Photo: USA Today Sports Images

by Vincent Verhei

In last week's edition of Quick Reads, we explored a few different methods of calculating quarterback streakiness, then opened the floor for comments and further discussion. As it turns out, the most illuminating comment had nothing to do with calculation or methodology, but simply with the visual display of data we used in our charts:

Dr. Mooch: Streakiness: first a comment on the graphs. It's easy to see differential performance on the graphs of DYAR, but we should call attention to the fact that the "right now" performance is actually the slope of the graph. A more precise representation of differential performance would be a bar graph of the DYAR values for each individual play.

Which provoked similar sentiments from another reader:

nat: Good point about the graphs. Using a cumulative DYAR graph obscures the streaks rather than highlighting them.

These observations are both spot-on. Looking at each play one at a time, rather than their running DYAR total, makes it easier to spot the streaks in each quarterback's play. For example, here's the cumulative DYAR chart we ran last week for Matt McGloin, who was found to be one of the streaker quarterbacks of Week 4:

And here's the same data, presented with the DYAR for each individual play, one at a time, on a column graph:

This makes McGloin's streaks of good passes (those with positive DYAR) and bad passes (negative DYAR) really stand out. He had nine streaks in all:

• One bad pass, which was followed by
• One good one, then
• Three good
• Four good
• One very slightly good (a 4-yard gain on first-and-10 that was worth 0.4 DYAR)
• And finally three bad ones.

And since we know McGloin had 20 dropbacks, we can say that his average streak lasted for (20 divided by 9 equals) 2.22 passes. Boom. Simple definition, logical calculation, can quickly be figured for any quarterback over any number of passes. We may want to do something where those close-to-zero plays don't end a good or bad streak, but this will certainly do for now.

Using this methodology, the streakiest starting passer of Week 5 was Green Bay's Aaron Rodgers, who had six streaks in 20 dropbacks, for an average streak of 3.33 plays:

Other streaky starters included Andrew Luck (average streak: 2.78 plays):

and Brian Hoyer (average streak: 2.71 plays):

If we're including reserves, we should point out that Logan Thomas had three streaks in ten plays, so he was just as streaky as Rodgers. However, his graph is far more hilarious, and too funny not to share:

On the other side of the spectrum we have Tom Brady, whose average streak against Cincinnati lasted just 1.54 plays:

Is it good to have a streaky quarterback? The five streakiest starters of the week (Rodgers, Luck, Hoyer, Austin Davis, and Jay Cutler) went a combined 3-2 with 195 total DYAR. The five least streaky passers of the week (Brady, Joe Flacco, Drew Brees, Philip Rivers, and Ryan Fitzpatrick) also went 3-2, with 109 DYAR between them. No strong correlation there, though it’s worth noting that there are some awfully good quarterbacks at either end of the spectrum.

We can also use this to measure the streakiest quarterbacks of the year. Among starters, that has been Hoyer; his average streak this season has lasted 2.47 plays. This makes sense when you consider the Browns’ penchant for late-game comebacks. Hoyer has -6 DYAR in the first half of games this year, but 254 DYAR after halftime, second only to Luck (352 DYAR). Here's what Hoyer has done each week so far this season:

Brian Hoyer's Streakiness By Week
Week Opponent Plays Streaks Avg. Streak
1
PIT
34
10
3.40
2
NO
42
23
1.83
3
BAL
27
10
2.70
5
TEN
38
14
2.71
TOTAL

141
57
2.47

This is a bit of a data dump, so we'll save the deeper analysis for a later date, but I think this methodology gives us something concrete to measure down the road.

Here are the leading running backs for Week 5, as measured by rushing DYAR only:

• 1. Eddie Lacy, GB: 13 carries, 105 yards, two touchdowns (48 DYAR)
• 2. Khiry Robinson, NO: 21-89-1 (41 DYAR)
• 3. Arian Foster, HOU: 23-157-2 (41 DYAR)
• 4. Ben Tate, CLE: 22-123-0 (40 DYAR)
• 5. Benny Cunningham, STL: 7-47-1 (32 DYAR) (All of his carries gained positive yardage, three of them gained at least 11 yards)

The least valuable rusher of the week was Chicago’s Matt Forte (17-61-0, -26 DYAR). Five of his carries were stuffed for no gain or a loss, and five more gained exactly 1 yard. A median gain of 1 yard is bad, very bad, and one of his “better” runs resulted in a lost fumble.

 Five most valuable running backs Rk Player Team RushYds RushTD RecYds RecTD TotalDYAR RushDYAR RecDYAR 1. Branden Oliver SD 114 1 68 1 68 30 38 Oliver had 19 carries against New York, and every single one of them gained positive yardage. That's exceptional. Only two of his runs gained more than 6 yards, but one of those was a 15-yard touchdown on third-and-2, the other was a 52-yarder. And three of his shorter runs also gained first downs. Oh, and he caught each of the four passes thrown his way, including a 50-yard gain on third-and-17 and a 9-yard touchdown. 2. Pierre Thomas NO 35 1 77 1 58 22 37 Thomas had eight catches in ten targets for a team-high 77 yards, including a 15-yard touchdown on third-and-12 and three other first downs. He only ran the ball four times, but that included a 27-yard touchdown and a conversion on second-and-1. 3. Eddie Lacy GB 105 2 27 0 56 48 8 Twelve of Lacy's 13 carries gained at least 1 yard, and five of them gained 10 yards or more, including two touchdowns and a 29-yard gain. He caught each of the three passes thrown his way, including a 21-yard gain on third-and-10. 4. Shane Vereen NE 90 0 18 0 45 31 14 Eight of Vereen's nine carries gained at least 5 yards, five gained 10 yards or more. He had five first downs, with a median gain of 10 yards. He caught each of the three passes thrown his way for two more first downs, including a 7-yard gain on third-and-5. 5. Khiry Robinson NO 89 1 8 0 42 41 1 Robinson was stuffed for no gain or a loss six times in 21 carries, but he had eight first downs, including a game-winning 18-yard touchdown in overtime.
 Least valuable running back Rk Player Team RushYds RushTD RecYds RecTD TotalDYAR RushDYAR RecDYAR 1. LeSean McCoy PHI 81 0 6 0 -31 -16 -15 McCoy was stuffed for no gain or a loss seven times in 24 carries, with only four first downs. He caught each of the four passes thrown his way, including a pair of first downs, but collectively they gained just 6 yards. One resulted in a 9-yard loss and a lost fumble.
 Five most valuable wide receivers and tight ends Rk Player Team Rec Att Yds Avg TD TotalDYAR 1. Demaryius Thomas DEN 8 16 226 28.2 2 59 Seven of Thomas' receptions gained at least 13 yards and a first down, including touchdowns of 31 and 86 yards. 2. Alshon Jeffery CHI 6 7 97 16.2 1 51 Five of Jeffery's receptions produced first downs, with three gains of 20 yards or more, including a 25-yard touchdown. 3. Brian Quick STL 5 9 87 17.4 2 50 Quick had touchdowns of 8 and 5 yards, plus gains of 25 and 43 yards. 4. Golden Tate DET 7 9 134 19.1 1 50 Tate had four gains of 18 yards or more, including a 55-yarder. 5. Emmanuel Sanders DEN 7 9 101 14.4 0 41 Five of Sanders' receptions gained at least 11 yards, capped off by a 30-yarder.
 Least valuable wide receiver or tight end Rk Player Team Rec Att Yds Avg TD TotalDYAR 1. Kelvin Benjamin CAR 3 11 40 13.3 0 -61 We all know that Benjamin is Carolina's best receiver, but that doesn't mean that forcing the ball to him on every play is a good idea. His longest reception was a 20-yarder, but it ended in a lost fumble. Only one of his catches resulted in a first down. He also drew a 6-yard DPI.

102 comments, Last at 08 Oct 2014, 4:22pm

### 1Re: Week 5 Quick Reads

I have actually defended flacco in the past and maybe it is just a function of not having Pita, but I thought flacco was really awful this week. Given the colts turnovers and the ravens rush success, anything above sub replaceable play and they win.

Since this was my first game watching the ravens - I'm going to assume it was just an aberration, but this really was about as a bad as I've seen Flacco since his days as a rookie.

### 39Re: Week 5 Quick Reads

I thought he did adequately, all things considered. A lot of bad things happened outside his control such as:

- Jacoby Jone's turnovers and poor returning decisions leaving terrible field position
- Steve Smith's first play fumble
- An UDFA rookie LT getting whipped and by all accounts the game plan was never updated to fix this (no extra blocking help brought in and no apparently hot reads or screens)
- Several drops by Torrey Smith, who's been pretty disappointing this year.

So far Flacco has had three good games and two bad ones, and both bad games had some huge issues with his supporting cast.

Now I'm coming at this from the perspective that Flacco is generally a pretty average QB, not the next Peyton Manning, like some Ravens fans will have you believe.

### 41Re: Week 5 Quick Reads

"Flacco is generally a pretty average QB, not the next Peyton Manning like some Ravens fans will have you believe."

Yeah, um, Ravens fans don't believe this. But we are accustomed to FLacco getting way more heat than he deserves and having his accomplishments routinely ignored. So perhaps we're a bit more outspoken about him because he's never credited for his success. He has the best playoff stretch in decades against some of the best teams in the NFL, and all people can offer up is "fluke". Yeah, I think you can understand why we tire of seeing him totally discredited.

Ravens O has been pretty good this year (6th in DVOA before Colts game) even though once upon a time, all the credit would have gone to Ray Rice.

### 40Re: Week 5 Quick Reads

I don't think you were watching this game closely enough then.

Flacco was under siege all day with relentless pressure in his face in less than 2 seconds with regularity. James Hurst got dominated and the Colts brought alot of blitzes that didn't get well picked up.

Not saying he played great, but I find it more a function of being under assault. I didn't see Flacco himself making the key mistakes. He didn't drop that great 4th and 3 pass to end the game.

### 57Re: Week 5 Quick Reads

I didn't think it was all his fault. I did agree the line failed on those blitzes in the A gap and there were plenty of drops. I just thought he was also inaccurate overall.

### 94Re: Week 5 Quick Reads

They don't have Pita Pit in Baltimore? Or is that too bourgeois for millionaires such as Flacco?

### 97Re: Week 5 Quick Reads

I agree with others here, I thought Flacco wasn't the main problem. His O-line evaporated under the rush, and his receivers didn't catch the ball. In general it looked like the Indy defensive game plan owned whatever the Ravens were trying to do. Aaron Schatz' comment about hot reads in Audibles seems spot on.

I guess if there's a bright side for Ravens fans, it's that after stumbling around for 58 mins the team still had a chance to tie at the end. That's something.

### 2Re: Week 5 Quick Reads

So Wilson was the most valuable runner this week too. Heh.

### 9Re: Week 5 Quick Reads

I am trying to figure out how 18/24 for 201 yards and 2 TDs is only worth 4 passing DYAR. Er?

I agree that Wilson had a great game on the ground, but I thought he did a fine job of throwing the ball, too.

### 14Re: Week 5 Quick Reads

He was sacked 3 times and I think he had a fumble as well.

### 16Re: Week 5 Quick Reads

Guessing:
1. Wilson fumbled twice at the snap. Were those charged against his run or pass totals?
2. Wilson was sacked 3 times, each time on 3rd down.
3. Only 8 of Wilson's completions went for 1st downs.
4. The Redskins have a below-average pass defense.
5. The Redskins were never flagged for pass interference.

### 24Re: Week 5 Quick Reads

Agreed, I would have expected him to at least be better than Locker.

### 32Re: Week 5 Quick Reads

Maybe they changed the rankings, but he is ranked higher thank Locker.

Locker, however, is a small sample size. Played out to a full game his numbers equated to 20/28 198 yds 2 TDs 0 INTs in the air and 10 att 85 yds 2 TDs on the ground.

### 37Re: Week 5 Quick Reads

Is there any way someone could recalculate Wilson's passing DYAR if both of the 26 & 41yd TDs to Harvin had not been negated? You know, just for fun.

### 42Re: Week 5 Quick Reads

I am not sure but based on the graph above where the 81 yd TD pass was worth about 45 DYAR I would make a WAG that those two passes would be worth 40 DYAR. This ignores any DYAR that should be removed, i.e. passes that would not have been thrown if the TD's stood.

### 46Re: Week 5 Quick Reads

Of course, every team suffers from drops, but it didn't help that Baldwin and Harvin both dropped two first-down converting passes.

### 3Re: Week 5 Quick Reads

I don't know if Austin Davis is any good or not, but it is kind of funny seeing him highlight just how much time and resources the Rams have wasted with Sam Bradford.

### 12Re: Week 5 Quick Reads

So far Davis looks like he just might be pretty decent, but you're right to be skeptical. I'm a Browns fan and I remember being excited by Derek Anderson, who was on fire for part of the 2007 season. I thought maybe we'd finally solved our quarterback problem. Six or seven quarterbacks later, we're hoping that Brian Hoyer has solved our quarterback problem.

### 4Re: Week 5 Quick Reads

What a difference a week makes.

Has anyone else besides Brady gone worst to first in QB DYAR from one week to the next?

Is this the largest week to week difference in DYAR?

### 5Re: Week 5 Quick Reads

I'm surprised there is anyone lower than the Jets QBs, let alone guys like Luck, Flacco, Brees and Stafford. They just looked so abysmal. In that game, they didn't even convert a single third down until the 4th quarter and it isn't like they were converting them on first and second down either.

### 8Re: Week 5 Quick Reads

That's why DYAR by itself doesn't tell the whole story - although in this case it's telling us that the Battle of Baltimore should have been one of those 35-31 games if either QB plays well. That was the shakiest 300yd game I've seen yet from Luck, and they still won.

### 20Re: Week 5 Quick Reads

Agreed. And it occurred to me after I posted that question that DYAR is a cumulative stat and neither QB had enough snaps to really let it go negative (also neither had any fumbles and only a few sacks despite constant pressure). If you combine their DYARs then Jet QBs are 4th from the bottom which is closer to where I would think they would be. Still seems a little high, but I suppose they only had 1 turnover between them.

### 11Re: Week 5 Quick Reads

It seems likely that Rex Ryan will lose his job by the end of the season, and if he does, he can trace his downfall to his quarterback situation. He wasted years propping up Mark Sanchez, then dumped him in favor of Geno Smith, who appears to be even worse. I understand that it can be extremely difficult to find a good quarterback, but still, the NFL is a pass/fail league, and Ryan has failed in that regard. He could have at least picked up a veteran journeyman who would have been halfway competent instead of these young gunslingers who clearly don't know what they're doing. But he chose to go with the young gunslingers, and he will probably pay for it with his job.

### 15Re: Week 5 Quick Reads

Good point. Teams that are heavily oriented toward defense rarely gamble their window on trying to develop a franchise QB. And rather stubbornly at that. They get some old veteran, ask him to take care of the football, and hope for a good run in the playoffs.

Here's the thing. If Sanchez had worked the Jets could have won the Super Bowl. But hitting on a 1st round QB is a risky proposition, the equivalent of going for the two-point conversion in the final play of the game. Defensive-minded teams don't take those risks on offense. All they want from their offense is not to lose the game.

The Jets don't know who they are and they haven't known in all of Rex's tenure.

------
Who, me?

### 49Re: Week 5 Quick Reads

Since the 2001 season, 7 Super Bowls have been won by quarterbacks drafted in the first round by that team. 4 Super Bowls have been won by quarterbacks drafted in a later round: Brady with 3, Wilson with 1. 2 Super Bowls have been won by teams with quarterback acquired through trades or free agency: Tampa Bay, and New Orleans. Only the 2000 Ravens and the 2002 Bucs fit the profile of a defensive team with an older veteran quarterback. Baltimore had more success once they drafted Flacco. The problem is not that the Jets tried to develop a franchise quarterback, it's that they thought Sanchez was one.

### 74Re: Week 5 Quick Reads

And if you take the fluke 6th round pick who won 3 superbowls out of your data, it becomes pretty clear that the 1st round is the only reasonable place to get a QB.

Yes, it is unlikely you're going to win a Superbowl by playing a first round pick at QB. It's much more unlikely that you're going to win a superbowl if you don't.

### 75Re: Week 5 Quick Reads

Not that I disagree overall, but there is also a selection bias at work. First rounders are pegged more often as starters and are given far longer periods of latitude. Mike Glennon had a very expected rookie season as a starter. He was summarily written off instead for an aged small sample wonder. It almost feels like you need to have Brady/Wilson success as a rookie late rounder for the team to give you a chance.

### 80Re: Week 5 Quick Reads

Even Brady needed some "luck" (the injury (iirc) to Bledsoe) before he even got his chance....

Duplicate post

### 82Re: Week 5 Quick Reads

Yes, the league has moved towards offense and maybe only two teams have won that way since the 2001 season, but if you're not aiming for that model, what do you want Rex Ryan as your head coach for? It's like a complete misunderstanding of your teams strengths and likeliest potential winning formula.

------
Who, me?

### 21Re: Week 5 Quick Reads

Not defend Geno, but who does he have to throw to? I think I rather have Denver's backup receivers than the Jets cast without Decker. Give Peyton this group and he wouldn't do a hell of a lot better.

### 28Re: Week 5 Quick Reads

It's true that Geno's supporting cast is awful but his inconsistency is what makes him the problem and not part of the solution.

### 44Re: Week 5 Quick Reads

The problem is that they shoved him into the starting line-up his rookie year. They didn't have a choice once Sanchez got hurt, but still, he needed a year holding a clipboard. He might be better off being benched now and coming back in when Vick gets hurt.

I might add that the entire team took the day off, including the run defense. How Harrison and Wilkerson graded out well on PFF, I'll never know. Before this game, you pretty much knew what Geno would do: If Decker is playing, he completes 60 percent, makes some good throws, and has two to three bad plays that killed you. Against San Diego, he completed 25 percent of his passes, but made no plays that killed them. His interception was basically a punt with no time left in the half, he didn't fumble (unlike Chris Johnson). The offense couldn't stay on the field against a team not known for its defense, but the entire offense played horribly, which is why the fanbase is flipping out right now.

### 50Re: Week 5 Quick Reads

Peyton Manning's 2010 season begs to differ.

### 54Re: Week 5 Quick Reads

You think Reggie Wayne and Pierre Garcon are comparable to Jeremy Kerley and Eric Decker?

### 67Re: Week 5 Quick Reads

While I am (and as far as I can remember always have been, even when he was playing for a competing team) a Manning fan, I think some of the discrepancy between Brady and Manning is that Manning has always had high expectations in part because he has always taken a salary high enough that people thought that some of the other positions (especially, the defense) were neglected because the Colts simply didn't have the cap room. Brady plays in that "we are all just workmen here" environment that Belichick has installed at the Patriots. Thus, his (Brady's) supporting cast is expected to carry more of the load and thus gets more blame when things go south. (I hope there isn't a FOMBC for mentioning the irrational Manning/Brady debate.)

I don't know what Manning's current salary is, but the Broncos seem to have as much auxiliary talent around him as they put around Elway. The "failures" (as if anything short of a SB win is a failure) of the last couple of years have not been only Manning's fault.

None of this is relevant to the Jets and whether they might not even be better off with Kyle Orton at their helm. At least their wagon isn't hitched to Tebow as QB. I'll let Will Allen opine on whether lazy Jay Cutler would be an upgrade for the Jets.... ;-)

### 76Re: Week 5 Quick Reads

The "l" should be uppercase.

I fear that Geno might share some of Jay's approach to his job, while not having as much throwing talent.

### 58Re: Week 5 Quick Reads

Not that I think Jets receivers are any good, but who in their right mind thought the bronco receivers were that great heading into the 2012 season? Mind you, this was before Julius Thomas and Welker - it was Dthomas and decker - both of whom finished near the bottom in dvoa. Now we all know tebow played a huge part in that, but whos to say that the same thing isn't happening to some extent right now? Do we really believe if Manning or Rodgers or Rivers was inserted into this lineup, we'd get the same results?

### 63Re: Week 5 Quick Reads

IIRC, a lot of people thought Thomas should become a star and that Decker was a serviceable #2 WR. They had low DVOAs because Tebow was their QB. The sentiment, when Manning was signed, was that they should shine with him as QB, and that they were certainly good enough to run a high-powered offense.

### 69Re: Week 5 Quick Reads

Yeah, I remember DThomas making some great plays and thinking this guy needs a real QB.

### 73Re: Week 5 Quick Reads

That might be fair - except I remember in FO almanac that even under Orton - the receivers weren't posting great dvoas either. I'm not saying they were bad and its far more likely DT improved a great deal in his third year as most top receivers would. And honestly, I think calling Decker a serviceable number 2 at that point in his career would be charitable.

Again, I'm not saying Manning created these players or that the jets receivers don't stink. I was responding to the above that Pm would look the same as Geno.

### 95Re: Week 5 Quick Reads

"Give Peyton this [Jets WR] group and he wouldn't do a hell of a lot better."

So you think Gino Smith would throw for 40+ tds if he were in Denver? I'm sorry but that's probably the stupidest opinion I've ever heard anywhere besides Fox News. Are you the same commenter who earlier described Vick as a "young gunslinger?"

.

### 22Re: Week 5 Quick Reads

...pass/fail...that's a great double entendre.

### 86Re: Week 5 Quick Reads

There seems to be a question of just how much a coach is involved in the development of a qb and if you take one side, you have to remain consistent. If you think rex "failed" Sanchez and Gino, then you have to believe that Mora, Belichick, and others aided the qbs who became successful. I don't really buy this argument.

Rex got unlucky in a lot of ways. Tannenbaum is the real culprit for the Jets disaster. Rex has just been given a suicide mission. Honestly - outside of two players on their d line and o line, who on the jets is any good?

### 98Re: Week 5 Quick Reads

Interesting question. I do think player development is a thing: putting the player in the right system or adjusting it to better suit his abilities, coaching (the right) technique (Waldman says surprisingly little of this takes place in the pros, and though it's a position coach's job, the head coach chooses his staff), providing opportunity (the biggest thing Belichik did for Brady, and it would be a mistake to assume every other coach would have given his 6th round QB the ball).

On the other side of the spectrum, a head coach can hang on to a failed prospect too long (even if the GM and the head coach don't get along, the HC can put up a fight, like when Philbin benched Vontae Davis to force Ireland to trade him), he can rush a prospect onto the field, and he can be too passive regarding the talent acquisition strategy. A HC needs to be firm in his demands, but first he must know what he wants. A HC was resources, he can stand for his convictions, go to the owner, make noise with the press, etc.

No one knows for sure, but Rex certainly gives the impression of failing in many of these points. For example, he always gave the impression Sanchez was his guy, never pushed for a veteran and always stood by him even when it was clear he wasn't the answer. Didn't he even get a tattoo of him or something? If you get that close to your players, you won't have the objectivity to do what needs to be done.

------
Who, me?

### 6Re: Week 5 Quick Reads

Funny that Dalton was 7th this week. I thought he was terrible. Not that I watched the whole thing through to the end.

------
Who, me?

### 33Re: Week 5 Quick Reads

Yeah, I'm surprised. However, I guess he did do some decent stuff in the 2nd half, and his terrible play (and WRs didn't help the times he managed to actually throw a good ball) in the first half (and sometimes in the 2nd half) were all incompletes rather than INTs, so he never managed to wrack up enough bad DYAR to offset his occasional good plays.

But from memory/feeling, he was responsible (or he and his WRs were very off on expected routes) for several drives dying, and that managed to reduce his number of plays to wrack up negative DYAR.

### 7Re: Week 5 Quick Reads

I think it's illuminating that Forte was the least valuable rusher. With all of the criticism that Cutler takes, the fact that the Bears have not established an effective running game is one of the (many, many) problems with their supposedly elite offense. It seems like every other running play I see with Forte ends at the line of scrimmage or behind; I don't know how much of that is him and how much is bad playcalling but I suspect the latter plays a big role.

That's not to dump on Forte, because outside of the anomalous but crucial fumble in the 4th quarter Sunday, he hasn't really been terrible - but I think a lot of Bears fans have gotten caught up in his yardage stats because of how the Bears have relied on him in the passing game.

### 26Re: Week 5 Quick Reads

I think injuries to the offensive line are a bigger problem than both. Bears were missing starting (pro-bowl) left tackle and starting center this game, and were missing starting left guard and center for a while now.

Of course, Forte has always been a little boom and bust.

### 10Re: Week 5 Quick Reads

7. Andy Dalton CIN 15/24 204 2 0 65DYAR
8. Russell Wilson SEA 18/24 201 2 0 4 DYAR

I haven't seen any of these games. But how come Dalton scored 60+ more DYAR than Wilson with equal number of yards and 3 less completions on same number of attempts? Did Wilson have a few long passes and loads of unsuccessful short ones?

### 17Re: Week 5 Quick Reads

Some of that may be opponent adjustments. Before this week, NE was ranked 2nd in DVOA against the pass, while Washington was 23rd. Dalton was also only sacked once for -8 yards, while Wilson was sacked three times for -23 yards. And Wilson had two fumbles, while Dalton had none.

### 18Re: Week 5 Quick Reads

EDIT: double post.

### 13Re: Week 5 Quick Reads

How much of this streakiness is actually the play of chance? I have some good plots of coin tosses that have some runs of 3, 4, 5+ heads (tails) in a row that I imagine you will be equally as interested in. And one outlier in the middle of 9 negative plays is not something I would call 3 streaks.

### 23Re: Week 5 Quick Reads

I think there is a fair point in this comment. I don't think a "streak" of 1 (or even 2) plays counts. Often times in math, one needs to see the pattern repeated 3 times before one takes it as a pattern. I think the same thing applies to streaks. If there aren't 3 plays in a row either good or bad, it isn't a streak.

Thus, the 1 outlier in Logan Thomas's play is not a short streak, but merely an outlier--an important one as it almost saved the Cardinals, but an outlier just the same. In particular, it didn't cause the Broncos defense to suddenly respect him and quit playing aggressively.

Similarly, I wouldn't call Brady's play "streaky" at all, simply "inconsistent". Only twice did he have 3 plays all in the same DYAR direction at a time (a set of 4 bad plays (11-15) and later 3 good plays (28-30)). There is also a stretch where he alternated good/bad DYAR plays (16-24) and that was as long as his two streaks combined. Interestingly, Brady's "inconsistency" didn't hurt him, because on average the good plays were better than the bad plays, which is why I put inconsistent in quotes, as I'm not sure it was real inconsistency as much as varying around a mean that had points both above and below the 0 DYAR line.

### 27Re: Week 5 Quick Reads

(Note this was originally part of my previous post, but I realized it was a slightly different topic, so I split it.)

BTW, I like the new graphic you evolved. I think it shows "streakiness" quite well, even if we haven't yet figured out exactly what streakiness is yet.

I would add another line (or two) to the graph though. You divide the graph at 0 DYAR which is reasonable.

However, I think there needs to be a second line where the play actually loses yards (and isn't simply worse that replacement level). Thus, there are some passes which are unsuccessful, but still positive (gaining yards, even if not enough to be effective). Those would be between the second line and your 0 line.

A third line, showing positive DYAR plays that don't meet the criteria for successful DVOA plays would also be useful. Those would be between the 0 line and the third line.

Passes in between those new lines are neither completely positive nor completely negative and can be discounted for breaking streaks. Well, actually we've seen some QBs like to throw all their passes in that range, gaining a yard or two, but not enough to matter.

I'd liken it to the stomps and guts. Plays above the 3rd line are like stomps. The QB has done more than enough to keep the chains moving. The QB is actually having "big" (or at least successful) plays where the passing game is likely to be influencing the scoring and/or opposing defenses style.

Plays below the 2nd line are stomps for the defense. These are the sacks etc. that stop drives.

Plays in between the 2nd and 3rd lines are guts or skates. The QB is making yards, but the drives are still stalling (well unless the running game is saving them).

Putting this together with my previous reply:

A positive streak would be at least 3 plays in a row above the 3rd line.

A negative streak would be at least 3 plays in a row below the 2nd line. (Hmmm, need names for those lines.)

A mediocre streak would be at least 3 plays in a row between the 2nd and 3rd line.

Once a streak was started it would continue until at least 2 plays in a row were outside the region (i.e. below the 3rd line for positive streaks, above the 2nd line for negative streaks).

Note, if you follow this model, there will be plays that aren't in any streak, at least not as I have defined them. That's ok by me, because I think those are times where the QB is "inconsistent" not being good enough to consistently move the chains, but not necessarily bad enough to stall drives.

### 51Re: Week 5 Quick Reads

This and similar websites have done a good job over the last couple years educating us on why "momentum," "clutch" and "hot hand" are overused concepts in sports. I think "streaky" needs to be subjected to the same scrutiny. My guess is that what looks like a streak or a trend in a players performance during one game probably is the result of natural randomness or other factors, such as defensive changes caused by the game situation.

### 66Re: Week 5 Quick Reads

I tend to think of "streak" as a descriptor of what happened, rather than a predictor of what will happen. If a coin comes up heads five times in a row, it's a streak. It doesn't matter that the streak was random.

### 77Re: Week 5 Quick Reads

Without trying to argue that momentum, clutch, streaky, hot hand, etc. are not overused concepts, I do think descriptive terms have their place and can add insight. Yes, football plays are mostly independent events and what happens on one play is not a strong indicator as to what happens on the next play (beyond what down and distance gives you). That said players (esp. QBs) do get hot and cold from time to time, either within a game or over the course of a season. It is something apparent to even casual viewers. Denying that because it isn't something one can statistically reason from, is to deny what we actually perceive. Yes, we shouldn't take the "man-in-the-moon" as an actual face, but our perception system still sees it, and it does tell us that certain patches on the moon are darker than others and are arranged like facial features.

Similarly, if a coin flips heads 5 times in a row, one cannot help but notice a streak, if it flips heads 10 times in a row, one had better question whether it is a fair coin, or whether there is some other factor influencing it. We should apply the same thing to football events. Fumble luck is non-predictive and teams regress to he mean on a regular basis. However, coaching ones players to create fumbles (or to avoid having them) is not a superstitious act. Therein lies the difference. Noting a pattern in random data and basing ones actions upon it without analysis that suggests their might be a causative reason is superstition. Noting a pattern and investigating whether there might be a causative reason is science. (Causing fumbles is something that increases one chances of recovering them, despite recovering thme being a random event. You don't recover 100% of the fumbles that don't happen.)

I have no data suggesting that QBs streakiness has any underlying cause other than random fluctuation. However, to determine if there is an underlying cause, one must define streakiness, formulate a hypothesis as to an underlying cause, collect data, and see if the data supports the hypothesis. We are still trying to define streakiness, one of the first steps.

Until we have a consensus definition of what we mean by streakiness, at best we will be talking past each other, with one person meaning one thing and another meaning something else.

With a definition of streakiness, we can rationally discuss whether some quarterbacks are more streaky than others and speculate as to why. Just like a coach might bench a running back who has caught "fumbilitis", even if it isn't a "real disease".

### 29Re: Week 5 Quick Reads

One could easily come up with an expected amount of streakiness. It would relate to the geometric distribution. If you have a p probability of completion, then after a given completion, the probability of seeing a streak of length k would be p^(k-1)*(1-p).
So the expected length of such a streak would be 1/(1-p). For p=.6, this works out to be 1/.4 = 2.5. And for the reverse streak, it's 1/.6 = 5/3 = 1.666.

Long streaks should definitely be the exception, if one assumes pass completion is i.i.d.

### 90Re: Week 5 Quick Reads

Brian Burke used something called the "runs test" when he was looking at momentum. The methodology is here:

### 101Re: Week 5 Quick Reads

Thanks, that was a good article and a nice pointer to a site I need to read more from. Moreover, the runs test at least partially answers the point made by yayFootball (#84 below).

In my opinion the test needs a couple of tweaks for reasons I'm going to outline here. They basically have to do with setting the zero point and better allowing for errors in the measurement.

Ok, point one, a team that does better than average (or worse than average) may appear streaky because their number of consecutive positive (negative) plays is larger than would be expected. That's because the zero point is arbitrarily set and does not actually represent that teams mean performance. With a coin, that works fine because we expect the coin to be "fair" and the odds of heads and tails to be equal. However, most football teams are not fair coins. Instead, they are better or worse than average, and may be a mixture of the above. It doesn't matter whether the zero is set from "expected success rate" or "replacement level play" or any other arbitrary metric--different teams will have different performance against that metric. Some teams will be better than average and some worse and thus will have an excess of runs on one side or the other.

Thus, to map the team to a fair coin, one has to adjust (and that is non-trivial) the play to compare to the teams average (and possibly accounting for the opposing teams average also, i.e. "opponent adjusted"). Thus, a successful play might better be defined by being above the teams average success rate for that play (in that situation). I don't know if there is enough data to compute that, there certainly isn't early in the season.

Perhaps, one could approximate that by computing a mean value of the zero point for the game. Lets assume we are using a stat like DVOA which has a built in 0 value which we can calculate a percentage difference from, e.g the Seahawks might have had a DVOA for a game of 35%. We simply adjust the meaning of success so that 35% above normal is the mean for that game. Thus, for a play to be successful it has to be 35% more successful than average. Note, you can even do separate DVOAs, so that you can separate good performance on the offense, from good defensive or special teams performance. You can also use other metrics like DYAR and tease out individual player performance, i.e. figuring out streaky QBs. The key thing is you need to compare to the individualized (i.e. just that team, just that unit, or just that player) average performance for the measurement to be "fair" and have equal numbers of successes and failures.

Now, point two, there is noise in the data. Players, even when on a streak, are not consistent (and other factors influence their results also). Tom Brady's numbers above suggest that well. There was a series of plays where he appeared to almost precisely alternate between good and bad plays. I don't think those were one play long streaks alternating, just simply the noise in the measurement. His average over those plays was slightly above zero, but his variance allowed individual plays to appear both above and below that average (and by enough that they appeared both above and below the zero line).

Perhaps, there is a version of the runs test that allows one to adjust for standard deviations. (Something like a 3-point running average, that smooths the curve a little. Of course, smoothing the curve is going to make it look like the streaks are longer because they will tend to erase jittery aberrations.) I'm not a statistician, so I'm not aware what tools are available in that box.

### 84Re: Week 5 Quick Reads

I agree that the definition of streakiness needs work, and it goes beyond the question of whether a single play should count as a streak. Consider the case where every single play a QB makes has positive DYAR. The current definition would label this performance as incredibly streaky, but I don't think any person would choose that adjective. When we apply this descriptor to a player, we are saying that this player prone to prolonged streaks of BOTH good and bad play.

I therefore propose a simple modification. Instead of looking at overall average streak length, we should track separately the average streak length for positive and negative plays. I would then only call a player streaky if both numbers are large. If one number is large and not the other, that would indicate consistent play (could be consistently good or consistently bad) while small numbers for both would indicate inconsistent (but not streaky) play. From this we can see that if we want to express streakiness with just one number, the minimum of the average streak lengths for positive and negative plays should also do the trick but without any distinction between consistent and inconsistent play for non-streaky quarterbacks.

### 19Re: Week 5 Quick Reads

I don't thnk you can capture the suckitude that is Christian Ponder in numbers. He literally sucks the life of the offence, the defence, the special teams, the coaches and the fanbase. I suspect the GDP of the state of Minnesota took a dip for a few days upon hearing about him starting and then actually playing.

### 30Re: Week 5 Quick Reads

He doesn't even put out the appearance of being an A-hole, thus denying fans the pleasure of viscerally gaining pleasure from seeing a jerk professionally fail. It's like watching your accountant, who always struck you as a pretty boring but non-offensive fellow, being thrust into the role of starting qb for the football team you root for.

### 31Re: Week 5 Quick Reads

Doesn't Ponder have a big dome/anywhere else split? Meaning that playing in a dome he's competent (or better) but outside for whatever reason he drops to poor if not inept.

### 35Re: Week 5 Quick Reads

I haven't seen that anywhere - it would surprise me because he's been awful so many times at home and in Detroit.

### 38Re: Week 5 Quick Reads

Ponder looked liked a decent NFL qb when playing at home, when Adrian Peterson was commanding attention from 10 of the 11 defenders on the other side of the ball. Unfortunately, half the games are away, and Peterson appears to be a feloniously violent crappy father.

I'll note again that Packer and Colts fans, they of the HOF to Pro Bowl quarterbacked squads, for what likely will end up being about 25 years straight, are, er, worthy of envy.

### 53Re: Week 5 Quick Reads

Ponder is Mark Sanchez. He's a very nice man who doesn't happen to be very good at football. However, after watching the New England game I'm convinced he's better than Matt Cassel.

### 78Re: Week 5 Quick Reads

Well, it is like arguing whether you want to hit your right or left thumb with a hammer, but I have to disagree. I think Cassell outplayed Ponder last year.

### 96Re: Week 5 Quick Reads

Huh? Ponder is INCREDIBLY GOOD at football. HE'S A STARTER IN THE NFL! I don't know where you get your standards but you should take them back.

### 99Re: Week 5 Quick Reads

Not anymore he won't be.

### 102Re: Week 5 Quick Reads

He hasn't been a starter since last year. He was clearly 3rd on the depth chart.

### 34Re: Week 5 Quick Reads

when the Vikings drafted him and the brain trust spoke of the reasons - I commented that those are all great skills if you were choosing a banker.

### 25Re: Week 5 Quick Reads

Something that might interesting to review (though far more time consuming) is to see if "streaks" continue less often if the offense is off the field for a longer period of time.

For example, a QB has a streak going and then goes of the field, then the defense gets a three and out. Is he more likely to keep the streak alive than if the other teams offense goes on a long, 12 play drive?

### 36Re: Week 5 Quick Reads

The blurb for Demaryius Thomas says he had three touchdown catches, but the chart (and the box score) say two. Or is he getting credit for one that was taken back on penalty?

### 45Re: Week 5 Quick Reads

Well, if THAT's the case, then Percy Harvin had a truly monster night! The penalty-nullified plays cannot count. Well, they shouldn't, at any rate.

### 43Re: Week 5 Quick Reads

This is a bit off topic. Last night Gruden said KJ Wright of Seattle was possibly the best LB in the league. Prior to this I was unaware of Mr Wright's alleged awesomeness. Anyone who watches a lot of the Seahawks agree or disagree?

### 48Re: Week 5 Quick Reads

He's probably the best at covering screens, and he helped shut down Graham in the first Saints game last year, but he is very slow for a LB and will get exploited because of this by faster receivers in coverage, so there's no way that he's anywhere near the best. Hell, his teammate Wagner is a better linebacker than him.

### 52Re: Week 5 Quick Reads

/agree.

Wright is a very good cover linebacker. But he's about the 8th or 10th best LB in the division:

Wagner;
Willis;
Bowman;
Aldon Smith;
Daryl Washington (when he's on the field, of course);
James Lauriniatas;
Sam Acho (Now on IR);

etc.

### 59Re: Week 5 Quick Reads

I don't think you should lump aldon smith into that mix - since he's really a linebacker in name only.

As for wright - he's really good, but there is still an element of "is it the players around him" thing going on. I suspect he's at minimum above average but proclaiming more than that definitively is just hard.

### 62Re: Week 5 Quick Reads

I'm not sure how much his teammates would be able to make him look better. Seattle doesn't blitz often, and Wright is probably asked to blitz the least because of his lack of speed, so it's not like he's getting pressures because of his teammates. He's in zone coverage most of the time, so there's no one helping him there either. Certainly a good pass rush can make any passes to his man less accurate, but that has nothing to do with his coverage ability itself.

### 64Re: Week 5 Quick Reads

Zone coverage effectiveness relies on players to either know when to pass off routes to another or when to jump a specific route. That relies a lot on play recognition and trusting your other defenders. I believe the linebackers are less exposed in matchups because there are other great players in seattle's secondary.

I am probably a heretic on this view, but I actually think Chancellor is the 2nd most valuable defender on Seattle and I do believe he helps the linebackers in their zone coverages.

### 70Re: Week 5 Quick Reads

Well, if you want to analyze someone's coverage ability, you look at how often they do get exposed, not whether those occasions get covered up by others. Here's an example: http://www.fieldgulls.com/football-breakdowns/2013/12/6/5176254/seahawks-earl-thomas-pete-carroll-nfl-saints

Look at the second play detailed there. Sherman and the pass rush cover for Wright getting beat by his man, but he's beaten all the same.

As for Chancellor, his role on the defense is so different from Sherman and Thomas that it's hard to say. Sherman and Thomas are there to prevent QBs from throwing at them (although Thomas is overrated in coverage, since QBs rarely challenge him; he was beaten on the second Jackson bomb and again on a potential TD pass that Cousins overthrew, and he dropped yet another interception) and Chancellor is basically a fourth linebacker.

### 71Re: Week 5 Quick Reads

That's fair, though I feel like Chancellor is the free lancer who allows the cover 3 to work as well as it does. Pound for pound, no one does his job better than Sherman on the defense, but I feel like Chancellor's work is highly under appreciated and his abilities make him a viable player regardless of what type of offense you are running.

### 72Re: Week 5 Quick Reads

Thanks for the feedback. I should have known better than to take anything Gruden says seriously.

### 65Re: Week 5 Quick Reads

Gruden is the worst commentator when it comes to over-praising whoever happens to be playing in front of him at the moment. I doubt it'll take many weeks for him to utter a completely contradictory statement.

### 68Re: Week 5 Quick Reads

Agreed. Gruden comes off like a complete moron 90 percent of the time. I know he's a very shrewd man so I can only imagine its something about being a broadcaster that turns your football brain into mush.

### 83Re: Week 5 Quick Reads

I recall Gruden being genuinely excellent his first year in the MNF booth, basically spouting off all sorts of proverbially hard-core football concepts. Blocking schemes, routes, all sorts of things that I loved hearing. That stopped in year two; I’m assuming somebody sat him down and told him to dumb things down. That, or I suspect somebody may have just transplanted Dan Dierdorf’s brain into his head.

### 85Re: Week 5 Quick Reads

Definitely the former. You see it in his QB camp as well. The first couple year's, it was hardcore football. A lot of play breakdown and film study (I believe the first year was 2010, when Bradford and Tebow were drafted).

Over the past couple years, it's become far less intense and more about Gruden and QBX talking while some clips of him being awesome go on behind.

### 55Re: Week 5 Quick Reads

I am trying to figure out how 18/24 for 201 yards and 2 TDs is only worth 4 passing DYAR. Er?

I agree that Wilson had a great game on the ground, but I thought he did a fine job of throwing the ball, too.

Three sacks and (officially) one fumbled snap on a passing play have a lot to do with it. He also had a lot of short-yardage problems. He had nine plays with 5 yards or less to go for a first down (seven of those on third downs) and picked up a first down just twice. That’s making the least out of a very good situation.

The blurb for Demaryius Thomas says he had three touchdown catches, but the chart (and the box score) say two. Or is he getting credit for one that was taken back on penalty?

Oops. I think my search for “D.Thomas” accidentally credited Demaryius with the touchdown that DeAnthony scored for Kansas City. Will fix.

### 60Re: Week 5 Quick Reads

While he had an excellent game, it is a little unfair to credit him with touchdowns scored by a different team. :-) So, yes, thanks for fixing that....

### 79Re: Week 5 Quick Reads

I'm also confused how the blurb for Jake Locker says that he had more value running than passing, but had 33 DYAR and 22 rushing DYAR.

### 81Re: Week 5 Quick Reads

I think it means he had a better DVOA running than passing.

### 91Re: Week 5 Quick Reads

That, to be honest, was written badly. His passing stats were nothing special, but his running stats were very good in a small sample size. He had more DYAR passing because he had more passing plays, but what he did in limited action as a runner was more impressive.

### 56Re: Week 5 Quick Reads

Cutler only threw four deep passes against Carolina.

Starting a rookie at LT (who hasn't even played that well at guard) will do that.

I personally thought Cutler was going to get seriously injured in that game when I heard that Bushrod was out and Ola would be filling in for him. But the Bears coaches did a fantastic job of calling almost all quick-hitting stuff and slowing down the pass rush with screens. It was only on the last drive, when everyone knew they had to throw downfield, that things fell apart protection-wise.

### 92Re: Week 5 Quick Reads

Interesting that the pressure on that last drive came up the middle, where they had gotten Slauson back.

### 93Re: Week 5 Quick Reads

The final sack was a twist on the left side. Slausen followed his guy, and Ola passed off like they would switch. Exact type of miscommunication you would expect with two guys who never played next to each other before.

### 88Re: Week 5 Quick Reads

Vince - It seems unusual for only two QB's in a given week to top 100 DYAR. How often does this actually happen?

### 89Re: Week 5 Quick Reads

There were at least three every week last season, usually about five or six. It was not a good week for quarterbacks, no.

DVOA DATABASE: Exclusive Access
• Historical Lookups
• Single Game DVOA by
• Stat Splits by
• Defense
NFL DVOA Database
• Overall DVOA
• Positions
• Line Stats
NCAA Football Stats
In-Season
Fantasy
NCAA
Offseason