

Posted by: Ned Macey on 31 Oct 2006

87 comments, Last at 15 Nov 2006, 1:09pm by Old Whippersnapper


by young curmudgeon (not verified) :: Tue, 10/31/2006 - 2:16pm

This article strikes me as fair. I have been saying that Roethlisberger has been playing like a guy that wasn't 100%. In addition, Ward and Polamalu have shown the effects of injuries. In today's NFL, if three of your key guys are not playing up to par, you're going to suffer. I understand the need to let Roethlisberger play himself back into shape, but it really looks like he could use some time off--I also wonder if Pittsburgh didn't underestimate the Raiders' defense and think that playing them might be sort of the equivalent of "some time off." As this article shows, it certainly didn't turn out that way. I think the Steelers are a good team having a bad season--this article reinforces that, and does a good job of suggesting some of the reasons why. (Hope springs eternal department--I have Roethlisberger penciled in as "comeback player of the year" next season.)

by B (not verified) :: Tue, 10/31/2006 - 2:20pm

If Ben Rothlesberger is living in a bad soap opera, when does his evil twin make an appearance? Or is he already here, and Ben has been locked in a closet somewhere since sometime after the superbowl. And of course his twin is played by "Brian St. Pierre" and can be recognized by his goatee.

by Ilanin (not verified) :: Tue, 10/31/2006 - 2:28pm

Great article, Ned, but one thing was missing - you didn't mention Casey Hampton's absence through injury. More than Polamalu or the LB corps, it is the behemoth NT that makes Pittsburgh's 3-4 tick, especially in the running game, which is where Oakland had moderate success. If he'd been able to stay in against Atalanta, the Steelers would probably have stopped Vick short on the OT drive, and the Raiders O would have been even more pitiful here.

Would they have won anyway? Maybe. But if Fargas hadn't sustained a couple of Oakland drives then the Steelers have the ball and driving in Oakland territory with possibly a minute or more left on the clock.

by Independent George (not verified) :: Tue, 10/31/2006 - 2:32pm

I think the best way of explaining the discrepancy between Pittsburgh's record and DVOA is this:

If it's week 17, and your team's playoff chances depended on the next game, who would you rather face: NYJ (4-4), KC (4-3), MIN (4-3), STL (5-2), or PIT (2-5)?

They might finish the season at 7-9 or 6-10, but does anybody in their right mind look at the schedule and think, "Oh, goody - we've got Pittsburgh this week"?

The same can be set of Philly (4-4), proving, once again, that God hates Pennsylvania.

by michael (not verified) :: Tue, 10/31/2006 - 2:33pm

evil twin? I think Ben already has a goatee, so you should be looking for his good twin.

by oljb (not verified) :: Tue, 10/31/2006 - 2:34pm

I'm waiting for the pendulum to swing back the other way. Get ready for the thirteen game winning streak, ending with Super Bowl XLI for the Steelers.

by Independent George (not verified) :: Tue, 10/31/2006 - 2:35pm

ack, 'said', not 'set'. Jeebus.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 2:37pm

6 The pendulum is swinging back..that streak happened last year... The steelers have a couple more losses to come.

I think this highlights something that Cowher did wrong, and (???) Herm Edwards did right: If you've got a competent backup, and your starter is too hurt to play competently (which Ben has been atleast 2 games), you go with the backup, not the starter.

by chris clark (not verified) :: Tue, 10/31/2006 - 2:38pm

From the article: Sometimes this is obvious, such as Philadelphia’s loss on an incredible field goal by Matt Bryant. But most of these meaningful yards end up hidden from view. At Football Outsiders, we convert special teams plays into points. Based solely on things outside of Pittsburgh’s control, opposing special teams have cost the Steelers 18 points. That is the highest total in the league.

So, do those things (outside Pittsburgh's control) count against Pittsburgh's ST DVOA? It makes sense for them to count toward the opposing teams' ST DVOA, because they are in the opposing teams DVOA, but for the Steelers they would seem like "luck" factors, i.e. just as fumble recoveries and interceptions returned for TDs are counted as random.

by Sean (not verified) :: Tue, 10/31/2006 - 2:38pm

This Steelers team reminds me somewhat of the 2003 Bucs. Tampa also came out playing at a high level but losing freakish games to Carolina and Indy. It seemed like it was just a matter of time until they right the ship, but instead they went the other way and started playing down to their record, until finally their DVOA accurately mirrored their season. I can see the same thing happening here, as Pittsburgh continues to be in a difficult division and has multiple tough matchups ahead. Once a team knows they are out of contention, it's hard for them to bring the same level of effort. I can see the season really getting away from Cowher and Co. pretty soon.

by admin :: Tue, 10/31/2006 - 3:14pm

Re: 9. It should say 11 points now. It does say 11, right? 18 was a miscalculation. And yes, that's right, kickoff and punt distance, as well as field goals, count only for the kicking team. Return teams are judged on returns alone.

by Brooklyn Bengal (not verified) :: Tue, 10/31/2006 - 3:48pm

In the past two years, the Steelers have traded Plaxico Burress to the Giants and Antwaan Randle El to the Redskins. Jerome Bettis retired and Hines Ward is playing like he's a 30 year-old wide-out who happens to be the team's only real receiving threat.

Maybe -- just maybe -- even with a healthy Roethlisburger, this team would not meet the DVOA's expectations. They SHOULD be 4-3? Come on. They COULD be 4-3. Hell, the Packers COULD be 4-3, too, were it not for a Brett Favre fumble here or a missed assignment there. But the Packers SHOULD be 3-4 just as the Steelers SHOULD be 2-5 based off of the way they have played under the given circumstances.

I love the DVOA rankings and think that they're a more accurate rating system than all those subjective "Power Rankings." But it drives me crazy that the system continues to favor teams like Pittsburgh and Philadelphia when those teams show a pattern of poor decision making.

Look at that wonderful stat called "turnovers." The Steelers are turning the ball over more than their opponents, so they SHOULD be losing. Fumble recoveries may be random over time, but consistently putting the ball on the ground and throwing interceptions will never win football games, no matter how many yards you gain.

And looking at the overall points for/against...well...I think that currently the DVOA is giving too much weight to the Steelers' ONE blowout win over those good-but-not-exactly-the-pinnacle-of-consistency Kansas City Chiefs. The Steelers' only other win came against a mediocre (at best) Miami Dolphis team in week one. One great performance and six average/below average performances so far this season. That KC game seems more and more like an abberration every week.

I think the Steelers will finish the season at or around .500 because they're a decent team, but the stats don't exactly paint the same picture that Mr. Nacy does.

by Sid (not verified) :: Tue, 10/31/2006 - 3:52pm

The key to Oakland is that they're just another West Coast team, in that they tend to play much better at home than on the road. That's why I picked them to cover each of the last two weeks (I had them beating Arizona straight up).
That and the fact that Roethlisberger has gone back to being terrible again. He finally showed signs of the old Ben last week, and then came the helmet-to-helmet hit.

by jj (not verified) :: Tue, 10/31/2006 - 4:18pm

Ben is/was/will be overrated. He was never more than the second coming of Kordell Stewart. He cherry picked his passes and waited quietly behind his O-line for a couple years.

Sorry, those days are over.

I don't care about DVOA on a guy that doesn't have to do much of anything but hand off a ball.

by Ryan Mc (not verified) :: Tue, 10/31/2006 - 4:24pm

Wow, you guys must have a lot of Steelers fans you need to keep happy.

"Against Atlanta, they played the Falcons to a draw on the road but lost the coin-flip in overtime." But only because Atlanta missed the potential game winning FG twice in the last minute. This is outside of Pittsburgh's control and therefore we must conclude that they were lucky not to lose in regular time.

"Against Cincinnati, they outgained the Bengals by over 100 yards, but Cincinnati fell on five of the six fumbles in the game." To some extent because Cincy got two short TDs in the 4th quarter after fumbles. In other words they didn't have to gain a lot of yards to score. Outgaining your opponent doesn't mean you were better, it can simply mean that your field position was generally worse.

I agree in principle that a team's record is not a perfect indicator of how good a team is, but you guys are treating it like it doesn't mean anything at all.

by Fnor (not verified) :: Tue, 10/31/2006 - 4:33pm

#15: Because, when evaluating teams, it doesn't. That's the whole point of... well... the FO stats. And this site. And most professional scouting.

Really, wins are extremely important- they keep you going towards the superbowl. But as a tool for evaluation, they're useless at best.

by calig23 (not verified) :: Tue, 10/31/2006 - 4:34pm


Actually watching some Steelers game over the last couple of seasons would be helpful.

And comparing Roethlisberger to Kordell Stewart is just about the stupidest thing I've ever heard.

by GlennW (not verified) :: Tue, 10/31/2006 - 4:36pm

I'd just stop trying to defend DVOA to people who want to believe that it's biased in favor of particular teams. I'm a Steelers fan and it's of absolutely no consolation to me that the team rates decently in DVOA. More than anything the turnovers have killed the Steelers, and yes, there's a lot of luck involved with them, but turnovers can and will kill your season in failing to even out over even a full 16-game schedule. The losses still count, and so be it. But if some people really want to believe things such as that a team can finish 15-1 in one season and win the Super Bowl in the next with a QB "that doesn’t have to do much of anything but hand off a ball", why waste your breath in arguing with them?

by AlexDL (not verified) :: Tue, 10/31/2006 - 4:47pm

Being the highest ranked / worst record team should be no consolation to any team. The way I see it here, the teams are measured on a wider base of information and are evaluated with that information. The football scourges of turnovers and other "luck induced" incidents are attempted to be pushed to the side when it comes to evaluating what the nature of the team is. ... Is it a team that can consistenly move the ball on offense? Is it a team that can consistently stop teams on defense? Does it's special teams have a positive or a negative effect on field position?
These are the questions that are attempted to be answered at Football Outsiders.

by chris clark (not verified) :: Tue, 10/31/2006 - 4:48pm

re 11: Thanks, Aaron. Then the DVOA stats are covering just what one would expect. It would be nice to have a complementary "LVOA" luck value over average that would "explain" (in the statistical sense of explain) the difference in the W-L record from what DVOA predicts. Didn't you do an article on that (luck) once?

by Ryan Mc (not verified) :: Tue, 10/31/2006 - 4:49pm

#16: right, but there's always the possibility that DVOA isn't perfect and isn't accounting for everything.

Isn't that why it was fine-tuned last year when the Colts were 10-0 and ranked 6th by FO? (ie. because a 10-0 record and #6 ranking simply didn't seem to go together)

Final question, when Pitt were ranked 5 last week and had a 2-4 record, did that also mean nothing? Because #5 ranking would suggest no way of losing to Oakland. 2-4 record would suggest that it was possible.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 4:56pm


This site increasingly dissapoints me. Not the site, but the commentators. Originally, DVOA was based on questioning everything, now it seems like everyone just accepts it as being correct, even when it obviously has trouble with things.

I still think there are plenty of aspects where we need to stop assuming things are luck, when theyre really not.

If fumbles by a RB are recovered 70% of the time by the offence, and a team is only doing it 10%, yeah, it could be luck. It could also be that the team has stupid lineman, who arent keeping their heads up. Problem? Theres no way to really tell, unless the trend stays that way.

As to Int return yardage being luck, thats a joke. If the ball gets picked off while a team is in either red zone, its run back for a TD a high percentage of the time. If teams aren't getting yards on those returns, maybe its not luck...maybe their corners arent that fast.

by Pat (not verified) :: Tue, 10/31/2006 - 4:59pm

but the schedule is too daunting for them to return to the playoffs. Six of their nine remaining games are against teams with winning records.

Carolina does not have a winning record. They're 4-4.

I still wouldn't be shocked if Pittsburgh ended up in the playoffs, but that's predicated on my belief that Baltimore's "hey we have an offense!" thing is extremely transitory.

by Pat (not verified) :: Tue, 10/31/2006 - 5:02pm

If fumbles by a RB are recovered 70% of the time by the offence, and a team is only doing it 10%, yeah, it could be luck. It could also be that the team has stupid lineman, who arent keeping their heads up. Problem? Theres no way to really tell, unless the trend stays that way.

Uh... so what would you suggest doing? "Maybe the team really is that bad... unfortunately, we'll never be able to tell." So... if you can't tell whether or not it's bad or not, wouldn't it make sense to regress it to the league average? What else could you do?

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:06pm


Whats best is to recognize that DVOA is missing something in cases like this, and realize that Pittsburgh maybe isnt as good as DVOA thought they were.

They most definitely aren't the 5th best team in the league, as DVOA thought they were last week.

by Pat (not verified) :: Tue, 10/31/2006 - 5:06pm

Because #5 ranking would suggest no way of losing to Oakland.

14-2 New England (ranked #3) lost to 6-10 Buffalo (ranked #22). By a score of 31-0.

14-2 New England (ranked #1) also lost to 4-12 Miami (ranked #23).

It happens.

by Fnor (not verified) :: Tue, 10/31/2006 - 5:10pm

#22: Uh, it's luck that the fumble happened in the red zone. It could have happened later on in the drive, or it could've happened after your own team's fumble.

Your thinking is fundamentally flawed for two reasons. First, you're saying that ints in the red zone are deterministically TDs, while at the same time talking about substantial probability. You can't have it both ways!

Secondly, you're completely ignoring the random aspect of it: field position. We can predict seasonal numbers for turnovers. We can't predict when those turnovers will happen. So if when it happens is random, how can you determine anything from it? Even aside from random bounces, what says that the interception happens near midfield on that same putative drive, rather than when the team is backed up? A confluence of playcalling, decision making and input from so many people calling it anything other than luck is rather laughable. So, yeah, RZ INTs are generally pick-6s, but the randomness is the fact that this pick happened in the red zone to begin with.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:12pm

Those are ass poor examples.

What are NE's DVOA for the first 5 games of that season where they lost to Buffalo 31-0 in the opener? They also lost to a poor Washington team in week 5. I'd bet that at the 6 week mark, their DVOA was ranked in the middle of league, not #3.

The Miami game was fluky, yes. It happened. It doesnt happen 3 or 4 times in the first 8 games though, and if it does, its no longer a fluke.

by Pat (not verified) :: Tue, 10/31/2006 - 5:15pm

Whats best is to recognize that DVOA is missing something in cases like this

Quoth I:

Remember, of course, that any statistical formula is not a replacement for your own judgment, just a tool to use in analyzing performance.

They most definitely aren’t the 5th best team in the league

1: Why? Because they lost to Oakland? Because upsets never happen?

2: They were also ranked 12th by estimated wins. On a play-by-play basis, they can look like the 5th best team in the league, but when it comes to doing the things that win you games, they're not doing so well.

You're getting far too wrapped up in "where a team is ranked." Football is far too complicated a sport to sum the entire thing up in a number from 1 to 32.

by Ryan Mc (not verified) :: Tue, 10/31/2006 - 5:17pm

re 26: notice how in your two examples the teams with the better records were ranked much higher than the teams with the worse records?

I'm not claiming that you never get upsets, but when you're 'upset' five times (OK, four, I guess San Diego was ranked higher than Pitt) in seven games you have to start wondering.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:18pm


Fnor, you're arguing against red zone picks costing more, and DVOA penalizes extra for turnovers in the red zone.

Saying its random that a pick occurs in the red zone is silly. If you're just going to assume that, you might as well just throw out the whole system.

Play in the redzone is given a weighting bonus by DVOA. Its considered more important.

Now, we can't see by play by play where the ball is being thrown to, but the fact that certain teams seem to consistently throw more pick-6s means something: It means theyre throwing more risky passes, IE outs to the RB/WRs on the sideline that are getting jumped. That tells us something. When those get picked off, theyre almost always 6 points.

by Will Allen (not verified) :: Tue, 10/31/2006 - 5:18pm

Rich, nobody ever claimed that DVOA was perfect, or that things weren't missed. However, if one is going to improve DVOA, or discount it in a non-arbitrary manner, one needs to have hard data with which to do so, and not just suppositions. I think it likely that the true nature of interception returns are still escaping us, but until we have a huge amount of detailed charting data to establish statistically, with a large sample size, that some qbs are throwing ints which are more likely to be returned for touchdowns, it is best to go with what one actually knows with a higher degree of certainty.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:19pm

"Remember, of course, that any statistical formula is not a replacement for your own judgment, just a tool to use in analyzing performance."

Exactly, because there are flaws. And Pittsburgh is clearly showing that it has serious flaws that DVOA can't see.

When you get upset 5 times in a row, theyre not upsets anymore, theyre losses, and you're a poor team.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:21pm


Will, I do have hard data. Pittsburg is 2-5.

If they were 4-3, and ranked #5, we could suppose that its a couple of unlucky bounces.

But theyre not, theyre consistenly getting beat by AWFUL teams. That makes them not so good.

by GlennW (not verified) :: Tue, 10/31/2006 - 5:23pm

First of all, Pittsburgh is no longer #5 (#6 actually) in DVOA. This article claims that they're now #11, which isn't completely ridiculous.

My newspaper this morning also says that the Steelers are 3-point favorites against Denver this Sunday. That's based on hard bettor preferences-- if you're in violent disagreement over the quality of these teams as applied going forward, this is one hell of an opportunity to load up on the "better" team. But some common sense is obviously in order here-- Pittsburgh simply isn't the average 2-5 team. We can split hairs over the numbers, but they're at least in the top half of the league, imo.

by Fnor (not verified) :: Tue, 10/31/2006 - 5:23pm

I see a huge disconnect in understanding in this thread. People are saying that the steelers aren't as good as their DVOA because they're losing games because of mistakes.

They're penalized for that. The point is that the effect those mistakes have are at least semi-random. That the bad thing happened can and should affect their record. It tells us nothing about how good the team is, however, because it's random! We can say it happened, but we can also say that it's not necissarily the team's fault.

Personally, I believe that all aspects of fumbles are random (including causation) and shouldn't be considered. That's neither here nor there, however, and I'm not going to talk about it until I've had a chance to discuss it (hopefully with Aaron).

by JJcruiser (not verified) :: Tue, 10/31/2006 - 5:26pm

I like the DVOA rankings as they are interesting even if they do not tell the whole story.

But I don't think they were meant to, right? The goal is not to say with mathematical precision what the best teams are -- that would be impossible. Too many variables that are not measurable: especially in-game coaching and planning, roster depth/injuries, emotion, etc. I don't buy that all fumble recoveries are totally random, for example, because these are professionals who train specifically to strip the ball and recover it. You can't tell me that plays NO role in who recovers the fumble. In another vein, there is simply no way to account for Cowher's bad decisions on when to start Roethlisberger, even though such things are a critical aspect of winning in the NFL. It's too subjective.

But I also don't think a lot of this criticism is fair: on whole it's a good system that tries to balance out a lot of complex factors to come up with formulas for evaluation. That's all. It's certainly better than College football's polls, which actually have an effect on the "Champion." And more interesting to me than "a poll of ESPN experts," or other power rankings that don't do as thorough of a job comparing statistics.

by Independent George (not verified) :: Tue, 10/31/2006 - 5:26pm

the upcoming games against DEN and NO (both at 5-2). But do you really think that Shannahan and Payton are looking at the schedule and smiling to themselves? Unless the players decide to quit on the season (a distinct possibility, I admit), I don't think anybody left on Pittsbugh's schedule is actually looking forward to playing them. Particularly if they don't start a literally brain-damaged QB for the rest of the season.

What last week's #5 ranking meant was that Pittsburgh was way, way better than Oakland - and, even with the loss, I think most of us would still stand by that assessment. What DVOA suggests - and I agree - is that even if Pittsburgh finishes at 7-9 or 6-10, they're likely to scare the heck out of a lot of teams on the way down. Or, in the parlance of our times, they were who we thought they were!

Just to be clear - I'm a Giants fan, and don't think I'm particularly biased for the Steelers. I just think that this remains a very, very dangerous football team, whatever their record says. (I also think DVOA give the Giants way too much credit, but that's another story entirely).

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:27pm


Right, theyre not the average 2-5 team, and I agree with that Glenn.

I just dont agree that they should be in the top 10. DVOA is seeing that theyre not good. I think they'll continue to drop. DVOA is just being slower than the conventional wisdom.

by NewsToTom (not verified) :: Tue, 10/31/2006 - 5:28pm

Who are the awful teams PIT is consisently losing to? Their losses have come against the 4-3 Jaguars, the 4-3 Bengals, the 5-2 Chargers, the 5-2 Falcons, and the 2-5 Raiders. Which of those first four teams do you believe to be 'awful'?

by Pat (not verified) :: Tue, 10/31/2006 - 5:28pm

The Miami game was fluky, yes. It happened. It doesnt happen 3 or 4 times in the first 8 games though, and if it does, its no longer a fluke.

Why does it matter when a fluke game happens as to when it can be discounted? Because it doesn't match up with the results from a team's record? Isn't it possible there are other ways to determine when a game should be called a "fluke"?

Here's a question: it's now about halfway through the season. You know the Steelers are 2-5. Without looking at their schedule, what do you think they're going to end up at? If you predict from their record, you'll get ~about 5-11.

If you predict from Pyth. wins (or estimated wins), you'll get about 7-9.

Now, go historically, and do the same thing after 7 games in a season. Estimated wins will do a far better job than just using the record will.

by mactbone (not verified) :: Tue, 10/31/2006 - 5:28pm

Re 33:
OK, so because one team is an outlier that means DVOA has serious flaws? You've made numerous suggestions, but everytime it boils down to you think whatever Pittsburgh does wrong conventionally isn't weighted enough or whatever NE does right isn't being weighted enough. Aaron has said before - and I've been around long enough to see him to respond to a lot of suggestions for improvements - that he's tried a lot of what you're asking but those suggestions don't improve DVOA as a whole. There are more improvements out there sure - but there isn't a quick fix that will get Pitt in line with perception without screwing with every other team.

I suspect Aaron will either be commenting on why Pitt's record is different from DVOA or Pitt's DVOA will nosedive and reflect what we think. I've seen DVOA and Aaron be right more than enough times to convince me that 90% of the time it's my perception that's wrong.

by jonnyblazin (not verified) :: Tue, 10/31/2006 - 5:30pm

re: 17
"And comparing Roethlisberger to Kordell Stewart is just about the stupidest thing I’ve ever heard."
OK, at the risk of incurring the wrath of Steeler fans, how about comparing him to Culpepper? Both have had seasons where they were statistically considered one of the most efficient in the league, but were under pretty sweet circumstances (Culpepper w/ great O-line play and Moss, Ben with dominant running game, O-line and also great D). Both are extremely tough to bring down, but they also hold the ball way to long and this has resulted in many injuries. Since the Steeler running game is no longer feared (Willie Parker is just an average back, at best), Roethlisberger has had to play QB in less than ideal circumstances, and it turns out he just isn't very good at it. He can certainly improve, but its also possible all those people saying Big Ben was overrated these past two years were correct.

by Joe (not verified) :: Tue, 10/31/2006 - 5:30pm

Really the Pittsburgh situation is showing the problem with all stats. Everybody's various slice and dicing of the stats is showing Pitt in or near the top 10 for both defense and offense, and yet their 2-5. When push comes to shove for the Steelers it's all turnovers, what happens after them, and where they're happening on the field. Cinci got 21 points off Pitt turnovers and Pitt lost at least 3 points due to those (counting the deep red zone INT as a FG if they keep the ball. Atlanta got 21 points off Pitt turnovers but I don't think any of them were in "garaunteed" point situations for Pitt. Oakland got 14 points costing Pitt at least 3 again.

That's a net reversal of at least 62 points in 3 games, that's the kind of red carpet losing seasons are built on. And I don't think there's any statistical analysis system that can compensate for that.

by Ryan Mc (not verified) :: Tue, 10/31/2006 - 5:31pm

re 36: I don't think there is a disconnect. (Some) people are simply sayng that randomness can't explain five losses. And remember, bad luck is only relevant if the game is close enough to make it relevant (ie. if you were really clearly better than your opponent then bad luck might mean you win by 10 points instead of 17; still a win) It simply isn't clear that the Steelers "should be at least 4-3" to some of us.

The only game I can buy that they 'should' have won is this most recent game, and considering that it was against an Oakland team that many expected to go 0-16 this season I don't think a moral victory is to Pittsburgh's credit.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:32pm

"Why does it matter when a fluke game happens as to when it can be discounted? Because it doesn’t match up with the results from a team’s record? Isn’t it possible there are other ways to determine when a game should be called a “fluke�?"

Because NE wasnt playing like a #1 team the first couple of weeks. They were playing poorly at that point. Look at those first 5 games, and NE wasnt playing that well. Losing to Buffalo wasnt that huge of an outlier at that point. Its only once they started destroying people later in the season that the buffalo game starts to look silly.

Again, if you want to use that argument, show me NE's and Buffalo's DVOAs after 4 or 5 weeks that season. I bet theyre nowhere near as far apart as Pitt and Oakland were.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:33pm

"I suspect Aaron will either be commenting on why Pitt’s record is different from DVOA or Pitt’s DVOA will nosedive and reflect what we think. I’ve seen DVOA and Aaron be right more than enough times to convince me that 90% of the time it’s my perception that’s wrong."

Pitt's DVOA nosediving would prove that DVOA WAS missing what was going on. If it nosedives, then Pitt was essentially a DVOA paper tiger in the beginning.

by GlennW (not verified) :: Tue, 10/31/2006 - 5:33pm

> Now, we can’t see by play by play where the ball is being thrown to, but the fact that certain teams seem to consistently throw more pick-6s means something: It means theyre throwing more risky passes, IE outs to the RB/WRs on the sideline that are getting jumped.

Unfortunately, as applied to your theory, you wouldn't have been able to classify either of Pittsburgh's INT-for-TDs last week as "risky" for TD returns. Both were passes closer to the middle of the field, and of medium distance downfield. The first INT was tipped by a leaping LB at full extension, and happened to find the right DB in the right (wrong?) place. Both INTs were marginal (the first) to poor (the final) decisions, but in both cases the return-for-TD aspect to the play amounted to dumb luck.

Throw in the fact that the sample size for INTs is small in the first place, and trying to rate a team's INTs for their "returnability" is just not going to get you very far.

by Raiderhatingisweaktit (not verified) :: Tue, 10/31/2006 - 5:35pm

#22: Similar point . I enjoy the site immensely as it disproves so many tired cliches of both mediots and casual fans. While I am by no means mathematically-inclined, the data is presented in a straight-forward manner and without bias. Facts are facts.
However, it appears the commentary (By FO and participants) can distort for different ends (factual vs. opinion). The "rankings" seem to be the biggest source of discussion/confusion. My interpretation is that the team ranked #1, #2, etc. is not necessarily the best team, but the team that executes best over the season. By charting as many of these aspects as possible, one can determine relative strength - not absolute. And, to your point, "luck" as an excuse may need to be revisited when it becomes a contention point (your fumble example). This site does a great job and its mods are continually upgrading the efficacy of the data. However, the vested fan (Pat, Steeler, or Raider) will tend toward bias in their verbiage when recounting the game. Are you gonna believe what you read or your lying eyes? Preparation, skill, and a dash of luck - such is the game. Raiders win this one.

by Pat (not verified) :: Tue, 10/31/2006 - 5:36pm

I just dont agree that they should be in the top 10. DVOA is seeing that theyre not good.

You could've moved Pittsburgh from #5 to #19 and still been within their variance. Hence the reason why they were ranked 12th in estimated wins.

Read here. Most of the criticisms that you're implicitly making (the fumble/interception bits are less important than the ones here) have already been made. That's why there's a second ranking, which shows the ability of a team to win, and Pittsburgh was ranked 12th there last week. And they'll probably be ranked lower this week.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:36pm

"The Steelers outgained the Raiders 360-98"

It bugs me that they use that as an indication that the steelers were the better team. How many times do they say that total yardage is a misleading stat.

How many yards did Joey Harrington throw for a couple weeks ago? 450?

Generally when teams outgain other teams like that, its because they had bad field position all game, and have been forced to throw all day... IE theyre losing.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 5:38pm


Glenn, I agree those two are dumb luck. But if Pittsburgh gives up 3 or 4 INT returns by lineman, is it still luck? No, it means theyre throwing the ball at lineman, which isnt good.

by GlennW (not verified) :: Tue, 10/31/2006 - 5:43pm

> OK, at the risk of incurring the wrath of Steeler fans, how about comparing him to Culpepper?

That comparison is not nearly so crazy as the one to Kordell Stewart, and is actually fairly reasonable. Under the right circumstances, Daunte Culpepper has posted at least three excellent seasons over his career and still has a career passer rating over 90. I want to believe that Ben is better because the Steelers' offense has not been as suited to a quality passing attack as the Vikings' was with Randy Moss in his prime, but I can't state that decisively. That comparison will have to play itself out...

by Sean D. (not verified) :: Tue, 10/31/2006 - 5:53pm

I've seen a couple people say that the Steelers can't be this, can't be that because they lost to the Raiders. But, the lost becuase of two bad plays. So basicall these people are saying this:

The Pittsburgh Steelers are clearly ranked too high because they lost to the Raiders. Judging a team by two bad plays during a 16 game season is way better than this. Plus, I prefer not to thunk and this site hurtz me head.

by GlennW (not verified) :: Tue, 10/31/2006 - 6:08pm

> The only game I can buy that they ’should’ have won is this most recent game, and considering that it was against an Oakland team that many expected to go 0-16 this season I don’t think a moral victory is to Pittsburgh’s credit.

It certainly isn't-- it's to the Steelers' eternal discredit. There still seems to be this sentiment here towards a system of "football justice", if you will, where a team's ranking must align with its record. I just don't think that's realistic. Look, I've watched every minute of the Steelers' seven games. They've been indisputably "outplayed", on the average-play, average-drive basis in only two of them-- against Jacksonville, and against San Diego. I'm fairly disgusted with this team, but that's not really the point.

The only statistical system that would properly account for what has happened to the Steelers this season as reflected in their record is one in which the 98% of mediocre-to-great plays is outweighed by the 2% of plays that are ridiculously stupid and/or unlucky (muffed punt returns with no defender within 10 yards, fumbled snaps, fumbles after minor contact, INTs for TDs, etc.). I think you could devise such a system where a team is severely penalized with "idiot points" in the extreme after a devastating play. I just don't think it would tell you all that much, and would certainly not be a good predictor of future performance. Today's sloppiest team in the league was yesterday's Super Bowl winner, in this case.

by Independent George (not verified) :: Tue, 10/31/2006 - 6:20pm

It's not that Pittsburgh backers are blindly following DVOA rankings, or that they're all just biased homers. My position is that: (1) aside from some flukey plays against good teams (and a couple flukey plays against a bad team), Pittsburgh has generally played pretty well. (2) a lot of those flukey plays come from a QB who has had one of the flukiest offseasons ever. I don't think Roethlisberger's regressing towards the mean so much as the fact that the two concussions have brought the mean much, much further down (so the Culpepper comparison actually sounds like a pretty good one).

I think what I'm saying is this - the Steelers with a brain-damaged Roethlisberger are a 4-12 team. The Steelers with a healthy Batch is a 9-7 team. The Steelers with a healthy Roethlisberger are an 11-5 team. Unfortunately for Steelers fans, in 7 games, they've had 1.5 healthy Roethlisbergers, 1.5 healthy Batches, and 4 brain-damaged Roethlisbergers. (And according to my flawlessly brilliant, objective, scientific calculations, that projects to an expected 2.21875 wins in the first 7 games.) If they go with Batch for the rest of the season, they might finish 7-9; if Roethlisberger can recover this season, 8-8 is within reach. And if he doesn't recover, but continues to start, they're looking at a top-5 draft pick.

by Rich Conley (not verified) :: Tue, 10/31/2006 - 6:30pm

I guess my whole issue is this: All these things that are considered random, are only really random if they dont repeatedly happen.

IE, if its just luck thats making the Steelers and Philly lose, then it should stop happening. (IE, the chance of them losing 5/6 fumbles is

by Rich Conley (not verified) :: Tue, 10/31/2006 - 6:33pm

losing 80% of your fumbles in a game is luck. If you continue to do it, its no longer luck.

Philly is losing because of long field goals... but we can't discount that they continually let opponents into fieldgoal territory at the end of the game, while only up a point or 2.

by Bionicman (not verified) :: Tue, 10/31/2006 - 6:39pm

>The Pittsburgh Steelers are clearly ranked too high because they lost to the Raiders. Judging a team by two bad plays during a 16 game season is way better than this. Plus, I prefer not to thunk and this site hurtz me head.

You aren't kidding when you admit that you prefer not to think, and if you said "I don't understand football," you would be right as well. Every week, there are several teams that commit two or more dumb plays or have bad luck go against them and still win. If the Steelers had played well for most of the game against the Raiders but still lost, you might have a point, but the fact is that the Steelers DIDN'T play well. The fact that they couldn't overcome 'two bad plays' (actually, there were four interceptions alone, not to mention a failed goal-line situation near the end) against a weak team is not a good sign.

To put it more simply, your post is an poorly thought-out strawman.

by GlennW (not verified) :: Tue, 10/31/2006 - 6:39pm

> IE, if its just luck thats making the Steelers and Philly lose, then it should stop happening. (IE, the chance of them losing 5/6 fumbles is

I don't disagree. I don't really know what the Eagles' basic problems are, but I think these events WILL stop occuring with the Steelers, at least at the ridiculous frequency that we've seen thus far (if they don't, it'll be about the damnedest thing I've ever seen, and I'll likely go crazy). However, with the injuries now mounting, there could be other problems. I don't think something like DVOA could ever hope to account for such factors, as well as the more motivational/emotional factors around a losing team, which I think is very real. Still, I'll bet this Steelers team finishes at 8-8.

by NewsToTom (not verified) :: Tue, 10/31/2006 - 6:42pm

Re #57
Unless they're really, really unlucky.

by GlennW (not verified) :: Tue, 10/31/2006 - 6:47pm

> The fact that they couldn’t overcome ‘two bad plays’ (actually, there were four interceptions alone, not to mention a failed goal-line situation near the end) against a weak team is not a good sign.

The "two bad plays" referred to here were the ones that resulted in 14 points for an opponent that had little chance of scoring that many of their own on offense, but your point is still a valid one, to an extent. Apparently just this one game dropped the Steelers from #6 to #11, which is appropriate. You're correct, it was not like the Steelers played well in this game overall; they were fairly ineffective on offense in their own right for at least three quarters.

by Jeff (not verified) :: Tue, 10/31/2006 - 6:57pm

"losing 80% of your fumbles in a game is luck. If you continue to do it, its no longer luck."

Yes. Let me know if this happens. It hasn't yet. Is DVOA supposed to predict that it will?

by Pat (not verified) :: Tue, 10/31/2006 - 6:58pm

losing 80% of your fumbles in a game is luck. If you continue to do it, its no longer luck.

I don't think you realize exactly how much small sample size affects things in this case. If recovering fumbles is a coin flip, if you lose 4/5 fumbles in your first 5, that'll happen 18% of the time - that is, it'll happen to 6 teams in the league.

Losing 4/5 fumbles twice happens about 3% of the time. Or, it'll happen to one team in the league. Each year.

by Brooklyn Bengal (not verified) :: Tue, 10/31/2006 - 6:59pm

RE: #27

You can't say that red-zone interceptions are completely random. I've seen Roethlisburger throw a couple of picks that could have been scores (v. Bengals and Raiders). That comes down to bad decision making. I give Kaesviharn credit for making a good read, but at the same time, Ben should not be throwing some of these passes.

This is similar to the terrible end-zone interceptions the Cowboys threw against the Giants; the plays called and the reads by the QBs were poor decisions. The teams are accountable for the way they play. Luck is certainly a factor, but consistent patterns display deeper problems.

RE: #29

You are dead on here. I think what some of the angrier readers are complaining about is mostly due to the editorial commentary. Football is very complex, and there are myriad reasons why the Steelers are losing. It is, however, unsatisfying for many readers (Steelers fans and haters alike) to see them not only ranked so high, but to read a commentary that says "The Steelers are GREAT" when it should say "Poor decision making and lazy mental errors are hurting a talented team" is a little disconcerting. Mr Macy does gloss over the Bettis issue and the terrible special teams, but that's just it...he glosses over it.

Still, overall I get my money's worth from FO. Keep up the good stuff!

by GlennW (not verified) :: Tue, 10/31/2006 - 7:09pm

> You can’t say that red-zone interceptions are completely random.

Maybe, but it's also a fact that behind the same QB(s) Pittsburgh did not suffer a single regular-season turnover of any kind in the redzone last year, which probably represents an element of luck working in the opposite direction. These are the kind of issues that we're up against in analyzing team performance. Maybe nothing is "completely" random, in the extreme, but some factors are obviously very random and therefore difficult to qualify.

by Will Allen (not verified) :: Tue, 10/31/2006 - 7:15pm

The problem, Rich, is thinking that a sample size of seven constitutes hard data. It doesn't. Now, if a NFL team played 162 games, and a team with a .290 winning percentage had the number 11 DVOA, THAT would indicate a serious problem with DVOA.

by B (not verified) :: Tue, 10/31/2006 - 7:20pm

66: It also reflects the fact that Pittsburgh no longer has Bettis to hand-off to in the red-zone, so they need to attempt more passes there, which is more chances for interceptions. That and their QB is completing a lower percentage of his passes.

by Oldcat (not verified) :: Tue, 10/31/2006 - 8:17pm

As a Bengal fan, I'd like to mention that the 'poor fumble luck' explanation is not as strong as you suggest, as in this game, two of the fumbles occurred on consecutive plays and were followed by an INT. You can only end a drive once!

Its also misleading to state unadjusted yardages in a game where the QB throws interceptions. The two pick-6 throws made Oakland lose 2 complete drive opportunities. Similarly in last years Cin-Pitt game in Heinz Ben threw for more yardage than Carson Palmer. But 3 ints and a KO return to the Pitt 3 yard line meant many of the Bengals drives were short ones. Its tough to build up 70 yards passing on a 3 yard drive.

In this year's game, a key to the game was the turnovers, and the fact that on the two 1 play TD drives in the second half the Steelers did not get pressure on Palmer, getting burned on 2 passes in less than a minute. If there was a time for them to 'dominate' the Bengals, that would have been it. Instead the defense seemed stunned and gave up quick scores.

by Kachunk (not verified) :: Tue, 10/31/2006 - 10:15pm

Perhaps turnovers aren't penalized heavily enough in DVOA. We keep seeing this explanation "such and such a team is rated higher than their record because they've had terrible luck with fumble recovery." After so many occurences of this perhaps DVOA needs to penalize each fumble more heavily.

by Pat (not verified) :: Tue, 10/31/2006 - 10:26pm

Or maybe, just maybe, Pittsburgh won't finish the season 4-12 or 5-11.

Does anyone really think that that's what Pittsburgh will finish at? Anyone?

by manning,e (not verified) :: Tue, 10/31/2006 - 10:30pm

i have to agree that ronde barber returning touchdowns on int's should not be considered luck. Some players do that and that has to be factored in. Anyone remember Neon Deon. Would you call his returns luck or a variable to be considered when playing against him, as well as when counting an int for a td as part of dvoa.

by Subrata Sircar (not verified) :: Tue, 10/31/2006 - 11:50pm

DVOA should correlate with wins. Ideally, it would correlate with wins over samples as small as seven games, but there is no statistic that really does so; I'm not surprised that DVOA misses a few things.

Essentially, what DVOA is saying is that if we played those first seven games an infinite number of times, with the same starters, Pittsburgh would end up 4-3 or better more than they'd end up 3-4 or worse.

It's also saying that going forward, the Steelers are more likely to play at a good level than a poor level.

I can see the former - every once in a while, a Raider will trip over the turf on his way to returning an INT to the house, for example, but I'm not sure I agree with the latter. The Steelers going forward have a number of issues which might stick with them (notably, the health of their QB). This is explicitly not considered in DVOA; it can only tell you what they're likely to do based on what they've done, and Big Ben's concussion is not yet fully evident (or fully played out).

by Justin Zeth (not verified) :: Wed, 11/01/2006 - 12:47am

I've watched all of the Steelers' games, and they're losing because of stupidity: Stupid coaching decisions, stupid QB decisions, stupid penalties, stupid special teams. Just sheer stupidity. The Steelers are a vastly talented team -- maybe the most talented in the NFL -- that's badly underachieving because they don't have their heads screwed on straight. And the head coach deserves a lot more blame for this than the quarterback.

It seems to my mostly uneducated eye that DVOA is being slow to downgrade the Steelers because it's having trouble distinguishing stupidity from random dumb luck. You can argue the Steelers have been unlucky if you want, and they have been a little. But luck isn't half the problem stupidity is.

This is supposed to be a veteran and allegedly well-coached team. They've shown every possible hallmark of bad coaching, all year.

Lest we forget, this is the team that had five AFC Championship games at home, was big favorites in all of them, and lost all but one (against a fluky 9-7 Colts team, on the last play). This is a team that's been killing itself with dumb penalties and dumb turnovers for years, and I highly doubt it's all the fault of (pick one: Neil O'Donnell, Kordell Stewart, Tommy Maddox, Ben Roethlisberger).

What we have here, folks, is a team that's been highly prone to stupidity for over a decade, that has always posted good regular-season records by consistently beating bad teams only to clam up against good ones, that had a team in 2005 so good that not even the typically bad coach could ruin it (and folks, the quarterback was reason #1 for that). And frankly, this is as messed-up as I've seen the Steelers since their very weird 2003 season, and it's starting to look like the coach, who, lest we forget, is not happy about his contract situation, is mailing it in.

There it is. Thoughts?

by Rob (not verified) :: Wed, 11/01/2006 - 1:46am

74: I don't think it's fair to say they've been 'prone to stupidity'. Every team is prone to stupidity. One or two stupid mistakes often decide the course of a game, and it can be as little as a receiver forgetting his route or a quarterback not seeing a blitzer coming at him. It takes a lot of luck to not make those stupid mistakes (or not to have them cost you the game), and thus to win a Superbowl, and the last Steelers team finally had the luck required to do that. But that's not damning them--they fielded a very good team, which yes, consistently beat bad teams (which is the mark of a good team: consistency, and beating the teams you should beat). It's also hard to say they 'clam up against good teams' when they went 15-1 only two years ago. Are you saying they only played one good team? I don't remember, but that seems unlikely.
All of that said, the current Steelers team is threatening my trust in DVOA to measure them. I'm not saying there is a systematic error in DVOA, at least not one which can be corrected easily. But one factor which we've glossed over, and which is not measured in DVOA, is the numerous injuries to the starting quarterback (and Polamalu, and Ward, and Hampton). I think that because we all saw what this team could do last year with health and good luck, we expect them to perform like that all the time. But their streak of wins was extremely unlikely. Also, if there is enough of a body of evidence to call any player in the league injury prone, it's Big Ben. Which is something to be considered, not just now, but in the future struggles of this team (competent backup or not, Batch will never be at the level of healthy Roethlisburger), and in the next couple years.

by jeff (not verified) :: Wed, 11/01/2006 - 2:35am

Let's talk about sample size and distribution for interception returns. What's the range of interceptions returns? About 0 to 100. (Though some even lose yardage.) What's the distribution of these returns? A fair number are clustered around 0-20 yards. But there's a reasonable number of longer ones, all the way up to 100. So it's a wide range of values, and fairly spread out along this range. My guess is the distribution is similar to punt returns, or kickoff returns.

How about sample size? The most interceptions thrown this year by a team is 12 with a low of 1. For KO returns, the low is 21 and a high of 50. Punt returns, 22 to 50. Can you really say anything with certainty of a sample size of 12 instances? If one goes 100 yards for a TD, is that random chance or is there some repeatable skill in that return?

How about the skill of the return itself. Do defenses work on being able to return int's for long gains? Not really. They work to get int's in the first place and stop the opposing offenses. Special teams - punt and kick return units - work on returning balls. If a defense returns an int for a TD is it because the returner found a seam? Or the defense set up a good wall for them? No. It was returned because the interceptor caught it in space, likely wasn't a D-lineman, didn't get tackled immediately by the receiver and his teammates gave him a few blocks maybe. But how much of this is actually controlled by the defense? Some? I suppose. All? No. Enough to determine from 20 returns? Unlikely.

Yes, some teams and players are better than others at returning interceptions. Is there anyway we can determine this amount with any certainty from 20 instances over the course of a year though? No. Does it have a substantial impact on the ability of the team? No.

An int return for a TD is the type of play that has a huge effect on any single game (a possible 14-point differential), yet has little to no bearing on the abilities of the two teams.

by Rick (not verified) :: Wed, 11/01/2006 - 8:50am

There's a bit of a trap that you don't want to fall into, and that is believing the statistic you are looking at more than other statistics. DVOA appears to be a good measure of "moves the ball well on offense, stops the ball moving on defense". And yes, that should obviously correlate well with winning. Also, depending on how DVOA accumulates from week to week, it seems that blowouts become disproportionately important.

I suspect that, if you look at a large number of games, when one team repeatedly gives up turnovers that lead into quick scores via long returns, you can see this team dominate time of possession and yardage accumulation (when looking at the offenses only) simply because their offense goes right back out on the field after each time it coughs up a turnover/TD. Thus it can be deceptive to simply say "Pittsburgh gained 360 yards to Oakland's 98 yards". In addition to 98 yards from the offense, Oakland got 125 yards from the defense and 106 yards from kickoff and punt returns. Pittsburgh only got 80 yards from their D and special teams. So the differential really was (so far) 440 yards to 319 yards. Now let's add in the penalties: Pittsburgh was penalized 6 times for 70 yards while Oakland was penalized 5 times for 45 yards.

When all that is added together, it's 370 yards to 274 yards. That's a differential of 96 yards. You would still expect Pittsburgh to be winning this game, but it's far from being as overwhelming as when you look at offensive stats alone. What happened? The last two drives of the game had Pittsburgh driving for 94 yards and getting no points out of it.

At some point, Pittsburgh needs to own up to gaining yards without gaining points. The discrepancy between the 2-5 record and their DVOA can be viewed as a function of their inability to convert yardage into points (KC game excluded).

Those are things to think about. Do I think Pittsburgh is "better" than a 2-5 team? Well, conditioned on a healthy QB, yeah, sure. Pittsburgh with a confused Roethlisberger in charge really is a 2-5 team, and that's what we saw against Oakland. I suspect he'll be back in form against Denver and even think they should win that game.

Finally, a few words regarding luck. The more any explanation relies on luck, the more we should be suspicious of it. As Bill Parcells famously said, "You are what you are". In science, randomness is never proven to be a cause, but only (roughly speaking) proven to not be a cause. If bad luck has really been holding the two PA teams back, we should see a reversal of fortune.

by Independent George (not verified) :: Wed, 11/01/2006 - 10:23am

#77 - So, are they who they are, or are they who we thought they were?

That's a question for the ages.

by B (not verified) :: Wed, 11/01/2006 - 11:54am

Looking at Pittsburgh's remaining schedule, I expect them to go 4-5 or 5-4, which is the performance you would expect from a .500 team. DVOA ranks them as a .500 team, so I feel it's accurate in this case.

by Pat (not verified) :: Wed, 11/01/2006 - 2:16pm

If bad luck has really been holding the two PA teams back, we should see a reversal of fortune.

Gambler's fallacy.

If luck has been holding the PA teams back, they'll play well over the second half of the season. They will not play well enough over to overcome the losses in the first half, and thus will end up, on average, with more losses than you would expect.

by Kyle (not verified) :: Wed, 11/01/2006 - 4:54pm

Anyone else amused at how repeatedly, how quickly, how valiantly, and how blindly Pat is willing to defend this website from any and all criticisms, ranging from the most blatant and uneducated troll post to the more well thought out questions of what DVOA could possibly be missing?

Its as if this person's football universe depends on DVOA being the universal truth, completely infallible, all-seeing and all-knowing. Yeesh.

by Pat (not verified) :: Wed, 11/01/2006 - 5:15pm

I've never even tried to defend DVOA against any criticism that I haven't seen before. Note the DB vs LB INT return yardage question, and you've never heard me mention anything about penalties.

But most of the criticisms that people have have been brought up before. Either that or they're cherrypicking.

by GlennW (not verified) :: Wed, 11/01/2006 - 5:17pm

> If luck has been holding the PA teams back, they’ll play well over the second half of the season.

Regarding the "gambler's fallacy", I took the original statement at face value and therefore equivalent to this one. Even 4-5 the rest of the way would represent a "reversal of fortune" but would still result in a lousy 6-10 finish.

The Parcellsian "you are what your record says you are" doesn't preclude improvement, or even some balancing out of more luck-based factors over 16 games, imo. I wouldn't read too much more into Parcells' logic other than that he's telling his team: "don't make excuses for losses". Parcells isn't so simple-minded as to believe that a team's record in a small sample size is an exact measure of what that team is or is capable of.

by Smartmonies (not verified) :: Fri, 11/03/2006 - 4:21am

People who are quick to criticize Roethlisberger, need to go back a few years to 2002. Which was the year after the Patriots and Tom Brady won their first SB. Listed Below are weeks 4 thru 8. The Patriots lost 4 games in a row and Brady was being written and talked about all over the country. The sentiment by most was that the Pats screwed up by letting Bledsoe go,while keeping Brady. A couple years later those same people are now saying Ben isn't no Tom Brady and Batch should replace him.

Here are those 4 games that I am referring too. Brady had 7 picks and 1 lost fumble. He also had an average passer below 50 thru those 4 games.

Brady had 2 interceptions and 1 lost fumble

Brady was 17-31 , 208 yards passing, 2 interceptions

Brady was 24-44 166 yards passing 3.6 yars per attempt, 3 interceptions

week 7 bye week

brady was 15-29 110 yds, 3.2 yards per pass

by steelberger1 (not verified) :: Fri, 11/03/2006 - 9:06am

RE 84:

I agree. Ben career is starting off in an eerily familiar way to Brady's. Ben just started earlier (rookie starter).

by dave whorton (not verified) :: Sat, 11/04/2006 - 11:52pm

i'm baffled by the ranking of the pitts def in the dvoa.their tied for 18th in total red zone scoring % their 3rd down def is 16 in % converted and yet this is a top 5 def.i know the def has been put in alot of bad situations but great def hold teams to fgs or cause turnovers in the red zone.

by Old Whippersnapper (not verified) :: Wed, 11/15/2006 - 1:09pm

So now we see the real Steelers emerge (albeit perhaps a bit too late to snare a playoff spot). Ben begins to have more games in the 100 rating area -- and by the way, jj, what an idiot you are for comparing him to Kordell. (Did Kordell ever become the #1 rated playoff passer?)

Last year the Steelers rode Ben's arm to the Super Bowl, so please, cool it with the "game manager who just hands off" crap.

The Steelers are in the top tier of stats in most of the key offensive/defensive categories, so please, stop whining about the consideration they're getting here. They're a butterfingered punt returner away from winning a couple games they lost -- which has nothing to do with the core team.

I'm sorry -- the loss to the Raiders was a JOKE (what, like 5 yards the Raiders had passing or whatever?). And so was the Cincy game -- thanks to Colclough, Carson Palmer escaped another of his patented big-game chokes (talk about your overrated QBs!).

The article was, and continues to be, an accurate assessment -- whether certain whiny Bengals fans like it or not.