Comparing the final Football Outsiders F/+ projections to the AP poll and conventional wisdom.
14 Jun 2012
by Mike Tanier
D.J. Williams tweeted an image of the Broncos playbook late last week, sparking an outcry from fans who feared that he betrayed precious Broncos secrets, as well as the usual "how dumb are you?" snark from those of us who trade in such things.
Williams deleted the picture and tweeted an apology, but you cannot erase the Internet.
Williams probably did not realize the image resolution was that good. He was making a joke about how he prefers writing out index cards to thumbing across a newfangled iPad when learning a new system. (Defensive system, not technological system.) He may have thought the image would be blurry, or guessed from the tiny thumbnail on his phone that the playbook would be unreadable. Or, judging from the page he photographed, he may have assumed that the plays he selected were no big deal.
Let’s take a look at those plays. "Sink Sam 1 Tite" is a simple blitz by the strong side linebacker. He is "Sam," and you can see him labeled with an S and lined up outside the tight end’s shoulder in each of these variations. The "1" stands for the type of coverage: man-to-man, with one deep defender, the free safety. Sink Sam 1 Tite starts with a five-man defensive line and two deep safeties, but the strong safety creeps forward and the free safety shades to the middle of the field. They can do this either pre-snap or at the snap.
|Sink Sam 1 Tite|
This is a pretty basic defense, something you might find in a high school playbook. The "adjustments" describe who covers whom when the offense goes in motion. In the base concept, the middle linebacker has the fullback, the weakside linebacker the halfback, the strong safety the tight end, and the corners lock down on their receivers. If the fullback motions right, outside the receiver to that side, the weakside linebacker follows. If he motions left, creating a "Trio" on that side, the strong safety picks him up, with the middle linebacker taking the tight end. By "right" and "left," I mean strong and weak side, since all of these concepts are probably mirrored.
None of these adjustments are particularly unique. There are probably about two more pages of them, covering empty backfields, spreads, and trips formations. The "Trio" motion in the middle right is the most interesting, as it forces the defenders to switch off. You can imagine an overmatched middle linebacker getting into Jimmy Graham trouble if the fullback motions left like he does in that diagram. When you see offensive coaches like Norv Turner use a lot of fullback motion, this is why: it can create favorable matchups by forcing the defense to make an adjustment it does not like.
Every system has its own little wrinkles. Williams is learning a new position and switching from a 3-4 to a 4-3, so he needs to study and review. He must also comb through all of those adjustments looking for variations unique to Jack Del Rio’s system. He probably won’t find them on the page that covers things like I-formations, but on the page that covers shotgun-bunch-right, back motioning to the left slot. That page was added 10 years ago, not 30 years ago. All of the adjustments are based on predictable patterns (the eligible receivers are numbered, based on their spacing from the quarterback), but Del Rio may have assigned the safety to cover a motioning H-back that Dennis Allen assigned to the middle linebacker. Finding these needles in a haystack is part of what a player goes through when he learns a new system.
So Williams did not tweet any top secrets. He did violate a non-disclosure agreement that players usually sign before getting their playbooks, and the Broncos may fine him and make him run some gassers. But no one in Oakland is cackling evilly and preparing to break the code of the Broncos defense based on this image. Williams might as well have been Tweeting the cover page, or a Madden screen shot.
The most interesting thing in the picture, for this long-time teacher, is the 75 cent pencil sharpener Williams uses. Anyone who knows pedagogy knows that the act of hand-copying those plays as study guides is the most important element of the studying itself: engaged note taking is so good for understanding and recall that it makes later study irrelevant. Keep those pencils sharp, D.J., and by September you will know the Broncos defense better than we now do!
1. Sam "Bam" Cunningham
Cunningham is another example of a running back who had the misfortune of playing his entire productive career between 1973 and 1977, the NFL’s Dead Ball Era. We have encountered a few others, like Chuck Foreman. Cunningham was not as good as Foreman, but his 1,000-yard 1977 season, and his 800-yard 1974 and 1976 seasons (with 4.8-to-4.9 yards per carry and some receiving value), would have been 1,100-1,300 yard seasons a few years later.
Outside of New England, Cunningham is most famous for a) having a nickname that Howard Cosell loved to say/mangle (calling him "The Bam" often), and b) being Randall Cunningham’s uncle.
2. Kevin Faulk
A long-time utility player for a great team, Faulk does not have many historical comps. Paul Hornung is one, but Hornung was a much better player. Dub Jones, the rusher-receiver-return man for the Paul Brown Browns, was another comp, and a closer fit. Faulk was probably more valuable to his Patriots than Dub was to his Browns.
In 18 postseason games, Faulk has rushed for 425 yards and averaged 4.8 yards per carry, caught 51 passes, scored two touchdowns, and added two two-point conversions. His postseason performance is essentially an extra season of third-down back play against the NFL’s best competition, and it is a very good one.
3. Jim Nance
A bruising fullback who is listed at 235 pounds on Pro Football Reference but as high as 260 pounds elsewhere, Nance was the AFL MVP in 1966, rushing for 1,458 yards. He had another 1,000-yard season in 1967, then had ankle surgery in 1968 and was never the same player. His yards-per-carry crashed from the high fours to the middle threes, his yardage dipped, and so on.
Two other interrelated factors affected Nance’s post-injury decline. First, the common draft started in 1967, and the AFL went from being a very skilled minor league to one more-or-less equal to the NFL quickly in the 1967 through 1969 seasons. Second, the Patriots went from a pretty good team to a miserable one, in part because some AFL franchises (the Raiders) were better run and better able to take advantage of the common draft than the Patriots.
Nance went on to be the all-time leading rusher in the World Football League.
Bill Parcells’ staff gleefully overused Martin as a rookie, ordering 368 carries from him. They toned it down to 316 carries in Martin’s second year, but with 48 catches it was still quite a workload. He posted negative DVOAs in those two seasons (-0.9% and -2.4%), as nearly any back would when plowing the whole field himself. By the time Pete Carroll took over, Martin’s DVOA was down to -7.3% and he missed several games. We will pick up the story again in a few paragraphs.
(Ed. Note: I will note that Mike is still using the currently posted DVOA ratings in these top five running back lists, not the new normalized DVOA v7.0 that we'll be unveiling in about a month. With the new DVOA, Martin's 1995 and 1996 seasons end up slightly above 0% rather than slightly below. All of the players Mike has written about from the mid-90s will see their rushing DVOA ratings go up by a couple percentage points. -- Aaron)
5. Mosi Tatupu
Tatupu was one of the all-time great special teamers, up there a notch below Steve Tasker and Bill Bates, on a threshold with players like Gary Stills. I am making that up, of course: there is no way to compare special teamers across eras. But Tatupu was very good, and if people can seriously suggest that Tasker was Hall of Fame-worthy, then I can suggest that Tatupu was the fifth-best running back in Patriots history. His qualifications include: a) a Pro Bowl berth as a fullback in 1986; b) two productive seasons as a change-up runner in 1983 and 1984; c) a long late career as the Kevin Faulk of his time; and d) a buried mention on The Simpsons.
Who ranks behind Tatupu? Don Calhoun shared the backfield with Cunningham for several seasons. Larry Garron made a bunch of All-AFL teams but played during the grain-of-salt era. Tony "Turbo" Collins had fine seasons in 1982 and 1983, but they weren’t outstanding years, and he fell off quickly. John Stephens led the team in rushing for several years during one of the Patriots hapless periods. All of them were better rushers than Tatupu in their primes, but none contributed so much that it obviously offsets over a decade of special teams service.
Antowain Smith and Corey Dillon were both productive for Super Bowl teams, but both were replaceable-part veterans for a team that was rapidly outgrowing the very concept of having a running back. Again, both deserve honorable mention, but they are stuck in a quagmire with the players listed above because their coaches clearly felt they were nearly disposable. And so, Tatupu.
There is one other, of course, but I shudder to type his name on the Internet.
As a thought experiment, let’s imagine what would happen if football were scored like boxing.
The teams play without knowing the score. Fans know that touchdowns, field goals, and safeties are worth points, but they are not sure exactly how many points. Announcers have to guess which team is winning based on their experience.
Teams vie to win quarters, not the full game. The score of most quarters is 10-9. Sometimes, there is a 10-8 quarter, like the second quarter of the Falcons-Buccaneers game in Week 17 (28 unanswered points by the Falcons) or the third quarter of the Packers-Lions Thanksgiving game (the Lions docked a point for Acts of Suh). But generally, the team that scores more gets 10 points, the team that scores less, nine points. Ties are scored 10-10, but just because both teams score the same number of touchdowns and field goals in a quarter does not mean the result is a tie. Judges could decide a quarter by other criteria, as we will see in a moment.
NFL judges work their way up through the ranks, from college and the CFL. They are experts, and there are rules and oversight, but there is a lot of subjectivity and wiggle room in the judging system, and the oversight is a legacy from an earlier time in sports history when scrutiny was not as harsh. No judge can make it through the ranks thinking that touchdowns don’t count, or that field goals are worth more than touchdowns, but little quirks can go undetected for years as a judge climbs from the FCS to Conference USA to the SEC to an alternate for Rams-Cardinals games to the Super Bowl. After all, the whole 10-9 system places most of the subtleties of the game inside a black box; all the judge has to be able to do is determine which team played better in a quarter. In this system, there is no objective reality to compare that judge’s choices to except the consensus of other expert opinions.
Let’s create three hypothetical football judges, each with his own criteria. They are all male, for the sake of pronouns.
Judge A uses our real football scoring: six points for a touchdown, seven with the extra point, three for a field goal, and so on. If a quarter ends in a tie, he marks it 10-10.
Judge B awards yardage bonuses. Twenty-yard touchdowns are worth seven points, 40-yarders eight, 60-yarders nine. Fifty-yard field goals are worth four points. Safeties are worth three points. He rose through the ranks with this system because a) he could judge dozens of quarters without anyone noticing he did it, and b) others use very similar systems, so his results matched theirs often enough that no one thought it was odd. And again, since our scoring system only exists in Judge A’s mind in this universe, there is no reason to think Judge B is being odd. He thinks of long touchdowns like basketball three-pointers. He scores a tie a tie.
Judge C thinks like a yardage-bonus, points-per-reception fantasy owner. Touchdowns and field goals are worth their standard values, but he awards a point per reception, a point per 50 yards of offense, a point to the defensive team for a sack, and two for an interception.
Now, let’s look at how this year’s Super Bowl would go, as judged by our expert panel.
By Judge A, the Giants won the first quarter, the Patriots the second and third, the Giants the fourth. He declares a draw. Judge B also declares a draw, even though he awarded three points for the Giants’ safety. There were no long touchdowns or field goals in the game.
Our third judge gives the game to the Giants by a score of 39-37. I did a quick estimation from the play-by-play, and the Giants won the first quarter (18-5), third quarter (15-13), and fourth quarter (18-6) by this fractured scoring. The Patriots took the second quarter 20-4. You may get slightly different results if you replicate; I did it in a rush, the way someone might if their "rules" are really just their own subjective guidelines.
So the Giants still win, except for one problem: would the game have played out differently if neither team knew the score?
If the Patriots thought they might be winning, or that the game was very close, they would not have attempted the Hail Myra. If the scoring was subjective, they might have played for a field goal. Similarly, the Giants might not have attempted a two-point conversion, which failed. Late-game strategies in this version of football would be a matter of guesswork.
Let’s add one more boxing element to football: the knockout. In this version of football, a Pick-Six or a kickoff return touchdown counts as a knockout. This wrinkle takes some of the wind out of Bill Belichick’s sails if his strategy is to pile on as many points as possible at all times. If he feels he has a commanding lead, he will not want to throw as often, and he will have incentive to slow down the pace of scoring so he does not give the other team kickoff return opportunities. This is like a boxer becoming defensive when he thinks he is well ahead in the scoring. In this style of football, you could imagine teams becoming ultra-conservative when leading against the Bears, a sloppy fighter with an incredibly dangerous left hook.
It would not take much juggling of rules to award the Super Bowl to the Patriots by decision. Or, for that matter, to change the whole complexion of a season by finding games that would have been decided differently if football were scored like boxing (Niners-Saints, for example). The point is not to change the outcomes of games, but to show just how bonkers boxing scoring is. Our Judge C is likely to inappropriately score many games, especially games in which one team completes a lot of passes to catch up to an opponent. Yet it would only take a few flukes for Judge C to slip into the NFL playoffs. I picture him as a top judge in the Big 12, where teams throw regardless of the score, whose high-profile judging in BCS games (and a connection or two) lands him in the NFL. It would take two of him to skew a result from the "true" scoring, anyway, or one of him and some other idiosyncratic scorer whose methods happened to line up with his for a particular game.
Thank heavens football is not like that. No one wants to wait around until five minutes after the Super Bowl to figure out who won. Or to learn that the game turned out completely different from the way you thought it did.
Martin arrived in New York just in time to be reunited with Bill Parcells and run into the ground for a few more seasons. Martin’s 1998 season, in which he averaged just 3.5 yards per carry, registered negative DVOA and DYAR. DVOA often frowns on grinding, high-carry seasons because it is a percentage stat tied to league averages. DYAR usually gives the grinder his due, crediting him for all the mileage he racked up that your replacement-level back could not have produced. By either count, 1998 was a bad year.
Martin had a 30-carry, 42-yard game in 1998. He also had a 20-for-38 and a pair of games in the 22-for-58 range. He was the only running back to carry the ball for the Jets in many of these games. The Jets running game was counterproductive in many of those games. Bill Parcells, bless him, made a career out of supplementing his or Belichick’s defenses with the kinds of offenses that make stat guys like us cringe.
Martin somehow plowed through the Parcells years and earned a slight reprieve when Al Groh and Herm Edwards took over. Martin’s DVOA and DYAR improved, in part because of circumstances, but in part because Martin became a unique running back in history, a Hank Aaron who defied age curves and usage curves. Martin somehow registered his highest carry total (371) at age 31, and recorded the best statistical season of his career to boot. Rising offensive averages helped a little, as did an offensive shift that fit his slashing style.
Martin’s career is backwards. Most great backs build their Hall of Fame resumes in their first six seasons, then accrue hang-around value. Martin hung around and had his butt worked off early in his career, becoming a Hall of Famer between 2001 and 2004, his seventh through tenth years. He would have been a much better per-carry back early in his career if Parcells were not so set in his ways.
McNeil was a Tony Dorsett-caliber runner who only played one full 16-game season in his career because of injuries to his ribs, ankle, elbows, and any other body part you can think of. He was also the lead plaintiff in the court case that struck down Plan B free agency in the mid-1980s. "Plan B" allowed teams to reserve the rights to 37 players per season, meaning that about eight bottom-of-the-roster guys could become free agents. It was essentially Veteran Kicker Free Agency: teams figured they could let their kickers walk and just sign some kicker another team let walk. The league seriously operated under these conditions for three years and called it "progress."
In a 1989 game against the Colts, McNeil blocked low on linebacker O’Brien Alston. Alston’s knee audibly cracked, and the defender was carted off the field. McNeil was so shaken that he blew a few assignments, then took himself out of the game: the thought that he may have ended another man’s playing career weighed so heavily on his mind that he could not concentrate. Coach Joe Walton responded to McNeil’s human emotions by criticizing him to the media. Even after McNeil had a face-to-face meeting with Walton to explain his reaction, Walton reprimanded him and expressed disappointment.
Some of these old coaches can be sons of bitches, can’t they?
3. Emerson Boozer
Boozer was a nifty-shifty back early in his career. After knee surgery in 1968, he became a tough 195-pound slasher known for his ability to run, catch, and block. He shared backfields with Matt Snell and John Riggins, regaining some of his big-play ability in the early 1970s. He was a vintage Other Back from the glory days of two-back offensive strategies.
4. Matt Snell
Snell was the true MVP of Super Bowl III with 161 yards from scrimmage and a touchdown. Traditionalists like to scoff at this blatantly obvious assertion, talking about how Joe Namath called plays and subsumed his ego by handing off to Snell, thereby earning the award. This is a simple example of an ad hoc justification: for Namath and only Namath, "subsuming ego and handing off" is given as criterion for a Most Valuable Player award.
Other arguments take a quasi-DVOA approach by combing through the play-by-play and noting the importance of many of Namath’s 17 completions, an approach which again is applied ad hoc to Namath and does not have any connection whatsoever to how voters actually made their decision (i.e. giving the award to the famous quarterback). Snell had 161 yards from scrimmage and a touchdown in a win. The Jets used a ball control offense, and Snell controlled the ball. It’s open-and-shut stuff, folks.
Snell ranks below Boozer because his best years were concentrated in the mid-1960s AFL, which was a wild and wooly era when the league was not quite on footing with the NFL.
5. John Riggins
Early Riggins was a 1970s fullback who shared carries with Boozer in a two-back offense and caught his share of passes. He was nothing like Famous Riggins, the one-man gang who plowed out 3.something yards per carry behind the greatest offensive line in history a few seasons later. Early Riggins, as discussed weeks ago, was probably a better player than Famous Riggins, though he played in poor conditions instead of ideal ones.
Johnny Hector was McNeill’s partner and stunt double for many years. He was very good. DVOA loved Thomas Jones in 2008 but was lukewarm to his other, similar-looking Jets seasons. Jones is something of a Ricky Watters, and the fact that he now has over 10,000 rushing yards shows just how important it is for us to reset our standards of NFL greatness.
88 comments, Last at 24 Jun 2012, 2:40am by Shattenjager