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.

21 Nov 2006

Any Given Sunday: 49ers over Seahawks

by Ned Macey

(Ed. note: First, let's answer the inevitable question -- if the #1 team in DVOA and the last undefeated team lost this week, why aren't we writing about those games? Short answer: We've written plenty about Indy's deficiencies, all of which showed up in that loss, and an Eagles-Titans article would consist of one sentence, "Donovan McNabb tore his ACL and the Eagles are screwed.")

The day EA Sports announced Shaun Alexander was going to be on the cover of Madden 2007, it was clear that the Seahawks were going to have a difficult time repeating as NFC Champions. The Madden curse plus the Super Bowl Loser's Curse posed seemingly insurmountable odds. Those powerful curses have taken their best shot, with both Alexander and Matt Hasselbeck suffering injuries, but entering Sunday's game, the Seahawks had a 6-3 record and were on track for the playoffs.

Most people assumed Seattle would repeat despite these well-known curses. Any challenge was presumed to come from either their recent rival in St. Louis or the hotshot skill players assembled in Arizona. Instead, it is San Francisco that stands just one game behind the Seahawks following a surprising win.

This might be called an exciting contest between the two best teams in the NFC West. But the truth is that the two best teams in the NFC West are among the worst teams in football. For San Francisco, a lucky .500 record gives credibility to a rebuilding plan that is starting to take shape. For Seattle, a lucky 6-4 record means they will be only the second Super Bowl loser to make the playoffs since 1999, despite taking numerous steps back on the field.

The Seahawks currently rank 25th with a DVOA of -15.1%. Both the rank and the DVOA rating itself are the lowest of any Super Bowl Loser since 1999 except for the 2003 Raiders. The Seahawks were blessed with 11 games this season against teams that currently have a below average DVOA. That schedule and two close wins over St. Louis may allow them to make the playoffs despite the Curse's best efforts.

The most surprising feature of Sunday's game was watching little-known Frank Gore run wild while reigning MVP Shaun Alexander struggled in his first game back from injury. Gore has emerged as one of the most dynamic running backs in football, always capable of breaking a big run.

Alexander has struggled all season, even when supposedly healthy. Running backs tend to take a substantial step back after having 370 carries in a season. Add Alexander as another data point supporting that theory. He has missed six games and averaged only 2.7 yards per carry. Alexander was hesitant at times and had few running lanes. He showed some flashes of his old form, but they were few and far between.

A frequent justification for the Seahawks' struggles on the ground is the departure of guard Steve Hutchinson. Widely considered the best guard in football, Hutchinson was not given the franchise tag and received an unmatchable offer from the Vikings. The Seahawks have been unable to develop a consistent ground game without him.

This theory overstates the impact of the admittedly valuable Hutchinson. Guards battle with centers for being considered the least valuable position on the offense. If the absence of even the game's best has this sort of impact, Hutchinson and his fellow elite guards deserve to be paid like Peyton Manning. Minnesota has one of the worst rushing offense DVOA ratings with Hutchinson, so he is hardly a panacea.

The Seahawks miss Hutchinson, but other injuries have upset continuity along the line. Center Robbie Tobeck, left guard (and Hutchinson replacement) Pork Chop Womack, and right tackle Sean Locklear have all missed games. 2005 first-round pick Chris Spencer is a more than adequate replacement, but he is only one player. Different players at different positions each week do not allow the line to play consistently.

More importantly, both left tackle Walter Jones and fullback Mack Strong have played well below their own high standards. Jones is widely considered the best left tackle in football. On Sunday, he was beat for two sacks. Jones is 32 years old, the same age as Jonathan Ogden, who has already lost a step. The transition of the game's best left tackle into merely an above average one could have as large an impact as the departure of Hutchinson.

Jones did play impressively at times, which is more than can be said for Strong. The venerable fullback finally started to get his due last year, but it appears his age has made him a liability. Consistently this season he has struggled both in clearing holes for running backs and in pass protection.

On the game's most pivotal play, a fourth-and-1 attempt in the fourth quarter, both Strong and Jones struggled. The Seahawks predictably ran Alexander behind left tackle with Strong as the lead blocker. Strong's block on a defensive back was ineffective, while Jones was also beat. The two defenders brought Alexander down behind the line of scrimmage.

Of course, while the run game struggles, nobody is talking about the elephant in the room: Matt Hasselbeck's injury. Almost everyone considers the Eagles' season effectively over thanks to an injury to Donovan McNabb. Hasselbeck is the same caliber quarterback, and the Seahawks have tried to play four and a half games without him.

Seneca Wallace has impressed with his poise, but his overall production has been substandard. He threw three interceptions on Sunday. The first one was a bad throw, the second one terrible, and the third one atrocious. Things should improve with Hasselbeck's return this week. Deion Branch is getting more comfortable in the offense, and along with Darrell Jackson will provide Hasselbeck with an excellent pair of starting receivers.

The massive injuries to the offense make it a little hard to evaluate that unit going forward, but more troubling for the Seahawks is their poor defense. A year ago, the Seahawks had one of the league's best offenses and an average defense. With some young players maturing, the return of safety Ken Hamlin, and the signing of Julian Peterson, it was reasonable to assume the Seahawks would improve. Instead, their defense ranks 24th in DVOA, equally bad against the run and the pass.

Unlike the offense, injuries have not been nearly as big a problem, with defensive tackle Marcus Tubbs as the only injured player of note. The seemingly no-brainer signing of Peterson may actually have been a mistake. Peterson provides excellent pass rushing skills, but he struggles playing the run. The Seahawks already were capable of getting to the quarterback without him. They lacked an elite pass rusher, but their overall pressure ranked among the league's best. And when Peterson was on the field Sunday, the 49ers ran at him constantly throughout the game.

Hamlin's return from an injury is an inspiring story, but he may be struggling to recapture his old form. Gore had two long runs on the day. On the first, Hamlin was easily blocked by a tight end leaving the hole to Kelly Herndon, who missed the tackle. Hamlin and Peterson both whiffed on Gore on his second long run.

Gore's two runs highlighted an amazing day that set the 49ers single-game rushing record. Football Outsiders predicted big things for Gore -- even comparing him to Priest Holmes in Pro Football Prospectus 2005 -- but not many others saw him emerging as one of the top backs in football. Gore now ranks second in the league in rushing yardage despite a suspect quarterback and unheralded offensive line. The offensive line looked impressive on Sunday, particularly the left side with tackle Jonas Jennings and future Hall of Famer Larry Allen.

Gore is not the top-ranked running back according to Football Outsiders' stats in large part because he intersperses great runs with poor runs. 11 of Gore's 24 carries went for two yards or less. Fortunately, he had four carries of at least 20 yards, including two of 50 yards. Gore has gained at least 20 yards on an amazing 6.3 percent of his runs this season. Tiki Barber, Larry Johnson, and LaDainian Tomlinson all gain 20 yards on fewer than 3.3 percent of their carries. Edgerrin James has yet to gain 20 yards on a carry.

Other than consistency, the only thing holding Gore back is a propensity for fumbles. His fumble on Sunday came in the fourth quarter and could have been extremely costly. That was the sixth time he put the ball on the ground this year, as many as the three elite backs mentioned above combined.

Gore may have exploded the past two weeks, but he has been running well all season. The big difference in the 49ers' current three-game win streak is the suddenly solid play of its defense. They have allowed only 30 points during the stretch after allowing more than 30 points in five of their first seven games.

A certain amount of skepticism is reasonable, given that the opponents were Minnesota, Detroit, and the Wallace-led Seahawks. At the same time, this improvement is so great compared to their previous level that it is worth considering if it is permanent.

The major personnel switch was the insertion of Ronald Fields at tackle. This move looked effective against the Seahawks, who had almost no success running up the middle. Fields is bracketed with solid defensive linemen in Marques Douglas and the estimable Bryant Young. Young is a known quantity, but Douglas was extremely active and difficult to block. Rookie sixth-round pick Melvin Oliver rounded out the defensive line and had a solid performance.

The secondary still lacks depth, but the apparently ageless Walt Harris made a fine play on his first interception. Sammy Davis, on the other hand, struggled when he was brought into the game. Teams are likely to spread San Francisco out in the future and attack their backup cornerbacks. They were protected against Seattle thanks to deep safety help and multiple looks that confused Wallace.

The emergence of Gore and improvement of the defense are nice stories, but the future of the 49ers rests on the development of Alex Smith. One and a half years into his career, he grades an incomplete. On Sunday, he was protected by short throws, many coming off of rollouts or bootlegs. The game plan featured a heavy dose of passes to his running backs and tight ends. He did complete one long pass to Antonio Bryant that was called back because of a holding penalty, but he actually underthrew his receiver where a better throw would have led to a touchdown.

Smith currently is an average quarterback based on our advanced statistics. That ranking is a bit disappointing based on his status as the first overall pick, but it is encouraging if we consider how bad he was a year ago. The good news is that he is only 22 years old, actually younger than Matt Leinart or Vince Young. Smith has never had great physical talents, but watching him on Sunday, it was easy to picture him developing into a quarterback much like Hasselbeck.

Sunday's win was certainly an important one for the 49ers, the kind of positive reinforcement the Mike Nolan regime needs. In the end, however, they beat a bad team whose record did not reflect its true ability. The current Seahawks with Wallace at quarterback are one of the ten worst teams in football. Even with Hasselbeck healthy, the problems on defense and in the running game make them average at best. To date, this team has played one very good half of football, the first half against the Giants way back in Week 3.

The Seahawks seemingly have only two difficult games the rest of the season, against Denver and San Diego. But, weren't the Vikings and the 49ers supposed to be easy games? Those two teams have combined for one other win against a winning team, Minnesota's victory over the Steve Smith-less Panthers. The Seahawks are not a good team, and they will be lucky to finish 9-7. Sadly, that is likely good enough for them to win their division.

The 49ers' arrival at 5-5 allows them to take the mantle from the Jets as the worst .500 team in football. Road games at St. Louis and New Orleans the next two week will test whether or not the improved defense is for real. Four road games out of six likely spell doom for any playoff dreams the 49ers may harbor. They are a combined 2-11 away from San Francisco the past two seasons. But as long as the defense does not return to giving up 30 points a game, this season has to be considered successful.

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 21 Nov 2006

61 comments, Last at 23 Nov 2006, 1:18pm by Joel


by David (not verified) :: Tue, 11/21/2006 - 2:25pm

An Eagles-Titans article would consist of one sentence, “Donovan McNabb tore his ACL and the Eagles are screwed.�

I disagree; that game was bordering on lost-cause status before McNabb went out (who committed the game's only interception? Not Garcia.), and quarterbacking had nothing to do with a 90-yard punt return, a 70-yard rushing TD, or a horribly off-target shotgun snap that turned into a TD recovery. The Eagles' offense was bad without McNabb, and they may well have been able to mount a comeback if he was healthy, but it wasn't his absence that caused the upset. they are completely screwed, though.

Okay, go back to talking about the actual article now. Sorry.

by Charles the Philly Homer (not verified) :: Tue, 11/21/2006 - 2:29pm

When I took Gore in the 5th round of of an 8-team fantasy draft, everybody thought I was an idiot. Of course, now I'm the idiot in 1st place.

Any commentary on why Gore fumbles? I only saw the one in the Eagles game where he just got hit funny at the end zone. That one was a seemingly random event where a helmet wound up in exactly the right place for Eagles fans, but his growing fumble rate is either emerging from very consistent bad luck or something systemic. Maybe a California football watcher has some idea...

by Pacifist Viking (not verified) :: Tue, 11/21/2006 - 2:33pm

Is there a place on the FO site that discusses what is meant at FO by the words "lucky" and "random"? I've been reading most of the explanatory stuff, but haven't found this. I would like to know specifically what is meant by these terms, because my first instinct when seeing a team's record called "lucky" is to recoil in horror, but I don't think my reaction is correct and want to know more about what is the meaning.

by hector (not verified) :: Tue, 11/21/2006 - 2:37pm

Running backs tend to take a substantial step back after having 370 carries in a season.

For future (and past) reference, is that a regular-season threshold or would it include playoff work? And if it's the former, is there any evidence to suggest that going deep into the playoffs eventually comes at a cost for your RB, or is that not really a big deal?

This just in: Gore plays for the 49ers. I'm a Peter King reader so I know things like this.

Nitpicking, but even with the McNabb injury I think the Titans Week 11 win was the shocker of the weekend over any other result. But if the SF-SEA game had more angles to write about (and it seems like it did), no harm in taking the better story. Nice piece.

by admin :: Tue, 11/21/2006 - 2:40pm

Re: 3, it's split into various articles in various places and I do plan on putting it all together in one place at some point. Also, remember that Football Outsiders writers do not share a hive mind, and therefore one writer may use the word "lucky" in a different way than another writer.

Re: 1, sure, Ned could have written about the defensive and special teams problems, but with McNabb done for the season how much does it really matter?

by Sean (not verified) :: Tue, 11/21/2006 - 2:40pm

Yes, you could see Gore was going to have a huge year just by watching him in the preseason (or by looking at his numbers last year in limited action, for that matter). I've got Gore on all three of my fantasy teams, and it wasn't by accident.

by Charles the Philly Homer (not verified) :: Tue, 11/21/2006 - 2:47pm

I think the lucky/random dichotomy works like this:

"Lucky" is when you have a positive differential in non-predictive events. For example, ridiculous ref calls, fumbles, bizarre circus interceptions, etc. have gone your way. This compounds to the extent that your team's performance is many deviations away from the predictive mean of a statistic.

"Random" is a reference to an individual occurrence of said non-predictive events. As in "Wow, that fumble was really random. It went 20 yards into the end zone and was recovered by the Giants, securing another heartbreaking 2007 Eagles loss."

by centrifuge (not verified) :: Tue, 11/21/2006 - 2:50pm

3: In FO terms, certain game factors are "random" -- fumble recoveries and return yardage on turnovers are the two biggest ones. "Random" here means that a team's performance in that field is nonrepeatable. A team can be great one weak, terrible the next, and great again further down the line, and probably will be, because there's little rhyme or reason to those performances.

A team is "lucky" when it does better than the league average in game factors that are out of its control. This includes doing well in the random factors above, but also certain plays that are solely in the hands of the other team, like opposing FG percentage and kickoff distance.

by Ned Macey :: Tue, 11/21/2006 - 2:57pm

Re Lucky: In the context of this article, I meant lucky only in the sense that the level of play on the field is not indicative of teams that are 6-4 and 5-5. Much of that is schedule.

The point is simply that according to DVOA they've both been among the ten worst teams in football. So, I consider it "lucky" that their records are what they are.

by Charles the Philly Homer (not verified) :: Tue, 11/21/2006 - 2:59pm


Random events aren't necessarily unrepeatable. I think it's more that the event has no predictive power in a metric. Meaning, if you played the game 10,000 times in a regression test, you would never be able to anticipate the frequency of the event.

by Pacifist Viking (not verified) :: Tue, 11/21/2006 - 3:05pm

Thanks, I was just curious if, coming from a language/literature/philosophy background, my use of the term is different than somebody coming from a mathematics/statistics viewpoint.

Ned, that makes sense.

by centrifuge (not verified) :: Tue, 11/21/2006 - 3:08pm

10: That is a much better way of saying what I was awkwardly trying to.

by navin (not verified) :: Tue, 11/21/2006 - 3:16pm

I agree that San Fran's DVOA shows that they should be terrible, but the expected wins actually has them at the top of the 10 worst teams (because of the extremely high variance). Outside of their blowout losses to Chicago, KC, and San Diego, they have played mediocre, which is a good step from what they were last year.

by Brooklyn Bengal (not verified) :: Tue, 11/21/2006 - 3:29pm

RE: #3

It's my experience here at FO lately that writers for the site and posters to the site both use the word luck as a way to ignore glaringly obvious negative statistics while favoring positive statistics for likeable players/teams.

For example: The Steelers have been extremely unlucky this year = Ben Roethlisburger has thrown 12 touchdown passes compared to 17 interceptions while former Buckeye saint Santonio Holmes drops more balls than he does domestic violence charges.

Or: The Philadelphia Eagles have been extremely unlucky this year = despite playing at an extremely high level on the majority of individual plays, the Eagles coaching staff is terrible at clock management, McNabb is prone to making poor decisions under pressure, the offense has a bad habit of turning the ball over and the defense gives up as many first downs as the offense makes. Also, that 60 yard Tampa Bay field goal as time expired really was incredibly lucky.

As for use of the word random, I think that the FO writers are consistent in their usage of the word pertaining to distribution of events.

For example, fumble recovery is random because over time, all teams recover 50% of fumbles. If the Atlanta Falcons recover 90% of their fumbles in the first 8 games of the season, they are likely to recover only 10% of their fumbles in the second half of the season.

However, a lot of us posters will say things like "turnovers are random" which is patently false. Turnovers are not random. Many teams have a propensity for interceptions and many players have a propensity to fumble, etc. Where and when those turnovers occur in a game are often randomly distributed.

Hope that helps!

by Andrew (not verified) :: Tue, 11/21/2006 - 3:33pm

David #1:

I disagree; that game was bordering on lost-cause status before McNabb went out

A game where you are trailing 7-3 on the second play of the 2nd quarter and driving in opponent territory is a lost cause? What football-free planet do you inhabit?

quarterbacking had nothing to do with a 90-yard punt return,

With McNabb in, the Eagles might not have been punting there, so there is no way of knowing that.

a 70-yard rushing TD


or a horribly off-target shotgun snap that turned into a TD recovery

McNabb is something like 6" taller than Garcia. I doubt the ball would have been over McNabb's head. This was a failure of Jackson to adjust his mechanics for the little man in at QB.

by centrifuge (not verified) :: Tue, 11/21/2006 - 3:35pm

For example: The Steelers have been extremely unlucky this year = Ben Roethlisburger has thrown 12 touchdown passes compared to 17 interceptions while former Buckeye saint Santonio Holmes drops more balls than he does domestic violence charges.

I think the Steelers' unluck had more to do with having recovered very, very, very few fumbles and having several INTs returned for touchdowns than having fumbled or been picked off in the first place. There's also the matter of losing the overtime coin flip to Atlanta, which is the very definition of bad luck. (Depending on the writer, they may also throw Roethlisberger's injuries under that heading as well, but that's another story.)

The Steelers have been recovering tons of fumbles the past two weeks. Coincidentally, they won both games. Hmm.

by Wanker79 (not verified) :: Tue, 11/21/2006 - 3:36pm

"Edgerrin James has yet to gain 20 yards on a carry."


by putnamp (not verified) :: Tue, 11/21/2006 - 3:42pm

More importantly, both left tackle Walter Jones and fullback Mack Strong have played well below their own high standards. Jones is widely considered the best left tackle in football. On Sunday, he was beat for two sacks. Jones is 32 years old, the same age as Jonathan Ogden, who has already lost a step.

Strong's been depressingly bad. Jones, however, has been playing on significant injuries all year. If he's lost a step, it's because he is physically hindered, not age (yet, anyway).

Seneca Wallace has impressed with his poise, but his overall production has been substandard. He threw three interceptions on Sunday. The first one was a bad throw, the second one terrible, and the third one atrocious.

You're a lot kinder than I am. The first one was into triple coverage, the defender on the second one had moved to get in front of the receiver before the pass had even left, and the third one.. I'm hoping somebody at some point can tell me who it was to, because there was nobody in a white uniform anywhere near that pass.

by jetsgrumbler (not verified) :: Tue, 11/21/2006 - 3:45pm

how is that my JETS manage to get slammed in an article about a game between to teams in the other conference? i prefer to call them the most improved team, rather than worst 500 team.

seahwawks may have been bad so far, but i can't imagine them not starting to gel as playoffs approach

by TomS (not verified) :: Tue, 11/21/2006 - 3:50pm

I just have one observation...

...it is all HawkSquatch's fault!


by Charles the Philly Homer (not verified) :: Tue, 11/21/2006 - 3:52pm

I think some sort of study should be done given Edgerrin James's huge performance decrease. How far back do your offensive line charts go, and could we compare similar free agent deals? That might have some real use to personnel guys making decisions about big-money running backs.

Basically, I want to know exactly how related offensive line performance is to running back performance. The hardest part would be determining which basket of offensive line data generated the most useful statistic without favoring the conventional wisdom too heavily...

by zenbitz (not verified) :: Tue, 11/21/2006 - 3:52pm

comments on 49ers also starting Brandon Moore at LB instead of Ulbrich and Mark Roman at SS instead of Parrish?

These guys (particularly) Moore seem to get all the press rather than Fields.

by dbt (not verified) :: Tue, 11/21/2006 - 3:55pm

#19: Teams below them in raw VOA (by last week's stats since this week's aren't out yet): Lions, Redskins, Bills, Browns, Texans, Titans, 49ers, Cardinals, Raiders, Buccaneers.

That statement seems exactly right to me.

by Andrew (not verified) :: Tue, 11/21/2006 - 3:56pm

Brooklyn Bengal #15:

For example, fumble recovery is random because over time, all teams recover 50% of fumbles. If the Atlanta Falcons recover 90% of their fumbles in the first 8 games of the season, they are likely to recover only 10% of their fumbles in the second half of the season.


Even if you flipped heads 9 out of 10 times with a fair coin, you are still likely to flip 5 heads in your next 10 flips.

A team that has recovered 90% of fumbles through 8 games is likely to recover 50% in the rest of the season.

by zenbitz (not verified) :: Tue, 11/21/2006 - 3:56pm

If the Atlanta Falcons recover 90% of their fumbles in the first 8 games of the season, they are likely to recover only 10% of their fumbles in the second half of the season.

Uh... no. They are still likely to recover 50% in the second half.

by DavidH (not verified) :: Tue, 11/21/2006 - 3:57pm

For example, fumble recovery is random because over time, all teams recover 50% of fumbles. If the Atlanta Falcons recover 90% of their fumbles in the first 8 games of the season, they are likely to recover only 10% of their fumbles in the second half of the season.

You've got the concept right, but the execution is wrong. If the Falcons recover 90% of their fumbles in the first 8 games, they are still likely to recover 50% of their fumbles in the second half. The football doesn't have a memory.

The point is that even 50% is drastically lower than 90%, and that will affect their results.

by DavidH (not verified) :: Tue, 11/21/2006 - 3:57pm

Haha, dogpile of the gambler's fallacy

by DavidH (not verified) :: Tue, 11/21/2006 - 3:57pm

ON, not of

by Thok (not verified) :: Tue, 11/21/2006 - 4:08pm

One thing to note about the Niners is that their improvement came when their schedule eased up. They had a five game stretch where they ran into 3 of the top 5 teams in DVOA and Kansas City on the road, with a breather at Oakland. They went 1-4 in that stretch, all of the losses being blowouts.

Basically, to assess the Niners properly, you need to decide if they are underperforming on the elite teams, or overperforming on the rest of the league. The latter is probably true, but it could be the former.

by Wanker79 (not verified) :: Tue, 11/21/2006 - 4:14pm

Re: 29

I don't think anyone is suggesting that the 49'ers are a good team. What people are saying is that compared to last year, they are playing remarkably well. That's a big difference.

by David (not verified) :: Tue, 11/21/2006 - 4:16pm

Bah, an awful article. No insight whatsoever into why the niners won this game, or what they have done to overhaul last year's win total with six games to go.

Sum total of analysis, paraphrased, this isn't really a shock, because Seattle actually suck, and so do the niners.


How about something on the three main personnel changes to the niners defense (Brandon Moore at linebacker, Keith Lewis at free safety), or the switch back to a 4-3, or Bryant Young's resurgence. How about the two free agency additions to the left side of the offensive line opening things up (or otherwise) for Frank. The emergence of Arnaz Battle, the disappearance of Antonio Bryant, the sudden inability of Joe Nedney to make chip shot field goals.

In fact, how about some actual analysis of how the niners won a meaningful game, against a division rival.

PS - And to answer the heckles I'm sure to receive, yes, I know it's a free site, and they can write what they want, and it's worth what you pay for it, and everything else, but this is a long way below the quality I'd come to expect from FO

by Nate (not verified) :: Tue, 11/21/2006 - 4:36pm

Wow, Mr. 31 - He mentioned the addition of Fields, the continuing good play of Bryant, and complimented Allen and Jennings on the run blocking. Did you even read the article?

by Nate (not verified) :: Tue, 11/21/2006 - 4:36pm

Bryant = Bryant Young, not Antinio Bryant.

by Wanker79 (not verified) :: Tue, 11/21/2006 - 4:42pm

Re: 31

Nobody's going to point out that it's a free site. What people may point out is that nobody gives a flying crap about the 49'ers and that the reason people would be interested in this game would be to see what has happend in Seattle to bring them to the point that they lose to the stinking 49'ers.

by Kaveman (not verified) :: Tue, 11/21/2006 - 5:05pm

#34: Speak for yourself?

Way to bring all the 49er fans into the fold. Personally, I like having fans from multiple teams posting on these boards. The 49ers may currently be a bad team, but there are a whole lot of people who do give a "flying crap" about that.

Why beat up on other people's teams? I like football and the Denver Broncos, not the other way around.

by tshipman (not verified) :: Tue, 11/21/2006 - 5:06pm

Some of us do care about the 49ers. Wouldn't it be more interesting to look at why a team that you haven't watched much won against a team you thought was superior?
Does it really take much analysis to point out that Seattle was missing Hasselbeck and that Alexander was limited? That seems really simplified analysis. Well, yeah, those guys were hurt, but what did the guys who actually, you know, won the game do? That would have been a more interesting article, whether you wanted to read it going in or not.

by James C (not verified) :: Tue, 11/21/2006 - 5:43pm


Look at a roster, McNabb is a whole inch taller than Garcia it wasn't just a bad adjustment it was a crap snap. Plain and simple.

by MDD (not verified) :: Tue, 11/21/2006 - 5:43pm

Half of the Seahawks offense is a MASH unit, the other half seems to have gotten old all at once, while the defense has been a bad sequel of 2004. That, folks, is your 2006 Seattle Seahawks.

by Brooklyn Bengal (not verified) :: Tue, 11/21/2006 - 5:44pm

RE: #24-#26

er...um...yes and no. the kind gentlemen here at FO have established that the probability of any team recovering any given fumble is about 50%, so we're going to compare it to an event of equal probability, like a coin flip.

yes, the probability remains at 50% for any individual coin flip (or in this case, fumble). so if one flips a coin ten times, one expects it to land on heads 50% of the time. however, if one has flipped a coin ten times and it has landed heads nine out of ten times, the chance expectation for the next ten flips changes in relation to the known probability of success. it's certainly possible to get another 9 heads and end up with 18/20 coin flips as heads. however, keep flipping that coin. the longer you flip it, the less likely you are to keep seeing so many heads, because ultimately the heads/tails ratio will even out.

just like fumbles. taken out of context, one would expect the falcons to recover 50% of their fumbles in the second half of the season...UNLESS one knows that the falcons recovered a disproportionate amount of fumbles the first half of the season. since one knows the known probability of success is 50%, one can expect the falcons' success rate of fumble recovery to drop in the second half and even out to 50%. if it doesn't drop, or even doesn't drop enough even out (and continues as a pattern season in and season out), then the known success rate is incorrect, FO has made a mistake and fumble recovery is not random (and tom stoppard prepares to write a new hit play titled "houshmanzadeh and tomlinson are dead").

so i'll maintain that the informed expectation holds that the falcons would recover significantly fewer than 50% of their fumbles the second half of the season PROVIDED they recovered greater than 50% of their fumbles the first half of the season.

by Pacifist Viking (not verified) :: Tue, 11/21/2006 - 5:51pm

Brooklyn Bengal, in all this talk of coin flips, I hoped somebody would mention that Stoppard play!

However, if a particular even (coin flip or fumble) has a 50/50 probability of going one way or the other, then each particular event has a 50/50 probability and previous occurances have zero impact on future occurances. Yes, over time you would expect the occurances to even out to a near 50/50 occurance, but the rest of a single season might not be enough time for that to happen.

It's like if a woman has eight boys, and you suggest that the probability shows the next child will be a girl. It really doesn't--the next child has a roughly 50/50 shot of being a boy or a girl. But in the whole scheme of human events, of course, this averages out to a "roughly" 50/50 standard.

by chris clark (not verified) :: Tue, 11/21/2006 - 5:56pm

re: random fumble recoveries and the gamblers falacy

One of the top FAQ discussed items, 2nd only to why some teams DVOA is too high or low.

As I recall, Aaron posted recently (week 9, link at my name) that comparing fumble recoveries for two halves of a season had a correlation coefficient of roughly "-.1". Thus, given the small sample size, it is appropriate to treat fumble recoveries as unpredictable. That's the right way to do one's statistics. This is often paraphrased as fumble recoveries are random.

However, the correlation coefficient was not 0. If fumble recoveries are truly random, the coefficient should be 0. If the correlation were 0, then an exact prediction of 50% (or whatever the underlying probability is) would be appropriate, and thinking otherwise would be the gamblers falacy, which roughly stated is "red" (some event) has come up too often, now "black" (its converse) is due.

However, with a (small) negative correlation, the gamblers falacy is actually the correct bet. A negative correlation means that things do tend to get corrected. A small correlation, simply means that the amount of correction isn't large, but still one should expect a slightly better chance of the event being reversed (i.e. the team will be unlucky at recovering fumbles) for the 2nd half of the season.

A small correlation and a small data set means it is hard to tell whether that correction is real, which is why Aaron is correct in discarding it, and replacing it with an assumption of randomness (non-correlation).

Note, before re-reading that article, I thought fumble recoveries has a small positive correlation. With a small positive correlation, one could expect a team that has above average luck on fumble recoveries to continue that luck. Moreover, with a small positive correlation, I suspect Aaron might have been tempted to pursue the issue further and to determine what makes some teams "lucky". Note, with a small positive correlation, it is possible to argue that fumble recoveries are not random and that we just don't have enough data to tease out the connection. However, with a small negative correlation, one isn't so tempted to figure out "why" some team's luck reverses, one can just ascribe it to "regressing to the mean".

And to pacificist viking's question, "random" seems to be used on this site in its "common usage" meaning unpredictable (or unsustainable), and "lucky" to mean random and advantageous.

This is different than how a mathematician would use the word. To a mathematician "random" means drawn from a probability distribution. Note a mathematical random event can have a probability (expected value) of 60% (or even 100% or 0%), it just has to have a distribution of values around that expected value. In some sense, all stats are "random" (in the mathematical sense), because one needs a distribution for statistics to make sense. However, this mathematical use of the term random is not how the word is used on this site.

Now, I hope I don't get penalized for "piling on".

by Brooklyn Bengal (not verified) :: Tue, 11/21/2006 - 5:58pm

RE: #16

yes, a coin flip can determine the outcome of a high-scoring game such as the Steelers/Falcons game a few weeks ago. but wasn't there a bit of luck that the game even went to overtime, with the Falcons making a field goal to put them up, only to have the field goal negated by a late time-out?

FO has stated many times that really good teams blow out weaker teams. i.e., really good teams overcome luck. take the michigan-osu game last weekend. osu was clearly the better team. they had a ghost PI call on a long pass which directly led to 7 michigan points. osu had bad luck with fumble recovery and some very ATYPICAL mishandling of the football which could be seen as luck. and they still outplayed michigan completely. yes, the final score was close but the outcome of the game was never in doubt. i think the steelers are a talented football team, and of course as their luck evens out they will get a few more wins. but i don't see how anyone can have a quarterback with 12 interceptions and 17 interceptions and expect their team to have a winning record or blame their team's losses on luck.

as the "any given sunday" article detailing the bucs' win over the bengals explained, while cincy lost on a "lucky" horrific call by the refs, the fact that that one call made the difference in the game only showed that the highly-touted bengals and the lowly buccaneers had played an even game.

coincidentally, the bengals had lost four of five and hadn't recovered a fumble the past two months. they got a fumble recovery last week and blew out the saints. do i blame their losses on luck? no. i blame it on their underperformance, on would-be interceptions dropped by nervous rookies, by blown pass coverage and a weakness against the run.

would the bengals have lost to the ravens without poor luck recovering fumbles? would they have lost falcons without poor luck recovering fumbles? it doesn't matter. they are underperforming. if they were significantly better team than the ravens or falcons, they would have overcome poor luck and dominated the games. as it is, all the bengals have proven is that they're about as good as the ravens and falcons.

also, let's keep in mind that when YOU DON'T FUMBLE (coughChrisPerrycough), you DON'T HAVE BAD LUCK RECOVERING YOUR FUMBLES.

by scott (not verified) :: Tue, 11/21/2006 - 5:58pm

I forget whether it FO, Bill Simmons or TMQ who mentioned it before the season, but here goes:

While the NFL is supposed to adjust for scheduling based on a team's performance (i.e. the super bowl winner gets a nasty schedule, the worst team gets a cupcake schedule), the NFL gave Seattle a relatively easy schedule becuase of how badly the Seahawks got screwed in the Super Bowl by poor officiating. And I agree with that decision.

As a result, let's give a round of applause for your soon-to-be NFC West div champs, the Seahawks!

by MDD (not verified) :: Tue, 11/21/2006 - 6:00pm

Um, you don't need to start recovering more than 50% of fumbles for the number to approach 50%. at 9/10 fumbles you're at 90%, +9 fumbles, but if the next 100 go 50/50 as expected, youre at 59/110 or 53.6%. The fact that you've recovered 9 of the last 10 does not change the fact that you are just as likely to recover as you are to lose 9 of the next 10. Most likely is a 5/10 distribution, putting you at 14/20.

The link on my name has a really sad looking depth chart if you're a Hawks fan.

by Brooklyn Bengal (not verified) :: Tue, 11/21/2006 - 6:00pm

RE: #43

it's my understanding that the majority of scheduling is done years in advance on a steady rotation, with something like four games tbd post-season.

by Mwana Uta (not verified) :: Tue, 11/21/2006 - 6:10pm

Have to agree with #31, this was a disappointing article. I was looking forward to it, as I can't remember the Niners ever being featured on Any Given Sunday or EPC (go figure, this is the first time we've won three in a row since this site was launched).

Anyway, you should have let Doug handle this one, he made more relevant analytical comments in his submissions to Audibles than the whole of this article.

by KurtisM (not verified) :: Tue, 11/21/2006 - 6:15pm

RE: #43

Scheduling talk may eventually cause me to lose it entirely... NFL schedules are almost set in stone now that we have 4 divisions per conference. Each team gets 6 games against division rivals, 4 games against another division in the same conference (on a set rotation), 4 games against a division in the other conference (also on a set rotation), and 2 other games. If I recall correctly, those last 2 games are against a team in each of the remaining divisions in your conference, and are against the teams that finished in the same spot as you the prior season. Aside from those 2 games, the only thing that determines ease or difficulty of schedule are the teams in the divisions your division is set to face that year.

by Brooklyn Bengal (not verified) :: Tue, 11/21/2006 - 6:34pm

RE: #41

Thanks for the link to Aaron's work. I must have missed the DVOA that week (or maybe I just skipped commentary with the obligatory zlionsfan warnings)!

by MFurtek (not verified) :: Tue, 11/21/2006 - 6:50pm

I agree this article overly focussed on Seattle.

17 paragraphs on Seattle's team, including a few on Hutchinsen.
4 paragraphs on Gore. I'm surprised no one has treaded Ned like Peter King, calling Gore "little-known".
5 paragraphs on the other 49ers.

I think many of us expected the focus to be on San Francisco. That's not an indictment on the article, it's just Ned chose to focus on something we didn't really expect.

It wasn't an easy week to come up with an AGS as it is.

by BillWallace (not verified) :: Tue, 11/21/2006 - 8:42pm

One thing I noticed watching this game is that the 49ers defense seemed to read Seattle's plays quite frequently. Much more often than I usually notice watching a game. For example.

On Walt Harris' first interception, he had double coverage deep, which helped, but he basically ran the EXACT route (buttonhook) that the WR ran, 1 yard shorter, and seemed to know exactly when to turn around to catch the pass. Also on a number of screens and other misdirection plays, a 49ers defender just seemed to know exactly where the play was going. On one 'fake handoff one way then pitch the other way' run, the DE on the pitch side was literally 2 feet in front of the RB, 5 yards in the backfield when the RB caught the pitch.

This could have just been a lot of luck but it was worth mentioning.

by Carlos (not verified) :: Tue, 11/21/2006 - 9:06pm

What's happening w/ Vernon Davis' return from injury?

by Luke (not verified) :: Tue, 11/21/2006 - 11:39pm

#50 - I think Trent Dilfer might have had something to do with this. That and the fact that Mike Holmgren is a very predictable playcaller after the 1st quarter. Where were the designed QB runs to take advantage of Seneca's amazing speed? Holmgren is a dinosaur.

by Polaris (not verified) :: Wed, 11/22/2006 - 2:48am


I am not here to defend Seattle. They played poorly and if anyone has been following the NFC, even a bad team can beat the best team on any given Sunday or Monday. Just ask Chicago!!!

That said, Seattle wasn't blown out and did hold the 49ners to ZERO points the second half...a point that Aaron chooses to overlook. I don't dispute his main points, but:

1. The real "any given Sunday" article really should be about the Eagles. They were the #1 rated team in DVOA in spite of many of us saying they were severely overrated. Philly was losing (and performing badly even if they had won) against the lowly Titans even before McNabb got hurt. Besides....I didn't think that DVOA was supposed to consider injuries.

2. I think that FO decided to focus on Seattle because the result instead of a simple upset allowed FO to foster the meme that Seattle is a terrible team regardless of record. That may be true...but frankly I think that FO decided to rag on a team and a game to better justify their own model while ignoring evidence that their model may need work.

Harsh, but that's the way I see it. The article reeks of this intellectual shortsightedness.


by centrifuge (not verified) :: Wed, 11/22/2006 - 2:50am

42: Well, that was awfully aggressive. I was just pointing out some reasons why the FO staff or others on the boards might apply the "unlucky" label to the Steelers. I'm not making the claim that they are a good team being held down by bad luck. I would say they are a slightly-above-average team held down by bad luck and some boneheaded coaching.

They certainly have at least one loss that can be attributed directly to bad luck -- the Oakland Raiders. Two INTs in the red zone returned all the way back for touchdowns is well above the league average, and the Oakland offense was certainly no threat to have capitalized on its own. Still, even that would put them only at 5-5, which is about what they deserve, all things considered.

by putnamp (not verified) :: Wed, 11/22/2006 - 6:28am

What's funny is that even in an article about the 49ers and the Seahawks, ppl can't stop talking about the Eagles and the Steelers and so on :P

by Rick (not verified) :: Wed, 11/22/2006 - 7:43am

Re: 25, 26
If the Seahawks recover 90% of their fumbles out of the first 8 games, it means that the hypothesis that their true recovery rate is only 50% is likely to be wrong. (Read some of the works of Rev. Bayes.)

The fact that a phenomenon can be modelled as a random event does not mean that it is a random event. (And even if it were a random event, there's no requirement that the random distribution be uniform. Consider the phenomenon of a jump ball in basketball. Now consider a jump ball between somebody 6'7" and somebody 7'2".)

by Wanker79 (not verified) :: Wed, 11/22/2006 - 1:09pm

Just for the record, my post in #34 was more of a lashing out at someone bitching and complaining that the article didn't take the perspective they were hoping it would. Sorry.

by centrifuge (not verified) :: Wed, 11/22/2006 - 3:24pm

56: If the Seahawks recover 90% of their fumbles out of the first 8 games, it means that the hypothesis that their true recovery rate is only 50% is likely to be wrong.

Not exactly. If the Seahawks recover 90% of their fumbles out of the first 8 games, it means that the hypothesis that every team's true recovery rate is 50% is still likely to be true. It's just that, with 32 cases running at once, random distribution requires that some teams fall towards the tails.

If that 90% rate were really representative of the team, wouldn't it carry over to the second half of the season? That -.1 correlation between halves of the season says that such an impressive percentage is rarely maintained.

by MFurtek (not verified) :: Thu, 11/23/2006 - 5:03am

Just got to halftime in the game.
1) 49ers did a good job faking handoffs in one direction and doing a screen throw-back or just running the screen in general. It wasn't all Gore, but mostly set up by his effective running.

49ers ran some effective end arounds and even a reverse was thrown in, with Alex Smith lead blocking.

2) Seahawks are down 20-0 and turned the ball over 3 times in the first half. Seneca Wallace was BRUTAL and when he completes a pass, it is fumbled by Jerramy Stevens (who had a 3rd down drop). Their passing offense was awful the first half.

3) Michael Robinson looks like he could be a good RB at spelling Gore.

4)I didn't watch with sound on but it looked like the 49ers should've been up 24-0. I have no idea how the TD was overturned at the end... NO IDEA!

I guess it could have something to do with your shin-calf being down...

by MFurtek (not verified) :: Thu, 11/23/2006 - 5:25am

2nd half thoughts.
1) Seattle opened with a good drive. Alexnander had a good run to get some first downs, there was a good bootleg, and Wallace his Branch on an 8 yard hitch that he turned into a 38 yard TD. Just a really bad job but CB (Spencer) and Safety who both over-ran Branch.

2) Celebration rule pisses me off. I was watching a special on Fox on "50 Greatest Football Plays" and I think most of the celebrations I saw would've been flagged under the new rules. Players should be flagged for celebrating inapproprately, like after a sack when you are down 20-7.

3) SF didn't do anything in the 2nd half. Their first drive was decent with a long Gore run, but when they could make another play it was called back.

4) Wallace looked hot and cold. He caught fire in the middle of the 2nd half, then lost it at the end. Not sure where that last throw was going.

Also, Alexander looked to be doing a lot better in the 2nd half, 4th and 1 not-withstanding.

by Joel (not verified) :: Thu, 11/23/2006 - 1:18pm

You forgot that the seahawks are also affected by the Chunky Soup Curse