Guest columnist Zachary O. Binney fact-checks a story in a national publication and finds that everyone makes mistakes.
03 Feb 2011
by Mike Tanier
The hookers are coming! The hookers are coming! Or maybe they aren't.
Last November, Dallas police sergeant Louis Felini announced that he anticipated between 50,000 and 100,000 prostitutes descending upon his city for the Super Bowl. According to the Dallas Women's Institute, 38,000 of those prostitutes would be underage. Dallas, like the host cities of past Super Bowls, braced for a Category 5 hooker storm.
The problem is that it is all nonsense. Peter Kotz of the Dallas Observer does such an incredible, hilarious job debunking the hooker myth that I don't dare steal one of his insights. OK, I will steal one: The Dallas Women's Institute calculated their underage prostitute estimate by looking at advertisements by back-page escort services and guessing the age of the ladies in the photos. It was well-intentioned but overwhelmingly distorted activism masquerading as data analysis.
The Super Bowl Hookers myth has been around for years. It spawned in the same nutrient-rich bunk propagation environment that made Super Bowl Sunday a "day of dread" for domestic violence attacks. That particular urban legend spread in 1993 and lingered for several years. Domestic violence and prostitution are serious social issues, but shoehorning them into the Super Bowl hype both trivializes them and focuses attention away from their real root causes and possible solutions.
Still, all it takes is one scientific-looking report with the words "Super Bowl" attached to it to grab national attention during hype week. Hand-wringing, uncritical recitals of the "underage prostitute" myth are appearing all over the media right now. A brief Google search should yield a few dozen, with one or two skeptical articles like Kotz's shouted down in between.
The Hundred Thousand Hookers headline was so crazy that it should have immediately raised doubts. Other Super Bowl-related stories are less sensational, but they still should be looked at more carefully. When an article about Super Bowl-related heart attack deaths started circulating this week, my skeptic senses started tingling. According to the study, cardiac events resulting in death increase 15 percent among men and 27 percent among women in the two weeks after the local team loses the Super Bowl. Most of the heart attacks occur among individuals over 65, according to the study, which reported an extra 2.6 deaths per 100,000 people each day for 14 days after a Super Bowl loss.
The heart attack study was published by Clinical Cardiology. A cursory look at their site suggests it is a legitimate, peer-reviewed journal. Still, the methods outlined by Dr. Robert Kloner do not sound very promising. Kloner and his colleagues counted heart-related deaths in the Los Angeles area during the two weeks after the Super Bowl XIV in 1980, when the Rams lost to the Steelers. For controls, he used Super Bowl XVIII, when the Los Angeles Raiders beat the Redskins in 1983, and seasonal averages for heart-related deaths in Los Angeles in non-Super Bowl years.
In other words, we are basing a national story on 30-year-old data from the mainframe era of data management, pertaining to one two-week period in one urban area. The 1983 Raiders data didn't stray much from the non-Super Bowl control group, according to abstracts, so all of the dramatic percentages reported are based on two weeks in Los Angeles in January of 1980.
Let's look at one major factor that was immediately overlooked. Pasadena hosted Super Bowl XIV, making it essentially a home game for the Rams and creating a unique situation for Los Angeles: It was both the host city and a participant. Without reading the full study, I don't know how Kloner adjusted for or acknowledged this unique situation. Right now, Dallas is choked with extra traffic (made worse by bad weather), people in service industries are working extra shifts, and thousands of travelers not used to the stresses of airline flights have descended on the city. Has anyone checked the heart attack data for a host city during the Super Bowl? The unique role played by Los Angeles in 1980 make it a bad choice for a study about Super Bowls and heart attacks.
The strange male-female split also raises doubts. "It may be the same emotional response as it is for men. Women root for their teams, too," Kloner told CNN Health. "Another possibility is that perhaps a mate's reaction adversely affects the female." That sounds a little like the Day of Dread myth, a convenient (and condescending) simplification of the husband-wife dynamic. John Stallworth scored a touchdown, Bob screamed at Hazel for the rest of the night, and Hazel keeled over from grief and sorrow. It sounds more like a morality play than a scientific explanation.
The CNN article includes a dissenting opinion from Dr. David Frid, who points to other Super Bowl-related factors that could result in a heart attack besides excessive rooting -- fatty snacking, anticipation leading up to the game, and so on. Frid never mentions the regional stress of hosting a game, but I do not expect him to -- he is a medical expert, not a football expert, so he does not know how unique the 1980 Los Angeles situation was. I do expect experts like Frid to point out that 30-year-old health data is extremely out of date.
The American Heart Association reports that the coronary death rates decreased 34 percent from 1996 to 2006. So even if Kloner's death rate spikes are accurate for 1980, and are completely caused by the stress of watching the Super Bowl as a fan, the mortality figures would be closer to 18 percent for women and 10 percent for men today, if not far lower. Remember that most over-65 adults in 1980 were lifelong smokers, increasing both their heart attack and coronary-mortality risks. Comparing an average 65-year-old in 1980 to one in 2011 is medically specious.
What about World Series games and NBA Finals? A close Rose Bowl featuring USC? Even assuming Kloner only had access to Los Angeles medical data of the early 1980s (and I am not sure why that is all he could get), there were plenty of other sports-related events to control for or analyze, to say nothing about other stress-producing regional events like political conventions. He could have checked Super Bowl XI or XVII, both hosted by Pasadena, to control for any "host city" bias. Nope. These are the questions I ask before I even start to worry about any counting biases that could infect the data.
(Super Bowl XVII may have gotten into the study, as Kloner checked the "corresponding days in intervening years" between the Rams and Raiders appearances. Note that he checked the corresponding "days," though. Super Bowl XIV took place on January 20, Super Bowl XVII on January 30, Super Bowl XVIII on January 22. His control studies probably included one week of a host city leading up to a Super Bowl. The moral of the story: Ot would be great to have some non-80s, non-Los Angeles data.)
My take on this data is that if we adjust down the fatality percentages to account for the differences between 1980 and 2010, then do a little dead reckoning to assume that hosting the Super Bowl leads to some small spike in heart attack rates, then compare the Super Bowl with other championships, political conventions, or other events, we might discover that any emotionally charged civic event causes a minor increase in a variety of health-related issues. Yes, I am making a lot of assumptions, but at least I am not passing along three-decade old data as a major health breakthrough.
Heart health is a serious issue, and I can certainly imagine a diehard fan doubling over in stress-related agony in the fourth quarter of a close game. For that matter, I can see an angry fan taking a bad game out on his spouse, or some Super Bowl revelers picking up a streetwalker who bussed into town just for the employment opportunities.
My doubts come not just from the nature of the Clinical Cardiology data, but also from the credulity with which they are being disseminated. The CNN article comes with a picture of two chunky Packers fans screaming at the television. An NPR blog ends with the admonition "so think of that when you are watching the Super Bowl." Everyone, including me, is quoting the same facts from abstracts and press releases, a sign that the story is just being passed along non-critically.
This is not medical reporting, it's finger-pointing and gawking at those silly football fans, who stuff their faces while screaming at the game, then commit acts of violence and perversion before writhing in agony and dying. The Super Bowl has become a repository for our national fears and obsessions. It's a media Monster from the Id, a catalyst for sex and death, a deflowerer of young maidens, a destroyer of marriages, a scourge of the elderly. It's the funhouse mirror of American excess. Soon, we'll hear of chainsaw-wielding lunatics with hooks for hands prowling the Super Bowl parking lots. A special-interest group will release a specious survey, 300 media outlets will run with it, and a portion of America will believe it.
Making sure grandpa takes his medicine before he goes on a Sunday bratwurst binge is a good idea. Distorting football viewing into a deadly pastime, based on a 30-year-old study with dubious controls is a bad idea. So is vilifying football fans as a bunch of obsessive jerks who assault their wives then prowl for prostitutes. Fans are better than that. And journalists are too smart to keep sloshing this bunk around.
Let's get to some football.
Defensive Player of the Year Troy Polamalu intercepted seven regular season passes. Some were great plays by Polamalu, others were great plays by other Steelers defenders, a couple were batted passes that just landed in the safety's hands. I decided to take a closer look at each interception, diagramming a few of them to see how the Steelers' blitz and coverage packages worked in concert to force turnovers. The following diagrams reveal a few things about Polamalu, but they reveal a lot more about how the Steelers approach pass coverage.
|Figure 1: Interception 1|
Interception One: Figure 1 shows the Steelers facing the Falcons late in the fourth quarter of a tie game. The Falcons are in two-minute mode, and the Steelers are in their 2-4-5 personnel grouping. The pre-snap read suggests a Cover-4 defense, with the cornerbacks 8-to-10 yards deep and the safeties out of the television frame.
The Steelers aren't in Cover-4 on this play, but in Cover-3 Cloud, meaning the cornerbacks are covering the deep sideline thirds of the field while one safety covers the deep middle. Polamalu has the underneath sideline zone to the offensive right. The underneath zones are rolled in that direction, a reaction to the fact that the Falcons have three receivers to that side. Roddy White (84) is both a dangerous boundary receiver and the Falcons' go-to target in most critical situations, and it's clear at the snap (as seen from a different camera angle) that Polamalu is anticipating an out-route by White.
Switching from a Cover-4 shell to three-deep coverage is not particularly exotic -- every team does it. The Steelers excel at it, in part because their pass rush limits the quarterback's ability to adjust, in part because of Polamalu's exceptional open-field range. Not every NFL safety can start 15 yards deep but still make a play in front of a receiver along the sideline. Polamalu's mix of speed and preparation allow him to read this play and jump this route.
Interception Two: Polamalu picks off Vince Young in the end zone. The Steelers appear to be in man coverage with Ryan Clark deep. Polamalu peels off tight end Bo Scaife, running a seamer, and cuts in front of a pass to Nate Washington, running a deep in-route in the end zone. It's not an interesting play from a diagramming standpoint, so I didn't draw it up. Polamalu makes a great read in the middle of the field, but this is really a poor decision by Young, who chucks the ball into a crowd.
|Figure 2: Interception 3|
Interception Three: Now, let's get to some serious Steelers deviltry. Figure 2 shows the Steelers facing Bruce Gradkowski and the Raiders on third-and-10. As shown in red, Bryant McFadden (20) and William Gay (22) leave their receivers to blitz from the offensive right. James Harrison (92) and James Farrior (51) show blitz but drop into zones.
The Steelers do not work very hard to disguise this blitz. McFadden and Gay start shuffling toward Gradkowski about a second before the snap. Clark (25) has deep zone responsibility to that side of the field, but an alert quarterback might have hit Johnnie Lee Higgins (the split end to the right) for a short completion in front of Clark. Gradkowski attempts instead to hit Jacoby Ford up the seam, leading Polamalu with his eyes the whole time.
Farrior does an excellent job on this play, chasing Ford up the seam and forcing Gradkowski to throw into a tight window. Farrior appears to have underneath zone coverage, but as an experienced defender, he understands his role. Thanks to the blitz, no running back will threaten an underneath zone, so he gets deep to help Clark and Polamalu. The double-defensive back blitz is a great design. The Packers frequently use similar blitzes to this one, so you may see both teams use a tactic like this on Sunday.
We will get back to Polamalu after a brief interlude. By placing the next segment here, it spreads out the diagrams and makes them easier to read!
Reader Allyson writes:
I have one of those fingernails-on-the-chalkboard moments whenever I hear an announcer say, "That ball was fair-caught!" Am I correct that the term "fair catch" is a NOUN and there IS no past-tense verb form? This drives me absolutely crazy, and I hear it often from well-paid, highly educated broadcasters.
Also, there were three words/phrases that would have been great for a drinking game this past college football season: "finesse," "poise," and "in space." One would not be capable of driving or operating heavy equipment/machinery after about one quarter of these over-used words!
Without consulting my grammar expert friends, I would say that "fair catch" is a noun. It is rarely used as a verb, and it sounds awkward to say that Blair White fair-caught 58 percent of his punts this season, even though it is true. The problem is that "fair catch" is a technical term that is not as pliant as it needs to be. We have to use verbs like "signals" to make the term work. Writing about fair catches gets tedious because the word lacks a good verb form. White executed 14 fair catches? Attempted them? Saying he "signaled" them is misleading when talking about season statistics, because I really do not know how many times he actually signaled.
Television announcers have it far worse, because they are describing action on the fly, and sometimes words get repurposed in sloppy ways. I am more forgiving of television announcers than many writers, possibly because of 17 years of speaking in front of teenagers. "Fair caught" does not bother me in speech because the announcer is trying to describe a rather dull event as quickly as possible.
I got to talk about "finesse" with Ben Zimmer in Sunday's The New York Times. "In space" has become a tedious college football buzzword because of all the spread-option plays and screen passing. "Poise" is one of those storyline words that is used to turn a five-yard completion into a statement about a player's character. I probably would not survive any drinking game based on buzzwords like those. Maybe 20 years ago!
|Figure 3: Interception 4|
Interception Four: The Bills empty the backfield on second-and-7 in Figure 3, and they send their tight end in motion to make the Steelers tip their coverage. The Steelers respond by sending everyone in motion. Pre-snap, Harrison, Timmons, and Farrior bunch up in the middle of the field, with LaMarr Woodley (56) over the tight end. When the tight end goes in motion, Woodley follows him but drops into coverage, while Harrison comes up to the line of scrimmage. Farrior, Timmons, and Harrison all blitz. Without knowing the Steelers playbook, I believe that they called a strong-side blitz and wanted two defenders attacking from the tight end side, one to eat up the tight end if he stayed in to block.
There is man coverage behind the blitz. Polamalu doesn't do much on this play except wait for the ball to bounce off Steve Johnson's hands and into his. Just like in the Raiders interception in Figure 2, the quarterback makes a bad read when faced with a Steelers blitz package, even when he isn't hurried by the pass rush.
Not shown in the already confusing diagram is the quick slant run by Donald Jones against Anthony Madison (37). Jones gets open, and considering the situation (second-and-7 at the 12-yard line in a close game), a five-yard completion would be useful. Ryan Fitzpatrick does not see Jones, and Gay does a fine job on Johnson.
Interception Five: Polamalu ran back a Carson Palmer pass intended for Terrell Owens for a touchdown in Week 14. The Steelers only rushed three defenders on the play, with man coverage and both Polamalu and Clark deep. The cornerbacks had outside technique on the receivers, meaning McFadden stays between Owens and the sideline and starts at a depth of eight yards.
Outside technique is common for a cornerback with safety help, but McFadden got caught flat-footed when Owens got up the field quickly on a post route. Polamalu undercut Owens and jumped in front of Palmer's pass. Most safeties don't make the interception on this play (which is not diagramed), but a good safety should be in position to at least break up the pass. The difference between a good safety and Polamalu was, in this case, the difference between a pass defensed and a touchdown.
|Figure 4: Interception 6|
Interception Six: Figure 4 also comes from the Bengals game in Week 15. It is late in the fourth quarter and the Steelers lead 23-7, but the Bengals are on the 17-yard line. Polamalu is not even visible on the television tape, even though the far corner of the end zone can be seen. The Steelers are playing Cover-1, with Polamalu in center field and Woodley lurking in the middle. Farrior and Timmons blitz from the linebacker position. What's great about the Steelers defense is that they can line up in exotic formations to confuse you, but they don't have to -- all four linebackers blitz very well, so they can give you a vanilla pre-snap look and still surprise you.
Carson Palmer appears to do a good job of not tipping his intentions on this throw. He looks left before turning to throw to the right, but a later replay shows that Polamalu only takes one step in the wrong direction before limping over (he got hurt on the previous interception) to jump the pass to Owens. Polamalu may have been reading route combinations and tendencies. When the slot receiver to the left pulled up for a short curl, he saw that he throw was more likely to go the other way. He may also have just been in a gambling mood, which is why he pitched the ball to McFadden after the interception instead of just going to the ground.
That play was an interception of circumstance. If you don't want to get picked off like that, make sure you aren't trailing by 16 points late in the fourth quarter.
|Figure 5: Interception 7|
Interception Seven: We have seen cool blitzes and crazy pre-snap motion. Let's finish with some funky hybrid coverage. Figure 5 shows the Browns in an empty backfield formation on first-and-10 at the beginning of the game. The Steelers appear to respond with man coverage, with Polamalu aligned over a slot receiver (tight end Robert Royal). But this is not man coverage, and Polamalu is not covering Royal. It's a three-deep zone, with some man coverage mixed in underneath.
Defenses typically assign numbers to the eligible receivers on offense, working from the quarterback out on each side of the formation. The inside-most eligible receiver to either side is No. 1, the next one out to that side is No. 2, and so on. Backs directly behind the quarterback (as in an I-formation) might both be declared No. 1, with assignments based on how they release. The numbering system allows the defense to account for multiple formations.
In this empty backfield set, Ben Watson (82) is No. 1 on the right side, while the slot receiver is No. 1 on the left. Farrior and Harrison are responsible for those two players in man coverage. Had the Browns lined up in a split backfield, Farrior and Harrison would be covering running backs. In any event, they are playing man coverage over the middle of the field, which allows the Steelers to fan their cornerbacks along the sidelines and place Polamalu and Timmons in the curl zones, roughly 10 yards down the field and outside the hash marks.
This coverage scheme would be hard for a veteran quarterback to read, and Colt McCoy does not see that Royal sprints right past Polamalu and straight up the seam, with Clark in poor position to stop him. (McCoy may have just known that it's Royal and decided not to bother.) McCoy does find an open receiver, as Watson shakes off Farrior with a little scat route. Unfortunately, the ball bounces right off Watson's hands and into Polamalu's. Polamalu has a habit of being in the right place at the right time on these tip-drills. There is some luck involved, but there is also some skill, both on Polamalu's part (he has good hands and gets close to the play) and on his teammates' (quarterbacks are often under duress, passing windows are tight).
The seven interceptions provide a glimpse of all the things the Steelers can do on defense, and we should see all of these things on Sunday. Does Polamalu deserve Defensive Player of the Year honors? I would have voted for Clay Matthews, but I am not picketing or anything. Great players, great defenses, great teams. Let's just watch them and not quibble about who won which individual award.
Looking for Walkthrough-style humor? Check out the Fifth Down blog at the The New York Times website. They have me chained to the desk all week, churning out crazy observations from the football realm. So far, I have written about Nick Barnett's decision to leave the Twitter realm, an abomination known as the Pittsburgh Cheese Steak, the Eminem iced tea commercial, Packers Media Day, a possible Media Day mode for Madden, and of course, hype itself. More is coming. It's probably more like Walkthrough than this article was this week, so head on over, then come on back for more fun here at FO!
48 comments, Last at 16 Feb 2011, 11:15am by armchair journeyman quarterback