NewsBot
New Member
- Messages
- 111,281
- Reaction score
- 2,947
After reading Football Outsiders 2017 Almanac, we had a few questions that needed answers.
As we do every year, before the season starts, we take a look at the Almanac produced by the guys at Football Outsiders. You can get a copy here. Once we read it we had a few questions and Mike Tanier, who wrote the Cowboys section, answered. (Why a lifelong Eagles fan is writing the Cowboys section remains a mystery!)
Blogging The Boys: You talked some about the debate of whether Dak Prescott’s sensational season was the result of the Cowboys offensive juggernaut already in place, or did he propel them to new heights. Based on his stats and comparisons to other quarterbacks, which side of the debate do you believe is most true?
Mike Tanier: That’s a needle thread. Not to cop out, but I think that the synergy between Dak’s ability and the support of the system around him is ultimately what is relevant. Great circumstances often develop great quarterbacks. (They also can conceal ordinary or even bad quarterbacks, but I don’t think that’s the case here).
One analogy I would use is Ben Roethlisberger. He led the Steelers to 13 wins as a rookie in 2004 when he did not even really have a second read: it was hit the primary receiver or scramble. The fact that he got exceptional support from the Steelers system isn’t really relevant 13 years later, because that support made his development smoother/quicker/easier/.
BTB: This year, you are projecting the Cowboys with 9.3 wins. And while that’s still the fourth-best mean projection in the league, it is also a 3.7-win drop from their 2016 record, and no other team has a bigger drop. Beyond simple regression to the mean, what factors does your model identify that would cause such a big drop, given that you wrote “there is no good reason to suspect a sophomore slump” from Dak Prescott?
MT: Transition on defense is likely to be a major issue. I think the Cowboys have the potential to be more talented on defense this year than last, but the team is replacing hundreds and hundreds of snaps on a defense which was not particularly strong.
The Cowboys also play a tougher schedule this year, and the NFC East is so strong overall that it tends to turn the divisional games into a .500 mush. Not factored into the calculations, but worth keeping an eye on, is the sheer number of primetime games the Cowboys play, which create some disruptive travel and “weird game time” sequences in the schedule.
There is also the simple matter that projection systems are conservative by nature. We have the Patriots at 11.6 wins. If you look at national predictions, I think most experts would consider 11 or 12 Patriots wins a pretty low total this year. Our season simulations often provide a little more illumination. We give the Cowboys a 33 percent chance of winning 11-plus games. That’s pretty great.
(Also: all of the calculations were done before the Ezekiel Elliott suspension. I am guessing Aaron will run separate projections for six games without Zeke. I am also guessing that there will be so many appeals and legal wranglings that this suspension might not be served until 2019, if at all).
BTB: The DVOA and DYAR rankings of Jason Witten have dropped off significantly in the last couple of years, while his raw numbers suggest he's still quite capable. Is he at the point yet where the Cowboys need to look hard for a replacement? Would his rankings be helped at all if he played something less than all the time?
MT: All of the loyal Cowboys fans reading this know who Rico Gathers is and saw his Hall of Fame Game touchdown. I am sure they also recognize that he is a project and a work in progress.
Our metrics show that Witten now does most of his damage on super-short passes (46 percent of his targets are now within 5 yards of the line of scrimmage) and on first-down play action (there are some nifty breakdowns in Witten’s player comment in the book).
Basically, Witten has gone from being an all-purpose receiver at tight end to the more traditional guy who catches the play-action waggle. It’s not a matter of resting him or limiting his snaps. He’s just getting old. That said, his value as a blocker makes him worth keeping in the game. Witten’s omnipresence disguises some of the team’s tendencies -- the Cowboys can execute nearly anything with him next to the right tackle -- which also helps him retain some value.
BTB: How would a fully healthy Jaylon Smith change your assessment of the 2017 defense, if at all?
MT: At linebacker (as opposed to QB or left tackle), a player like Smith would have to be projected at roughly an All Pro level to really move the needle. I think all of us are optimistic that Smith can be a great player, but right now he is just another young player with no track record (and not even first-round pick status) projected to play a role.
BTB: What was the biggest surprise you found when doing the preview for the 2017 Cowboys?
MT: I was amazed at just how clearly the stat breakdowns tell the story of the Cowboys offense. Readers will have to check out the chapter for details, but it boils down to the fact that being able to run for five or six yards consistently on first downs created great second-and-short opportunities, great play action opportunities, and minimal third-and-long situations, which of course allowed the Cowboys to keep running the ball, which set up even better play action, which made everything easier for Dak. The chapter may be interpreted as HERE IS PROOF THAT DAK HAD IT EASY or something, but that is not the case. It’s a snapshot of how most NFL coaches, particularly run-oriented old-fashioned coaches, dream that their offenses should operate.
BTB Bonus Question: There were 15 teams with nine or more wins last year. Every single one of them is projected for a lower win total in 2017. There were 13 teams with less than eight wins last year. Every single one of them is projected for a higher win total in 2017. Is that just an oddity or is regression to the mean the single most important factor in the projections?
MT: To be clear, there is no special “regression” variable or weighting. The preponderance of hundreds of variables just tend to force extreme highs and lows back toward the middle, as they do in just about any projection system.
Our system often projects MORE surprises than conventional wisdom or the “poll of experts” who make season predictions usually project. If you look at expert picks across the media landscape, you will probably see tons of “chalk:” not just the Patriots winning the East, but teams like the Steelers, Seahawks, and Packers chiseled in as division champs, the Jets/Jaguars/Browns in their familiar cellars, and so on. DVOA projections in past years served up many more surprise risers and fallers than the polls did.
We have the Texans, a regular division winner, at 6.7 wins this year. We have the Panthers bouncing back at 8.6, highest in the NFC South. We also have a possible playoff bounceback for the Cardinals and some other surprises. The projections are not as wild and wooly as in years past but still contain some revelations.
Continue reading...