Every year, Blogging The Boys gets an opportunity to peruse the annual Football Outsiders Almanac from the fine people over at Football Outsiders. 2016 was no different. The bottom line for the Cowboys is a predicted 9-win season (technically 8.8 wins). They also give Dallas a 50% chance at hitting the postseason. For comparison, New York is projected for 7.8 wins and a 34% postseason chance, Philadelphia is 6.9 wins with a 23% chance, and Washington comes in at 6.8 wins and 21% chance of a postseason.
In addition to digesting all the information in the Almanac, you can get one here, we also ask five specific questions. This year Sterling Xie answered them and the results are below.
Blogging The Boys: How much of a correlation is there between winning Time of Possession on offense, and a team’s defensive success. That has been an argument for the Cowboys emphasis on the running game, but do stats show this to be true, that winning TOP is good for the defense?
Football Outsiders: It's definitely not as strong a correlation as you might expect. The idea makes sense on paper: If your defense is shaky, it makes sense to keep them on the sidelines as long as possible with an offensive approach that naturally shortens the game. But last season, the correlation between time of possession per drive and defensive DVOA was just 0.03. That means there's essentially no correlation at all between the two. The Cowboys themselves should serve as a prime example -- Dallas' offense ranked third in TOP/drive last season (3:00 exactly), but it's not as if that helped the defense much in the end.
The more critical factor for the Cowboys will be how often their defense gets to play with a lead, not how long they're on the field. This was the key difference between the 2014 and 2015 offensive units: While both ate up sizable chunks of clock, the one with a healthy Tony Romo and Dez Bryant was also successful at putting up enough points on the board to increase the defense's margin of error. So the real place you want to focus on is defensive DVOA vs. average lead. The correlation between those two stats in 2015 was -0.60 (remember, more negative DVOAs indicate better defensive play. So this suggests good defenses last year typically got to play with a lead). For the Cowboys, keeping their defense off the field isn't as important as keeping the opposing offense in a bind.
Shoutout to our own Jim Scott, who put together a stellar analysis of this phenomenon earlier this offseason!
BTB: What was the biggest surprise you found when doing the preview for the 2016 Cowboys?
FO: The biggest surprise was how spectacularly unlucky the Cowboys were, and I'm not even referring to the Romo and Bryant injuries. That definitely represents a type of bad luck, but important players get hurt every year, and at this stage of Romo's career, assuming he'll stay upright the whole season is dangerous.
However, Dallas' turnover luck was quite literally as bad as you'll ever see. The Cowboys finished last in turnovers per possession on offense AND defense. The latter was especially snakebitten: Dating back to 1997, the 2015 Cowboys defense had the fourth-worst turnover per drive rate in our entire database. Doing some quick math, that means the 2015 Dallas defense ranked 598th out of 601 defenses over the past 19 seasons at taking the ball away.
Forcing and preventing turnovers isn't entirely luck-based, but turnover margin does tend to regress to the mean between seasons. The Cowboys weren't good last year, but they weren't 4-12 bad either.
BTB: You’re forecasting 8.8 wins for the Cowboys. Can you briefly explain how you arrived at that projection, and what are the Cowboys’ best- and worst-case scenarios for 2016?
FO: The way we come up with our DVOA and win-loss projections is actually explained on page xviii of the introduction. For win projections, the basic gist is that we come up with a range of possibilities based on DVOA over the past three seasons, quarterback performance, and the value of offseason additions and subtractions. Since it's a mean projection system, it'll naturally tend to cluster around 8-8. No team has a higher projection than 10.5 wins or lower than 5.2 wins. Based on wins, the Cowboys project as the No. 3 seed in the NFC, behind Seattle and Green Bay and ever so slightly ahead of Carolina. However, the Cowboys rank sixth in the NFC in projected DVOA, behind those three teams plus the Cardinals and Vikings.
As for best and worst case scenarios, I think we've seen an example of each the last two seasons, haven't we? Dallas probably won't be picking in the top five again next April, but everything will need to break right for the Cowboys to become Super Bowl contenders like they were two seasons ago. Our projections found a middle ground which leans closer to 2014 than 2015.
BTB: Your numbers show that the Cowboys had one of the lowest play action percentages in the league in 2015, when they ranked 30th with 15% play action. And that's not just a function of missing Tony Romo, the Cowboys have been at the very bottom of the league for the last half dozen or so years, even when DeMarco Murray was running opponents into the ground in 2014. Any explanation for that?
FO: I would think Jason Garrett's offensive system provides the best explanation for that. While the Cowboys have attempted to provide Romo more support from the running game as he ages, they haven't exactly changed their passing principles when he does drop back. Romo's deep ball percentage has always lingered at around 10 to 12 percent based on our tracking data. Dallas was a pretty successful play-action offense in 2014 (38.6%, ninth overall), but that was likely just the product of an excellent offense rather than anything to do with play-action specifically.
BTB: The Cowboys had an above-average defense on second and third downs, ranking 10th and 13th, respectively. However, they ranked 31st on first downs. Also, the defense was extremely good in the first half of games in 2015, extremely poor in the second half, in almost every category. Is there an explanation for these discrepancies?
FO: These types of splits tend to be more of a one-year blip than anything. They help explain why things happened the way they did in 2015, but it's not necessarily a good idea to read them as indicative of anything to come in 2016. Take the first half-second half split for the defense. In 2014, it was reversed: The Cowboys ranked 27th in first half defensive DVOA but 15th in the second half. What was so materially different about the Dallas defense between those two seasons?
Most teams have weird splits like this somewhere, and I would suspect a defense like Dallas', which profiles as a middle-of-the-road unit, will see splits like this crop up again in 2016. As alluded to in Question 2, though, creating more turnovers is probably the primary objective for the Cowboys defense this season.
Thanks to Football Outsiders for the knowledge.