2017 Patriots Defense
- Thread starter Greekca
- Start date
Well, they started the season, over the first four weeks, with the worst defense in NFL history by the numbers. They ended up 5th in the league in points allowed, which is remarkable given where they started. Clearly this defense is not the same as it was those first four weeks.
First Four Weeks: 32.0 points, 456.8 yards, 24.3 first downs
Last 12 Weeks: 14.0 points, 335.8 yards, 19.0 first downs
Last Four Weeks: 18.3 points, 337.0 yards, 18.3 first downs
The last four weeks included two touchdowns scored by opposing defenses, so that scoring average really is more like 14.8 points a game.
And it's not like the offense has been helping the defense more the past 12 weeks, because they haven't. The offense has performed WORSE.
First Four Weeks: 32.3 points, 423.8 yards, 24.8 first downs
Last 12 Weeks: 27.4 points, 384.3 yards, 22.8 first downs
So the offense is still doing its job, but it's not carrying the team like it was the first four weeks. It's played worse, which has made the defense's job harder. And yet they've played 7 out of 12 games on the road and performed significantly better than in those first four weeks.
And I was discouraged looking at the last four weeks until I remembered that there were two defensive touchdowns scored by the other team that make the scoring numbers look much worse.
First Four Weeks: 32.0 points, 456.8 yards, 24.3 first downs
Last 12 Weeks: 14.0 points, 335.8 yards, 19.0 first downs
Last Four Weeks: 18.3 points, 337.0 yards, 18.3 first downs
The last four weeks included two touchdowns scored by opposing defenses, so that scoring average really is more like 14.8 points a game.
And it's not like the offense has been helping the defense more the past 12 weeks, because they haven't. The offense has performed WORSE.
First Four Weeks: 32.3 points, 423.8 yards, 24.8 first downs
Last 12 Weeks: 27.4 points, 384.3 yards, 22.8 first downs
So the offense is still doing its job, but it's not carrying the team like it was the first four weeks. It's played worse, which has made the defense's job harder. And yet they've played 7 out of 12 games on the road and performed significantly better than in those first four weeks.
And I was discouraged looking at the last four weeks until I remembered that there were two defensive touchdowns scored by the other team that make the scoring numbers look much worse.
5th in points allowed, but 29th in yards allowed -- I don't think I've ever seen a ranking spread like that.
How did this happen?
- Some of this is likely due to special teams (1st in opponent's average drive start).
- Turnover differential (18-12) helped a little
- There were definitely times when the opponent passed up a late FG opportunity because they needed a TD. But five or six more FG allowed would only drop them from 5th to 7th in PA
- What else??
Combination of a lot of things....
- They give up a lot of yards but stiffen in the red zone. The Pats are 4th in the NFL in opponent red zone TD percentage (43.8%).
- They don't get many turnovers, so it takes them a while usually to get a stop on downs. The Pats are 21st in the NFL in opponent 3rd down conversion percentage (39.4%). And they're 25th in takeaways.
- Opponents have missed a LOT of field goals against the Pats. Opponents have tried an average of 1.9 FG per game against the Pats, but have made just 1.4 per game. That right there is an average of 1.5 points per game. Now nobody is going to have their opponents make every field goal, but the Pats are #2 in the NFL in opposing FG% at 70.1%. Houston's opponents, in contrast, made 94.3% of their FG attempts this year. And there seems to be no correlation between quality of the defense and opponents' FG%. For example, Jacksonville, one of the best defenses in the NFL this year, finished 31st in opp FG% (91.7%).
So some of it is luck - the Steeler and Bills and first Jets' games show how close they were to yielding at least 21 more points, and the missed FGs are mainly just random. I really would like to see this defense either generate more turnovers (McCourty missed an easy red zone INT yesterday, for example) or stiffen more on third down. Because I don't think you can expect the luck to hold up all playoffs long.
But a lot of it is scheme - they must do SOMETHING different to be tougher in the red zone.
I see myself talking about this a lot over the next few weeks. I don't think the missed field goals is a byproduct of "luck." Something people consistently fail to think about when it comes to the NFL and statistics is that not every game is played on a neutral field in a dome. If there is one thing that affects kicking more than any other, it's weather. Kicking field goals in New England, Buffalo, New York is significantly more difficult in November onwards than it is in Jacksonville, Houston, or New Orleans or Miami, etc. The weather, the weather, the weather. It explains a lot of things when looking at the numbers that aren't otherwise explainable. There is a reason why we've basically had two kickers here over the past 20 years, and why a guy like Ryan Allen (who many think suck) had one of the best games of the season yesterday in brutal punting conditions. The Pats offense has "struggled" the past 6 weeks, but they are playing games in cold rain storms, negative wind chills, etc.
This is why it is such a huge advantage to play at home in January. Nobody plays better in every type of weather than Tom Brady, and BB coached teams. Shit, Brady's first ever playoff game was the Snow Bowl, so it's probably a bit of destiny. We watched Peyton Manning come in here almost year after year and shit his pants in the playoffs because he couldn't get it done in the cold. The bottom line is it's so hard to extrapolate conclusions from the numbers without going back and actually looking at the conditions at the time of the game (or even the actual play, given that wind and weather can change by the minute during a game). Were guys kicking into the wind, with the wind, etc. It's also why Ghost has so few attempts from 50 plus yards in his career. BB almost never, ever trots him out there to kick a 50+ yarder at home unless it's the end of the half, or game, etc. If it's 4th and short inside our opponents 35, we go for it, if it's 4th and long, we punt, etc.
Anyway, I'm not saying the weather explains everything, far from it, but I feel like it's constantly discounted or ignored altogether in way too many conversations when it comes to NFL statistics. Houston's opponents have made 94% of their kicks, and I would guess it's in large part due to the fact that since their game on November 5th (when the weather starts to kick up around the country), they've played 6 of their 9 games indoors. The other three were at Jacksonville, at Baltimore and at Tennessee, and the last of those games was on December 3rd against Tennessee (unsurprisingly, the Titans missed a field goal in that game and Houston missed two of them).
There is obviously a reason the Pats consistently are among the league leaders in yards given up and fewest points as B.B. long ago determined that that is what wins games. Having your opponent “waste” energy during inefficient drives while your offense “rests” works. If it was all 3 and outs on D there would be a lot more opportunities for injuries on offense
Also, I disagree on the luck factor in FGs as the previous poster mentions the conditions in NE are much different than most other stadiums.
Also, I disagree on the luck factor in FGs as the previous poster mentions the conditions in NE are much different than most other stadiums.
One of the factors that opponents points are down and have always been down relative to the yardage given up in the BB / TB12 era is the pressure exerted by the Pats offense. They force teams to forgo field goal attempts in order to score TDs, either by virtue of the score / time left in the game, or just by the fact that the Pats offense is so good. If you are down 10-0 and have 4th and 1 from the 20 or whatever, a lot of teams would kick a field goal under normal conditions to make it a one score game. Against Brady, if you kick that field goal then 10-3 can become 17-3 in a heartbeat and you are in danger of getting blown off the field. I would be curious if there is any way to back this up statistically.
I'd love it if someone could find the stats on the Pats defense with respect to yardage given up by quarter. I've been saying for weeks now that they seem to give up a ton of yards early in games (usually in the 1st quarter), and then they start to make some adjustments and put the clamps on. I think they are coming out very, very vanilla on defense in games, figuring out what the opponent is doing, and then making in game adjustments. Its probably why they've also only given up 3 points in the fourth quarter over the past 5 weeks. Yesterday was more of the same. They gave up a couple long drives early, the sky was falling in the game thread, and then they turned Bryce Petty into Bryce Petty and it was over. Lather, rinse, repeat, week after week.
I agree with some of your points about weather, but a lot of the missed FG is luck. The game where the Bucs kicker missed 3 FG was in Tampa. Most of the misses were earlier in the year before the weather turned.
This sounds like nonsense, and moreover the Patriots have been good at yards prevented and at generating three-and-outs in previous seasons. In 2015 and 2016 they were top 10 in yards and 2014 was 13th. They were top-10 last year at forcing three-and-outs.
In fact no team has given up more fourth-quarter yards than the Patriots. They improve to 28th if you discount the first four weeks.
https://www.pro-football-reference.com/teams/nwe/2017_opp_splits.htm
Rushing:
1st: 4.5 ypc (101 attempts)
2nd: 4.4 ypc (108 attempts)
3rd: 4.8 ypc (102 attempts)
4th: 5.1 ypc (80 attempts--makes sense, teams trailing, Pats very willing to let them run)
1st half: 4.5 ypc
2nd half: 5.0 ypc
Passing:
1st: 86.4 rating
2nd: 95.3 rating
3rd: 94.2 rating
4th: 78.5 rating (time to clamp down?)
1st half: 91.4 rating
2nd half: 84.1 rating
YPA way down in second half. Also, Patriots only have 4 sacks in 1st quarter--they average about 13 in the other 3 quarters. If they had 9 more in first quarter, bringing them up to the average for the other quarters, they'd have 51 on year, good for 3rd in NFL.
Other nuggets:
--When teams are in their own territory, they average about 5.3 ypc. Once across midfield they average 3.8 ypc. Passer rating drops from about 89 to 80.
--Patriots defensive YPA, passer rating, and YPC all move in proportion to Game Win Probability. Pats give up some plays early (before adjustments?) and some plays late (game out of reach?)
--Opponents under center average 3.7 ypc against NE, and nearly 6 ypc when in shotgun.
The Bucs game was one of the windiest days in the month of October in Tampa. Just saying: https://www.timeanddate.com/weather/usa/tampa/historic?month=10&year=2017
I also don't really believe it's "luck" that a guy like Nick Folks misses kicks. He missed 2 the week before, and he was cut after the Pats game, and nobody has beaten down his door since. Was it "luck" that we held the Jets to 6 points yesterday, or is it more that Bryce Petty sucks?
Thanks for that Drew. I think the sacks goes to what I was saying about the 1st quarter. The Pats defense seems to be pretty vanilla early on in games. They take very few chances, seem to keep plays in front of them, and then ultimately, make some adjustments from that point forward. The 6ypc out of the shotgun is a pretty interesting stat. I'd love to know what that number is usually across the league. I think that's a function of garbage time, and allowing the opponent to run out the clock late in games if they want to do so.
If the Bucs have a crappy kicker, that's poor roster construction by them. But it's good luck for the Patriots - they didn't pick Tampa Bay's kicker, and nothing they did made him miss (unlike the Falcons kicker where we blocked one, for instance). With Petty at least you can argue how much is bad QBing and how much is bad D; if the kicker misses, that's good fortune.
This is silly on a million different levels. If you think the Pats over the course of the last 10 years are better off at any point in a game with Tom Brady and his awesome offense on the bench and the (generally speaking) much worse defense on the field, you are smoking crack.
I guarantee you BB would take a 3 and out every single time on defense to get Brady and the offense more drives, and is not the least bit worried about his offense being on the field too much or that creating more opportunities for offensive players getting hurt.
The Patriots are 29th in offensive drives, and first in points per drive. If they were top 10 in drives, they'd likely be in "historically great" territory with all the points they'd score. The thought that having your defense on the field for an extended amount of time is somehow beneficial is just another case of the belief that everything that happens on the field must be by Belichick's design, even the obvious flaws. He's not infallible.
The "bend but don't break" thing gets bandied around a lot, but I think it's pretty interesting how successful they've been with it relative to the league. If you looks at plays/drive, the pats are 2nd worst at 6.36, behind only Atlanta, and worst in yards allowed per drive (36.24). Most of the other teams near the top of that are... not very good defensively, which makes sense, since they're giving up a lot of long, scoring drives. But for the pats, you can see this manifest in terms of plays per point allowed, where they end up 4th, behind only Minn, Jac, LAC, who are the top 3 in ppg allowed. So somehow, while they're allowing a lot of long drives, they're not letting that translate into points.
As for why, I guess the hope is that it's the combination of limiting big plays, giving poor starting field position, and good red zone defense (as opposed to just being lucky). Thus, they're making teams run a lot of short plays (which they do successfully), but don't give up big chunks, meaning that teams need to string a lot of them together, and eventually they screw up or get stopped. If you make them get 7 first downs to score, your odds of stopping them once are better than if they only need 4 first downs to score.
The stat I can't seem to find, but that would be a prediction of this, is the variance of yards per play. The pats are 3rd worst in average yards per play (5.7), but if this is right, you'd expect they'd have a low variance, meaning they're forcing teams into long, plodding drives, and limiting the big play/quick score drives.
As for why, I guess the hope is that it's the combination of limiting big plays, giving poor starting field position, and good red zone defense (as opposed to just being lucky). Thus, they're making teams run a lot of short plays (which they do successfully), but don't give up big chunks, meaning that teams need to string a lot of them together, and eventually they screw up or get stopped. If you make them get 7 first downs to score, your odds of stopping them once are better than if they only need 4 first downs to score.
The stat I can't seem to find, but that would be a prediction of this, is the variance of yards per play. The pats are 3rd worst in average yards per play (5.7), but if this is right, you'd expect they'd have a low variance, meaning they're forcing teams into long, plodding drives, and limiting the big play/quick score drives.
Two quick points on the defense:
1) While they are 31st in DVOA, they are 22nd in weighted DVOA, which attempts to quantify how a team is playing recently. I think this is s significant difference.
2) They are 21st against the pass, but 30th against the run. It's really the run D we need to be worried about in the playoffs IMO.
1) While they are 31st in DVOA, they are 22nd in weighted DVOA, which attempts to quantify how a team is playing recently. I think this is s significant difference.
2) They are 21st against the pass, but 30th against the run. It's really the run D we need to be worried about in the playoffs IMO.
Is there any place to sort rushing stats by direction? It seems to me that they’ve been pretty stout up the middle but have been hosed around either end.
I was going to post this. Part of this is the eyeball test too. I think 22nd is about right and if KVN and Branch are healthy for the post season run they have the potential to be average to below average. With their offense and special teams we're in a better position than 2011. The defense is still going to give us all stokes/heart-attacks but they have a higher ceiling than some of their predecessors. YMMV.
To the credit of a few that’ve mentioned this, Charlie Weiss, yesterday, said that people are looking too much into Brady’s somewhat poor recent play and not giving enough thought to what the weather’s done to stifle offensive production. He seemed amused at the idea that people aren’t talking about the major impact the weather has on an offense.
Well, opposing teams will be exposed to the same elements. He wasn’t using it as an excuse and saying things will get better, he was just saying that the cold has a real effect on performance.
Yeah, I think Weis' point is that the proper comparison in bad weather is more "Brady vs Opponent QB" than "Brady vs Brady's better weather numbers" which is what most are doing when they cite last 4-5 Brady games vs first 11-12 Brady game stats.
I personally think that's less true about the interceptions, but still at least partially true. And ultimately, I'll take my chances with Brady rather than anyone else in the playoffs regardless of how much of last few weeks is slump, how much is injury, and how much is weather/weapons.
It does, though, also undercut some of the purported defensive improvement during those same weeks as well.
I personally think that's less true about the interceptions, but still at least partially true. And ultimately, I'll take my chances with Brady rather than anyone else in the playoffs regardless of how much of last few weeks is slump, how much is injury, and how much is weather/weapons.
It does, though, also undercut some of the purported defensive improvement during those same weeks as well.
I think this is right. A game like Sunday's played in those conditions isn't going to make me concerned about the offense, nor am I going to take it as a sign that defense has taken things up a notch. It's hard to play offense when it's that cold. They played it pretty conservative knowing the only way the Jets would make up that deficit is if the Pats screwed up. Obviously a team like Pittsburgh has better chances in cold weather than a punchless squad like that Jets, but you can't be disappointed about scoring in the 20s in New England in December / January.
On the flip side, it wasn't cold in Miami, and Ben had better overall numbers in Pittsburgh. There are reasons for both, of course. I think they need Hogan back if they are going to win it all.
Miami was a disaster, but that was really Brady's only bad game all year (he had a couple other not great games). I disagree Ben had better numbers. He had an extra drive to work with, got more out of his running game, and still threw for fewer passing yards and the Steelers scored fewer points. Brady was great against the Steelers, one bad pick excepted. A lot of the "bad" in Brady's current streak is the Patriots have been running it in near the goal line and it's depressed his touchdown pass numbers. They ran it in twice against Pittsburgh in drives where they predominantly threw the ball to move it down near the goal line. The offense is still putting up points.
Brady played fine against the Steelers, but his completion % was worse (62.9% to 73.3%), his YPA was worse (8.5 to 9.4), and they both threw a pretty bad pick. IIRC Brady started the game 10 for 12 and finished 12 for 23, something like 5 for his last 14 with a pick
The attempt is not the unit of opportunity in football; the drive is. The Patriots had about half as many rush yards and one fewer drive, but they scored more points and Brady had more passing yards.
As for the numbers you quote, they don't look pretty but they're a little misleading. Brady had a bomb to Cooks wiped out because Cooks stepped OOB, which didn't matter because the next play Gronk drew a DPI, and neither of those plays shows up in the stats. Bottom line, they scored on both fourth-quarter drives, including a TD in the two-minute drill for the game-winning score.
Yep. His 17 playoff catches for 332 yards and 2 td in 3 games were helpful last year as were Edelman’s 21 for 342. Cooks can help replace one, but not both, even with Mitchell back.
I'll wait to make plots until the post-week 17 numbers are up, but through week 16, in the points/drive vs yds/drive plane the 2017/8 Patriots had the most bend-but-don't-break-y D in the NFL over the span 2001-present, just ahead of the 2001 Bears. Accounting for average starting field position, they were 3rd, behind the '01 Bears and barely behind the '01 NEP. Factor in TOs, and they were back in 1st barely ahead of the '01 Bears.
I'd like to look at the distributions of starting defensive LOSs. I wonder if a team like the Patriots might have such different of a distribution that using *average* starting field position in the control sample to predict what the points allowed per drive given their yards allowed per drive and average starting field position isn't capturing the whole story.
I'd like to look at the distributions of starting defensive LOSs. I wonder if a team like the Patriots might have such different of a distribution that using *average* starting field position in the control sample to predict what the points allowed per drive given their yards allowed per drive and average starting field position isn't capturing the whole story.
Long-term (as in winning the SB), I think I’m more worried about the potential to be gashed by a Kamara/Gurley in the passing game than directly on the ground. McKinnon, McCaffrey and the ATL backs are all dangerous too.
Just thinking about Elandon Roberts in coverage.
Firmly in our tradition — the level of outrage occasioned by the misspelling of a binkie’s name is directly proportional to the degree of underperformance by said individual. Clay Buchholz, Charlie Weis. Though it must be acknowledged that Charlie was phenomenally successful separating colleges from their money.
True, although the emergence of Marquis Flowers has helped stabilize that area a bit
Hard to believe how much KVN actually matters when he’s in there. In addition to making plays, the defense looks markedly more coordinated when he’s holding things together.
I know we all talk about “next man up,” but being down three mics (DH —>McClellin—>KVN—>???) matters.
Personally feel he’s the guy we need back most for the playoffs.
I know we all talk about “next man up,” but being down three mics (DH —>McClellin—>KVN—>???) matters.
Personally feel he’s the guy we need back most for the playoffs.
I'm curious which way you think this breaks. It seems to me the Patriots do everything well that helps field position. They score on the highest percentage of their drives and have a great kickoff unit (and Gost has been amazing on KOs this year). They turn the ball over very infrequently (second-least), so they're not giving the D many short fields to have to defend. You can imagine some teams might do well scoring and kickoff well but turn the ball over too much and distort the starting field position average with a few short fields, but it seems to me the Pats do everything well in this area (which is why they are far-and-away the best).
I posted this last night on Twitter:
Here's an insane statistic and I don't know what it means. The @Patriots defense has faced 92 pass plays in the red zone, most in the league. They've faced 36 run plays in the RZ, tied for least. And they're actually better against the pass in the RZ, statistically
Most of that appears to game-script-related; it's 70 passes, 16 runs when they lead and 22 passes, 20 runs when they don't. This is despite the run game being more efficient than passing in the red zone in general and against the Pats in particular (they average 3.25 yards per carry in the RZ and 2.55 yards per pass attempt). So I think even the red zone success is due in significant part to the offense.
Blue are the control sample of non-NEP team-seasons 2001-2017 which are used for the regressions. Red are NEP team-seasons 2001-2016, green is the NEP 2017. The dotted lines are the fits, the dashed lines show the slope scaled to the NEP 2017 point.
Right, I'm starting to wonder if it would be better to look at the effect of a team's offensive turnovers on the defensive pts-v-yds allowed.
That's what I always assumed, but then it occurred to me that offensive turnovers could in principle affect scoring more strongly than DLOS. However, I'm not seeing anything in the data so perhaps not.
The scatter in the control-sample residuals to the pts-v-yds fit is reduced by:
21% when including only DLOS
7% when including only defensive TOs
28% when including both dTOs and DLOS
13% when including only offensive TOs
23% when including both oTOs and DLOS
indicating that, as you say, oTOs and DLOS are less orthogonal than dTOs and DLOS.
I came to this thread because I was just having this exact conversation with someone on Twitter. By DVOA, the Pats have a bad defense, yet they have a top ten defense in points per drive. You would think that this would be unsustainable, however the Pats have had a top 10 scoring defense the last 6 years while having varying levels of success in yardage and DVOA.
I wasn’t able to put my finger on how this could be apart from them just playing better in the red zone, but a lot of the things mentioned previously seem to make sense, with the cumulative effect being fewer points allowed than you would expect.
I wasn’t able to put my finger on how this could be apart from them just playing better in the red zone, but a lot of the things mentioned previously seem to make sense, with the cumulative effect being fewer points allowed than you would expect.