Possibly the closest Super Bowl matchup in history also poses the question: how much does it mean when certain aspects of an NFL team improve dramatically in the second half of the season?
15 Jul 2005
by Aaron Schatz
Also check out the previous edition of Four Downs: AFC East.
Travis Henry is going to Seattle. Travis Henry is going to Jacksonville. Travis Henry is going to Tennessee. Travis Henry is going to the Red Sox for Kevin Millar and he's going to the Knicks for Stephon Marbury. We don't know which one of those trade rumors is going to prove true, but everyone is pretty sure that Travis "Little Hands of Concrete" Henry is leaving Buffalo.
What does the future hold for Mr. Henry? Let's take a look at the most similar players since 1978 using our system of similarity scores (explained here). First, here are Henry's conventional numbers for the past three seasons:
Because Henry had such a huge drop-off in usage between 2003 and 2004, the list of similar players changes dramatically depending on whether you list the most similar players over one season, two seasons, or three seasons. The list of similar single seasons has plenty of players who never came close to 1,350 yards; the list of similar two-year stretches features a number of players whose 1,300-yard season was a one-time event, not a repeat from the previous year; the list of similar three-year stretches isn't necessarily that similar, as you get a lot of players who were similar the first two years and not quite as bad in year three.
Let's start with similar players in a single season. One of these guys is pretty familiar to Buffalo fans, and he also switched teams the next season:
"SIM" is the similarity score -- the closer to 1000, the more similar. (When we do multiple years, it is the harmonic mean of two or three similarity scores.)
If Henry wants to prove to people that he can still be a starter somewhere, he's got a lot of good evidence on this list, starting with Antowain Smith. Like Henry, Smith was a former starter sent to the bench, but his previous seasons weren't similar to Henry's -- they were worse. He had gone from 1,124 yards to 614 to 354, and 2000 was Smith's third straight season below 3.75 yards per carry. The Bills released him, the Patriots signed him, and a year later he was coming off a career-best 1,157 yards and a Super Bowl championship. Smith's not the only player on this list who had a 1,000-yard season the next year; Cleveland Gary did as well. Gary Brown lost 1996 to injury but had 945 yards for the 1997 Chargers and 1063 yards for the 1998 Giants. On the other hand, Reggie Brooks never played again, and Carson Kressley will have a 1,000-yard rushing season before Travis Minor ever does.
Next, the most similar players over a two-year span. For space purposes, we'll leave off receiving numbers here, but they are reflected in the similarity computations. Age reflects the more recent season.
|First Year||Second Year|
The first thing that should jump out at you is this question: "How on earth do you rush for 1185 yards and only score three touchdowns?" Erric Pegram was playing on Jerry Glanville's last run-n-shoot team, and the 1993 Falcons had 28 passing TDs but only 4 rushing TDs. The 1992 team was even worse, with 33 passing TDs and only 3 rushing TDs. Pegram had just 89 yards rushing that season, and his career wasn't much like Henry's. He was a backup who had one big year, and then reverted to his previous backup status. He did change teams in 1995, like Henry probably will, and he put up 813 yards sharing time with Bam Morris for the AFC Champion Steelers.
Levens is more like Pegram than he is like Henry -- his career consists of one huge season (1997) and one moderately good season two years later (1,034 yards in 1999). The 1998 season similar to Henry's 2004 was caused by injury, not benching.
Errict Rhett is Henry's worst nightmare. He was a two-year starter in Tampa Bay, but like Henry he tallied up lots of yards and touchdowns because he carried the ball so much, not because he was very good at it. In each of his 1,000-yard seasons, he had just 3.6 yards per carry. He spent the first half of 1996 in an extended contract holdout and when he came back he was horrible, with 3.1 yards per carry, and had to share the load with a rookie named Mike Alstott. Then the Bucs drafted Warrick Dunn to compliment Alstott and that was pretty much it for Errict Rhett.
Both Rhett and Mike Rozier show up again on the list of the top three-year comparables:
|First Year||Second Year||Third Year|
As you can see from those similarity numbers, we're stretching things at this point, because Henry's career path is abnormal. Henry has only one player with a three-year similarity above 815, and to get even that one guy we had to pro-rate stats from the 1987 strike year. By comparison, Shaun Alexander 2002-2004 has 21 players with a three-year similarity above 815. (Alexander's top three, for the curious: Curt Warner 1985-87, Tony Dorsett 1979-81, and Walter Payton 1978-80.)
Is Corey Dillon similar to Henry? Well, each player lost his job to a younger teammate and then became trade bait. But Dillon was three years older when he lost his job to Rudi Johnson (the age difference is reflected in the similarity score) and was traded into the perfect situation, a defending champion with only one hole on offense: running back. Unless Ryan Moats and Correll Buckhalter both get injured and Philly is desperate for someone to match with Brian Westbrook, Henry's out of luck.
Duce Staley comes out as similar to Henry just because nobody else does. His bad third year was caused by injury, not benching, and while the rushing numbers are similar, the receiving numbers are very different, marking Staley as a very different kind of player.
No, Henry's future is likely closer to Rozier and Rhett Rozier had two average years as part of a committee in Atlanta and then retired. Rhett piddled around, got one more year as a starter in Baltimore in 1999, played at replacement level (3.8 DPAR), and was replaced. Greg Bell, by the way, never played again -- he is listed as "LA" because he had two big years with the Rams, was kicked to the curb in favor of Cleveland Gary, moved across town to the Raiders for a year as a backup, and was gone.
The best statement about Henry doesn't come from similarity scores, but from our advanced metrics which take into account opponent and situation. Here are Henry's rushing DVOA and DPAR for the past four years (those stats are explained here) along with rank among backs with at least 75 carries:
Dr. Z says that Henry is "a terrific runner," but I politely object. I understand not wanting to pay Shaun Alexander when there's a market glut for running backs, but if the Seahawks are going to give up a draft pick for this instead of playing Maurice Morris, they're crazy.
In the last Four Downs, I pointed out that the Dolphins have spent the off-season collecting defensive linemen. Not yet satisfied with their haul, the Dolphins selected USC defensive tackle Manuel Wright in the fifth round of the supplemental draft yesterday.
This move, however, is difficult to criticize. The Dolphins haven't just been collecting defensive linemen -- for the most part, they've been collecting old defensive linemen. If they do move to a 3-4 alignment, they will likely be starting 36-year-old ex-Patriot Keith "I Was Not Traded For Dirk Nowitzki" Traylor at nose tackle. Kevin Carter will be 32, Vonnie Holliday 30, Larry Chester 30, and Jeff Zgonina roughly 137. Wright was considered a potential first-round choice if he had stayed at USC for two more years, probably a second-rounder if he stayed for one more year. It's not such a bad thing to take a player with that much talent, especially when he costs just a fifth-round pick and a roster spot for the next couple years. From what I've heard, Nick Saban isn't so bad at that teaching and development thing.
On the other hand, Wright is yet another defensive player added to a team that already had a good defense, while the offense has received less attention than the last Macy Gray record. They have Ronnie Brown now, and Ricky Williams coming back -- although as we've said numerous times, it is doubtful Williams will really make a difference. They apparently are going to give the starting quarterback job to Gus Frerotte, which doesn't seem like a huge step forward. The major problem with this team is offensive line, and with only one change -- free agent tackle Stockar McDougle -- the Dolphins are counting on new offensive line coach Hudson Houck to make lemonade out of the excrement of cows that may have eaten lemons last week.
Will Dolphins fans have something to cheer in 2005, rather than a season spent looking to the future? But no matter how much work Nick Saban does with the defense, the offense has to improve for the Dolphins to become contenders again. It doesn't have to be good, but it has to be average. Last year's Dolphins had a DVOA of -28.5%, worse than every team except Chicago.
We have DVOA ratings from 1998-2004, and in that time eight teams increased their offensive DVOA from one year to the next by 20%, relative to the league. (The baseline for DVOA is based on multiple seasons, so because the league's offensive environment changes slightly each year, the average DVOA for a specific season isn't necessarily 0%.) For the Dolphins to have hope of competing this year, they'll need to have an offensive turnaround similar to one of these teams.
1998 Raiders: -27.7% DVOA (29th)
1999 Raiders: +15.9% DVOA (2nd)
Reason for Improvement: The Rams won the Super Bowl, but the Raiders were actually the team which had the biggest offensive surge in 1999. The offensive personnel were almost entirely the same both years, with one major exception: at quarterback, Oakland switched from Jeff George and Donald Hollas to Rich Gannon. With this newly potent offense, the Raiders improved their record from 8-8 to ... wait for it ... 8-8. These may be the two most deceiving 8-8 records of all time. The 8-8 Raiders of 1998 were outscored by their opponents 356-288. The 8-8 Raiders of 1999, however, outscored their opponents 390-329. That's a 129-point switch with no additional wins. When the Raiders went 12-4 in 2000, it should have surprised nobody.
Similarity to 2005 Dolphins: Unless Gus Frerotte is Rich Gannon, none.
1998 Rams: -17.8% DVOA (26th)
1999 Rams: +18.3% DVOA (1st)
Reason for Improvement: You all know this one. The Rams traded for Marshall Faulk, drafted Torry Holt, finally got a healthy year from Isaac Bruce, and found Kurt Warner's space pod in a Kansas cornfield. I'll have much more to say about this team when we finally run articles with the full 1998 and 1999 numbers over the next few weeks.
Similarity to 2005 Dolphins: Yeah, right.
2000 Chargers: -39.0% DVOA (31st)
2001 Chargers: -3.8% DVOA (16th)
Reason for Improvement: The Chargers replaced Ryan Leaf and Terrell Fletcher with Doug Flutie and LaDainian Tomlinson. The Chargers outscored opponents by 332-321 and still managed to go 5-11.
Similarity to 2005 Dolphins: Pretty good, actually. This is the move the Dolphins would need to make to be in the playoffs -- not from horrible to good, but from horrible to average, letting the defense win games. The Dolphins, like the 2001 Chargers, replaced a committee of mediocre veterans with the highest-drafted rookie back. The harder move to duplicate is at quarterback. As bad as Fiedler and Feeley were last year, they weren't as bad as Ryan Leaf, and Gus Frerotte is no Flutie.
2002 Texans: -38.0% DVOA (32nd)
2003 Texans: -11.0% DVOA (24th)
Reason for Improvement: David Carr improves, Andre Johnson and Domanick Davis arrive.
Similarity to 2005 Dolphins: Miami's offense may be expansion-level bad, but it isn't expansion-level young, so no.
1999 Eagles: -31.9% DVOA (30th)
2000 Eagles: -1.8% DVOA (16th)
Reason for Improvement: Donovan McNabb's first full year as starting quarterback, maturing offensive line.
Similarity to 2005 Dolphins: No, this team is the exact opposite, with a great quarterback and offensive line to go with no running game and a receiving corps that didn't have anyone with talent even close to Chris Chambers.
1998 Panthers: -16.8% DVOA (24th)
1999 Panthers: +4.8% DVOA (10th)
Reason for Improvement: George Seifert arrives as head coach and brings a San Francisco-style offense which drives Steve Beuerlein to an inexplicable career year, with 36 touchdown passes and 4436 passing yards (previous career highs were 18 TDs, 3164 yards). Tim Biakabatuka has half of an inexplicable career year, averaging 5.2 yards per carry when not injured. Every position on the offensive line except left guard turns over.
Similarity to 2005 Dolphins: None, I think. You have to forgive me, as I wasn't paying much attention to the NFC back in those days, but looking back this season seems to make no sense whatsoever. I'd love to hear from some Carolina fans who can explain what the heck happened in 1999 and then what happened in 2000 to make it stop.
2003 Patriots: -0.8% DVOA (13th)
2004 Patriots: 26.3% DVOA (4th)
Reason for Improvement: Patriots finally get a running game.
Similarity to 2005 Dolphins: This is an offense that needed one last piece to go from average to spectacular, not lots of pieces to go from horrible to average, so no.
2003 Steelers: -8.4% DVOA (19th)
2004 Steelers: 16.6% DVOA (7th)
Reason for Improvement: Offensive line was wracked by injuries in 2003 but healthy in 2004 (except RG Kendall Simmons, but he went out early enough in training camp that his replacement Keydrick Vincent was fully ready when the season began). Jerome Bettis discovers fountain of youth. Highly efficient rookie quarterback.
Similarity to 2005 Dolphins: Yes, in that an offensive line that was horrible one year became spectacular the next. But the Pittsburgh linemen had been good for years. The aberration was the bad 2003, not the good 2004. That's not the case in Miami. I don't think there's a Ben Roethlisberger on this roster, either. On the other hand, we thought Jerome Bettis was toast, and his 2004 season certainly gives hope to other running backs trying to come back from declines caused by high usage, Ricky Williams included. Especially if Ronnie Brown can be Duce Staley.
The NFL is a copycat league, or so the cliche goes, and 31 teams are constantly trying to copy whatever strategy helped the 32nd team win the Lombardi Trophy. That's certainly the case with the New England Patriots, winners of three of the past four Super Bowls.
A number of NFL trends can be tied, at least in part, to a desire to do things the Patriot way: defensive coordinators flirting with the 3-4 alignment, front offices cutting costs on the offensive line (except at left tackle), and coaches using defensive players on offense to create size mismatches (Julius Peppers as a wide receiver, for example). But one of the most important aspects of the Patriot dynasty is also the most difficult to copy: New England's outstanding depth.
Most NFL teams will begin the 2005 season believing they can contend for the Super Bowl if they can just avoid injuries. But that's just not good enough in the NFL. A well-built team must be able to excel even after suffering numerous injuries. No team has demonstrated this over the past two years more than the Patriots.
In 2003, New England started more than 40 different players, the most of any Super Bowl team in history. In 2004, they started 39 different players. Both starting cornerbacks were lost for most of the season; number one receiver Deion Branch missed half the year, as did starting right tackle Tom Ashworth; Pro Bowl defensive lineman Richard Seymour missed the first two playoff games. And yet coach Bill Belichick was able to fill every one of these holes with a competent backup, even if it meant using wide receiver Troy Brown as a nickelback.
Not content to rely on such stopgap solutions, the Patriots have spent the off-season signing even more veterans to create their deepest roster yet. Most teams fill out the training-camp depth chart with undrafted rookies, but that's not the case in New England.
For example, all eight defensive backs who played in last year's Super Bowl are still on the team. So are two players who lost most or all of last season to knee problems: Tyrone Poole, who began the year as a starting cornerback, and Guss Scott, a third-round pick who spent his rookie year on injured reserve.
To these 10 players, the Patriots added three veteran defensive backs who were starters for other teams for at least half of 2004 -- cornerback Duane Starks came in a trade, cornerback Chad Scott and safety Antuan Edwards in free agency. The Patriots also spent a third-round draft pick on cornerback Ellis Hobbs and a fourth-round pick on safety James Sanders.
As a result, the Patriots will bring to camp 14 defensive backs who either have NFL experience or were 2005 draft picks -- 15 if you include Brown. Nine of those defensive backs have at least two years' experience.
Wide receiver is another example. David Patten's departure for Washington opened one spot on the depth chart, but the Patriots signed two free agent veterans, David Terrell and Tim Dwight. They also have 2004 fifth-round pick P.K. Sam, who spent most of his first season inactive but was specifically drafted as a developmental project. The Patriots even have three different players who were used as an NFL team's primary kick returner last season: holdover Bethel Johnson and free-agent signings Dwight and Chad Morton.
Where are all of these players going to fit? The answer is that they aren't. The Patriots have moved on to the next logical step. Not content to withstand injuries during the season, they have built up so much depth that they can withstand the inevitable injuries that come during training camp.
If a receiver or defensive back gets hurt during a preseason game, that spot on the depth chart will be taken by an experienced veteran, not undrafted rookie filler. And if nobody gets hurt during the preseason, the Patriots will just cut one or two of these veterans and escape relatively unharmed. Morton's signing bonus was just $60,000, so that's all the Patriots lose if he does not make the team. Dwight and Edwards received no signing bonuses, and cutting them before the season will cost the Patriots nothing (Scott did not receive a signing bonus either, but some of his salary is guaranteed).
The Patriots aren't taking more players to camp than other teams; they're just taking more players who are more than just fodder for the fourth-quarter of preseason games. It also doesn't hurt that the Patriots prefer to offer good contracts to numerous quality veterans rather than a few gargantuan contracts to a handful of superstars, leaving very little money to fill out the depth chart.
Can other teams recruit a similarly deep pool of veteran backups? Some might, but not all can, because depth is a zero-sum game. Every veteran going to camp with New England also means one less veteran available to other teams. For example, although New Orleans offered Brown more money in free agency, the versatile veteran decided to return to the Patriots (an endorsement deal with a local bank, it turns out, will make up most of the salary difference). That meant that the Saints had to turn to their second option, free agent Az Hakim, which meant in turn that the Kansas City Chiefs -- who wanted Hakim -- had to turn to their second option, the erratic Freddie Mitchell.
Of course, it's not enough to simply accumulate veteran players; teams have to know how to use those players in a way that masks their declining skills and accentuates their remaining strengths. Baseball writer Bill James used to say that the best managers were the ones who concentrated on what players could do rather than what they couldn't do. Earl Weaver would take a guy like John Lowenstein, who nobody wanted because he couldn't hit lefties, and pair him with a kid who could hit lefties (Gary Roenicke) to build one important piece of a World Series champion. Belichick does the same thing with his football team, juggling multiple players at each position and using each one in the right situations.
Conventional wisdom says that the Jets went 10-6 last year thanks to a great defense and their running game, so all eyes this off-season have been on the passing game and special teams. The big name acquisitions are a wide receiver, a tight end, a kicker, and a new offensive coordinator, and the biggest question mark of training camp revolves around the shoulder of the quarterback.
But as we've pointed out numerous times, and will explain further in our upcoming book Pro Football Prospectus 2005, conventional wisdom is wrong. The Jets did not have a good defense. They had a defense that was on the field for fewer plays thanks to the team's conservative offensive philosophy, and the advantage of a schedule filled with poor quarterbacks. Going into this off-season, the biggest problem for the Jets wasn't finding a wide receiver or a kicker; it was finding a professional secondary to play behind their talented front seven so that teams couldn't just pass on them all day. Check out the team's DVOA against different types of receivers, and you can see how the linebackers keep the New York passing defense from looking completely porous:
|Year||vs. #1 WR||Rank||vs. #2 WR||Rank||vs. Other WR||Rank||vs. TE||Rank||vs. RB||Rank|
With two weeks to go until training camp begins, the secondary is still unsettled. In fact, it became even more unsettled yesterday when starting cornerback Donnie Abraham retired. David Barrett is a nickel-quality back who was severely stretched last year as the number two corner. Now the Jets are stuck with him as the number one corner. Who will start at the other corner? Justin Miller fell to the Jets in the second round of the draft, but rookie cornerbacks generally have a steep learning curve (just ask the Packers). The other candidate is Pete Hunter, who the Jets acquired yesterday from the Cowboys for a conditional draft pick. Hunter finally became a starter last year, his third season in the league, only to blow out his ACL in Week 3. This year the Cowboys wanted him to move to safety, which isn't exactly the biggest vote of confidence in his coverage abilities.
Then again, there is a third candidate -- the one star cornerback still remaining on the free agent market, Ty Law. Law hasn't signed because no team has been willing to pony up the money that matches his resume, and no team is willing to pony up that money until their doctors have made sure that they're getting the same cornerback who put that resume together. Abraham's retirement is the best thing that could have happened to Law, because the supply of number one cornerbacks is one, and the demand in New York just shot through the roof. Law would like to be able to play against the Patriots twice a year, to show them up for giving up on him, but do the Jets have the cap space? According to Askthecommish.com, the Jets are in the bottom five in available cap space (as of May 21).
How about safeties? Free safety Erik Coleman may have been the best second-day draft pick of 2004, and he's the one solid piece in the secondary. But as good as his rookie season was, he did have a weakness that became clear near the end of the season. According to our new tape-analyzing buddy K.C. Joyner and his book Scientific Football 2005, Coleman was one of the top safeties at preventing complete passes deep, but he had a problem with late reaction time, so when deep passes were completed against him, they went for huge yardage. This was particularly a problem in the final two games of the regular season against David Givens (29- and 35-yard gains), Kevin Curtis (a 34-yard gain), and Shaun McDonald (a 33-yard gain).
Who will start next to Coleman, after last year's starter Reggie Tongue was waived? I decided to ask someone who knows the Jets better than I do, Jets blogger Brian Bassett. Before he went off to get married (congratulations, Brian!) he sent me these notes:
I think that Jon McGraw will get the nod for the starting job. He has battled injuries over the past few years, but it looks like the staff is ready to give him a shot. I think it is his job to lose.
Fifth-round pick Andre Maddox from North Carolina State is a confident player (thinks he is better than Brodney Poole) and has been known for jarring hits and his ability to tackle, so I think he would show promise in a limited run-stopping role for now, but it looks like he will be playing gunner on special teams initially, and the Jets will expand his role from there.
The big question mark is fourth-rounder Kerry Rhodes from Louisville. He was a quarterback and transitioned to safety in college. Apparently, he has great athletic abilities, but he has some rough edges that he was rounding off in college. He's more aggressive in coverage, the opposite of Maddox. The Jets traded up to get Rhodes, so they see the promise. Rhodes is a rangier player than Maddox, and Herm is big on players with the ability to close on the ball in the air, because he says this is an innate skill that is hard to teach.
It sounds like the rookies are pretty raw, which doesn't help a team that needs a couple of players to step up. Even if the Jets can give our friend Brian a big wedding gift by signing Law, they still have a defensive hole at the other corner that is easy to exploit unless Hunter or Miller can play at a high level and replace Barrett. This hole is an even bigger problem if Coleman still makes the occasional mistake reacting late on a deep route. As Joyner writes, "He has good pass recognition overall, but all it takes is a few times when you don't react quickly enough and you'll be beaten deep."
If Law joins the Jets, it will cause big problems for their opponents who depend on one great receiver, such as Miami, Jacksonville, and New Orleans. But I foresee some big fantasy football performances against New York by strong number two receivers, particularly those who like to go deep: Jerry Porter or Ronald Curry, David Givens or David Terrell, Joey Galloway, Ashley Lelie, and above all Lee Evans.
Note: An edited version of the New England section appeared in Tuesday's New York Sun.
Next week: NFC East by Al Bogdan.
55 comments, Last at 12 Aug 2005, 1:48pm by Led