Shotgun used more than half the snaps in 2007-2008

[unoriginal]

1. I wouldn't assume that. We don't have those 06 stats, and it's also possible the 06 stats are skewed because Brady took more time looking for receivers (and taking hits) since they weren't as good. If this were the case, it's odd that the hit number would still be high in 07, but it was...suggesting it wasn't a talent issue but a style/scheme issue.

We must assume the Pats took more shotgun snaps in 2007 than in 2006, because Football Outsiders said it, in plain English.

2. I disagree about run/pass ratio determining whether a team is a pass-emphasis team or not. Almost all teams pass more than run, even the run-emphasis teams. I would argue that those other QB's on that list also use predictable (meaning defense knows it's a pass) spread formations often, and just as when we are in shot gun, the opposing defense knows that it is highly likely to be a pass play and can tee off on the pass rush.

Yes, we agree most teams pass more frequently than they run, but that does not mean the stat is not informative. It is much more informative than the stats regarding number of shotgun plays and number of runs from shotgun, which we don't have and therefore tell us nothing.

But on a football fundamental level, the offense's set is less an indication of run and pass then it is of a play tree, or side tendencies. As a defensive player in high school and college our scouting reports included the favorite plays a team liked to run out of each set, along with play percentages to strong and to weak, but I don't recall ever being told "if they line up in unbalanced right with a wingback don't bother playing pass, they're running the ball."

As a defensive coordinator, if you are gambling on pass you are basing that off of pass/run ratio, down and distance and personnel groupings: things you know before the teams line up.

Do you agree, or disagree, that it's possible that when the opposing defense knows a formation or play has an extremely high chance of being a pass, that it makes their pass rush more effective?

In lieu of more specific formation stats, we can't pretend to scout the Pats on this. Here, incidentally, are the pass/run ratios coupled with the % assaulted stat I just made up:

2006		Pass	Run	Pass %	% QB Assaulted
[B]New England	527	499	51%	14%[/B]
Cincinnati	523	435	55%	16%
Arizona		545	419	57%	21%*
Green Bay	630	431	59%	10%

2007		Pass	Run	Pass %	% QB Assaulted
[B]New England	586	451	57%	11%[/B]
Seattle		590	430	58%	12%
Cincinnati	575	417	58%	10%
Arizona		590	402	59%	16%
Green Bay	578	388	60%	9%

*stat only for Warner, who had less than half the QB snaps this year

Among the teams we have stats for, the Pats passed the least each of these years. We also see that all these teams proportionally passed more in 2007 while their QBs were less frequently hit. There does not appear so far to be a strong correlation between how frequently a team lines up in the shotgun and how frequently the QB is assaulted, nor does there appear to be correlation between how frequently a team passes and how often the QB is assaulted.

What we do have is correlation that an increase in shotgun snaps means a decrease in running plays. But this correlation isn't much use, because we aren't arguing whether running plays are more effective than passing plays.

 

[unoriginal]

Yes. Deep passes are more like the home run analogy, and we do oftentimes run slow developing plays out of multiple formations. The shot gun, the formation we use by far, has more slow developing deep plays than quick ones. Why do I suspect the shot gun is mostly deep or slow-developing passes? It makes no sense to throw a quick slant from shot gun when you can just do it from under center, with negligible risk of being sacked/hit that fast, and with more space for the receiver to operate since more defenders are on the line.

Like most assumptions, this one turns out to be incorrect:

New England led the NFL with 6.2 average yards after catch, more than half a yard better than the Saints, who finished second. This had a lot to do with the one area Cassel found exceedingly difficult all year — the long bomb to Randy Moss. To compensate, the Pats led the league in receiver and tight end screens, using the plays 30 percent more than any team. This should even out with Brady's return — New England was a YAC team before, but not to such an extreme degree.

Given that we know the New England average 2008 YAC (6.2, above) and we know yards per reception was 11.2, we know the depth the average 2008 New England pass reception was caught at: 5 yards downfield. Not terribly deep.

Furthermore, in 2007 and 2008, our leading receiver was Wes Welker. Here are his yards per reception, yards after catch, and calculated yards before catch:

Year	Y/R	YAC	[B]YBC[/B]
2007	10.5	5.7	[B]4.8[/B]
2008	10.5	6.8	[B]3.7[/B]

It would seem much of our shotgun offense is geared towards passes at or under 5 yards.

 

[RDS11]

This isn't categorically true. This is a theory you have which can't be backed up. I can point to the "total QB hits while throwing" stat and you can say the absolute number doesn't matter, only the rate matters. I could argue the 4-WR or shot gun formations are not better than I-formation or single-back, then you'd point to some superior yards-per-average-play number, which I don't think is the right metric or end-all for offense. Under your 'efficiency means everything' philosophy, the Pats should just run shot gun 100% of the time since it gets more average yards... which is silly.



I agree with you on theory. Unfortunately, I operate in the realm of reality, and in the face of facts and statistics, don't close my eyes or dismiss things that go against my worldview. Brady has gotten hit while throwing the most, and it just so happens the Patriots run the most shot gun out of anybody. Most normal people would want to investigate this instead of simply dismiss it. I do acknowledge that there aren't enough stats going back from 2001-2009 which can fully explain what I am saying (if there were stats on it, it wouldn't even be debatable).



Again, this is theory that isn't backed up by reality as it pertains to the Patriots. I agree with you that shot gun lets the QB see more of the field more quickly, but it doesn't necessarily protect the QB more, especially against defenses which KNOW that when we are in shot gun we usually pass it. When you have a fearless QB like Brady, he's going to take the hit if it means letting a play develop more and even if he sees where the blitz is coming from.

First, I must say that I'm a statistician, that's what I do for a living. So when you say 'I operate in the realm of reality, and in the face of facts and statistics' I'm sorry, but I have to laugh at that.

First of all, volume statistics are useless if you don't put them in context. Just a small example so you'll understand : take 2 towns, and look at the number of car stolen :
Town A : 250 stolen cars in 2007
Town B : 9,500 stolen cars in 2007
Some people might think than town B is a lot less secure. But now if we say than town A is Walla Walla and town B is New York, it gives better perspective.

Second, one cannot take a fact and pinpoint the cause without looking at different aspects. That's what we call multivariate statistics. Again, getting back to the stolen cars, we might want to look at the crime rate, the average income, number of policemen per 1000 lives and so on. Then we would look at all those variable and see if they are related (and if they are, to which degree) to the final results.

Now, getting back to football, assuming that the shotgun formation is causing the high amount of hits Brady is taking is wrong on 3 levels :
1- The data for the numbers of hits you used is incomplete and and not well defined, which is a major sin in statistics. Before saying you got that from a reliable soucre (ESPN) I must tell you that in college one exercise we used to do was to take the newspapers and write about the lack of methodology, misleading information or conclusion that were absolutely not related to the data. Believe me, I can now look at data provided by the media and see those who are useless or misleading...
2-You are using volume statistics without giving a proper context. Again a big no-no.
3- You are assuming that the shotgun formation is the cause of the number of hits, which again can't be proven and moreover, might not be the case.

NAd finally, no I'm not saying that the Pats should run the shotgun formation 100% of the time since it produces an higher yards per play average that from under the center. In order to know what is the right balance, one would have to start at the game by game level, look at the percentage of shotgun snaps and and the average yards per play it generated. That would be a good start.

In conclusion, football is a contact sport. Hits are unavoidable. So that said, I will always go for the offense which produces the highest yards per play average. Always.

 

[maverick4]

As a defensive coordinator, if you are gambling on pass you are basing that off of pass/run ratio, down and distance and personnel groupings: things you know before the teams line up.

I appreciate your evidence-based approach, let's see if we can figure this out.

Regarding the quoted portion above, this is essentially what I am saying, or trying to say. Looking at run/pass ratio of ALL plays is meaningless in figuring out which teams are "run-emphasis" teams, because almost all teams pass more than run, so even most run-emphasis teams won't be found just looking at total pass/run ratio.

However, look at the pass/run ratio for personnel groupings. The article tells you how much we passed, and how many pass attempts were out of shot gun. It shows that an overwhelming ratio of the time out of shot gun, we PASS the ball.

 

[maverick4]

First of all, volume statistics are useless if you don't put them in context. Just a small example so you'll understand : take 2 towns, and look at the number of car stolen :
Town A : 250 stolen cars in 2007
Town B : 9,500 stolen cars in 2007
Some people might think than town B is a lot less secure. But now if we say than town A is Walla Walla and town B is New York, it gives better perspective.

Second, one cannot take a fact and pinpoint the cause without looking at different aspects. That's what we call multivariate statistics. Again, getting back to the stolen cars, we might want to look at the crime rate, the average income, number of policemen per 1000 lives and so on. Then we would look at all those variable and see if they are related (and if they are, to which degree) to the final results.

I am actually educated in stats as well and use it when looking at clinical research studies. I agree with you that in general there are many other factors that need to be looked at to reach any definitive conclusion. Unfortunately as you've discovered yourself, there isn't that much information on the subject, and I'm only presenting it as something interesting and something to look at.

For the same reason it's very difficult to statistically prove the link between Brady leading the league in "hits on QB while throwing" and with the fact that the Pats were the only team ever to have over 50% of snaps from shot gun for at least two years, it's also difficult to disprove it as well.

 

[maverick4]

Given that we know the New England average 2008 YAC (6.2, above) and we know yards per reception was 11.2, we know the depth the average 2008 New England pass reception was caught at: 5 yards downfield. Not terribly deep.

It would seem much of our shotgun offense is geared towards passes at or under 5 yards.

More nice analysis from you, which I appreciate. I appreciate that you are also trying to figure out this problem, rather than debate from a point of personal defensiveness or to point the blame at any individual. It isn't the goal, I'm just trying to figure out if there is a problem that needs to be fixed.

Regarding the quoted portion, we don't know whether average depth of 5 yards is above, below, or at the league norm. We need context as to what that number means, which we don't have. We also don't have info about whether the average depth is longer or shorter when we use shot gun. Those numbers are like the total run/pass ratio numbers, it doesn't tell us much, we need a break down of what it looks like under shot gun.

 

[maverick4]

Among the teams we have stats for, the Pats passed the least each of these years. We also see that all these teams proportionally passed more in 2007 while their QBs were less frequently hit. There does not appear so far to be a strong correlation between how frequently a team lines up in the shotgun and how frequently the QB is assaulted, nor does there appear to be correlation between how frequently a team passes and how often the QB is assaulted. What we do have is correlation that an increase in shotgun snaps means a decrease in running plays. But this correlation isn't much use, because we aren't arguing whether running plays are more effective than passing plays.

1. I agree with you, which is why in the original statement I argued that it had to do with style of play, rather than some simplistic measure of total pass/run ratio cross-matched with QB hits. As an example, the reason the 01 Pats beat the Rams is because they noticed that the explosive stats-breaking Rams ran a disproportionately high number of plays for slants, in-cuts, and passes to Marshall Faulk. They chipped Faulk using linebackers, and destroyed receivers catching slants and in-cuts in the Superbowl. The reason that offense was defeated was because they kept using the same formations/plays over and over, even though the defense knew what was coming. I argue that this same stubborn-ness currently applies to the Pats as well, in which opposing D's KNOW we will pass when we are in shot gun.

2. The problem is you're still looking at things from overall run/pass ratio, which is mis-guided, rather than trying to figure out whether it is obvious for an opposing defense to know what when an offense lines up in a certain formation or personnel, that it makes things easier for them to pass rush or hit the QB. Pass-emphasis teams try to win the game by passing, then running to kill the clock. We don't have the in-game data over all recent games and years, but I would argue that the Patriots (as are the other teams on that QB hit list) are pass-emphasis teams which try to win games by passing firs, and run just for balance.

 

[Rob0729]

I am actually educated in stats as well and use it when looking at clinical research studies. I agree with you that in general there are many other factors that need to be looked at to reach any definitive conclusion. Unfortunately as you've discovered yourself, there isn't that much information on the subject, and I'm only presenting it as something interesting and something to look at.

For the same reason it's very difficult to statistically prove the link between Brady leading the league in "hits on QB while throwing" and with the fact that the Pats were the only team ever to have over 50% of snaps from shot gun for at least two years, it's also difficult to disprove it as well.

If you agree that there aren't enough stats or context around the stats to support your argument, why bring it up? You are arguing like it is strong stats and now admitting that the stats could be out of context and if put into context could be the exact opposite what you are arguing.

 

[maverick4]

As a defensive coordinator, if you are gambling on pass you are basing that off of pass/run ratio, down and distance and personnel groupings: things you know before the teams line up..

Wouldn't you agree that opposing defensive coordinators do have this data, and do know (as we do based on the little info we have) that based on how many times we pass, and how many times we pass out of shotgun (provided in the original article), that they ARE able to key on us when we are in shot gun?

 

[maverick4]

If you agree that there aren't enough stats or context around the stats to support your argument, why bring it up? You are arguing like it is strong stats and now admitting that the stats could be out of context and if put into context could be the exact opposite what you are arguing.

It's an interesting statistic, I think the "hits on QB while throwing" stat and the "over 50% of all plays out of shotgun" stat and the "most Patriot passes are out of shotgun" stat are legit. Those stats are strong, but the other supporting stats are not strong enough or non-existent. On the flip side there isn't enough information for you to disprove it, and I acknowledge there isn't a definitive way to prove it either.

However, I don't see you applying the same rigorous statistical standard to pretty much any other thread or topic on this board.

 

[maverick4]

[For the Pats in the recent past, it’s been all about the shotgun formation. In 2007, New England became the first NFL team to run the shotgun formation on more than 50 percent of their plays, and that trend continued in 2008. Our data shows that teams are more effective and efficient in the shotgun – over the last two years, teams have averaged 5.9 yards per play from the shotgun, and 5.1 under center. Cassel threw 433 passes out of the shotgun and amassed a DYAR [Defense-Adjusted Yards over Replacement] of 531, and a DVOA of 7.8%. Under center, he threw 124 passes for a DYAR of 124 and a DVOA of 2.2%. ]

In 2008 we passed 534 times, and from the article 433 of those passes were from shot gun. We ran 513 times, for 1034 total plays. When we are in shot gun, the vast majority of plays are passes. Is this, or is this not, a huge giveaway to the opposing defensive linemen and linebackers?

It's as big a giveaway as when the Pats defeated the 01 Rams based on their repeated tendency for slants, in-cuts, and outlets to Faulk.
New England Patriots Stats

Shows how much we passed even in 2005 and in 2006, with less talented WR's:
http://www.patsfans.com/new-england...d/10/45643-run-attempts-vs-pass-attempts.html

[With the reemergence of the 2007 playbook and 500-plus yards in each of the last two games, the Patriots offense, again almost entirely a PASS-FIRST OFFENSE, is looking very much like a record-setting group once more.]
Foresight: Josh McDaniels passes Patriots by vaunted Steelers defense - Eye on Foxborough - MassLive.com

 

[Rob0729]

It's an interesting statistic, I think the "hits on QB while throwing" stat and the "over 50% of all plays out of shotgun" stat and the "most Patriot passes are out of shotgun" stat are legit. Those stats are strong, but the supporting stats are strong enough. On the flip side there isn't enough information for you to disprove it, and I acknowledge there isn't a definitive way to prove it either.

However, I don't see you applying the same rigorous statistical standard to pretty much any other thread or topic on this board.

Except you are ignoring the analysis from some of the sources that you quote those stats and form your own directly opposite opinion. The proof against some of your arguments is that statistically, the Pats have had the most successful offense by most statisical breakdowns over 1 1/2 of the last two years. As for the number of times hit, we have repeatedly argued not that it is wrong, but it is a bogus stat since there is no context to them.

 

[efin98]

This thread still being alive has me thinking these questions that I demand an answer to from the resident McDaniels basher-

Since you are so against the shotgun formation yet have yet to provide reasoning nor facts to back up your stance: what are Brady's completion percentages under center vs. in the shotgun? What are his hit numbers in shotgun vs. under center? what are his percentage of running plays vs. passing plays?

Answer those questions. If you can't answer them with actual numbers then you have no basis for the stance against the shotgun except your own personal bias against it

Bringing this back up for an answer...I'm waiting

 

[jmt57]

The way I see it, Ernie Adams has probably spent the last five months compiling a database on offensive formations, down-and-distance, success rates, chance of Brady being touched, likelihood of Brady being hit, possibility of failure for any given play, best and worst play call versus specific opponent's defensive alignment, and dozens of other categories against each of this year's thirteen opponents.

In other words, rather than assuming the offense will be identical to 2007, perhaps the Pats have taken Brady's injury in to consideration in planning for the 2009 season.

 

[Deus Irae]

The way I see it, Ernie Adams has probably spent the last five months compiling a database on offensive formations, down-and-distance, success rates, chance of Brady being touched, likelihood of Brady being hit, possibility of failure for any given play, best and worst play call versus specific opponent's defensive alignment, and dozens of other categories against each of this year's thirteen opponents.

In other words, rather than assuming the offense will be identical to 2007, perhaps the Pats have taken Brady's injury in to consideration in planning for the 2009 season.

I sure as heck hope that they haven't. If they're taking the injury into consideration, it means that he's not 100%.

 

[efin98]

I sure as heck hope that they haven't. If they're taking the injury into consideration, it means that he's not 100%.

Agreed.

The only things that should be taken into consideration are his personal views on the plays- how would he have called it, how would he have adjusted, how would he have done the same.

 

[jmt57]

I sure as heck hope that they haven't. If they're taking the injury into consideration, it means that he's not 100%.

I was looking at it from the perspective of reducing chance of a new injury while still maintaing high probability of a successful play, as part of their self scouting and preparation for the upcoming season. I wasn't suggesting the Pats would be adjusting the playbook due to a player's being less than 100% healthy, though I can see how my words came across that way.

 

[efin98]

I was looking at it from the perspective of reducing chance of a new injury while still maintaing high probability of a successful play, as part of their self scouting and preparation for the upcoming season. I wasn't suggesting the Pats would be adjusting the playbook due to a player's being less than 100% healthy, though I can see how my words came across that way.

But that's assuming they didn't do that already after the Chiefs game...which they almost certainly did as part of their basic game tape breakdown sessions. He might have a different perspective of the adjustments needed, but they had to have already made adjustments.

 

[jmt57]

But that's assuming they didn't do that already after the Chiefs game...which they almost certainly did as part of their basic game tape breakdown sessions. He might have a different perspective of the adjustments needed, but they had to have already made adjustments.

After the Chiefs game adjustments were made for a Matt Cassel-led offense. Now the team is back to a Tom Brady-led offense, which is what this very long thread has been going back and forth about. There are a couple of forum posters who very strongly believe the Pats should line up in the shotgun less often, and use four or even three WR formations less often as well. I was subtlely attempting to suggest that the Pats have probably already looked back at the pros and cons of those (and many other) ideas for the upcoming season - with a whole lot more data available at their disposal for analysis than what we have.

For some reason there seems to be a lot of assumption by people on both sides of the debate that the Pats absolutely will do this or won't do that. I don't recall Belichick and his staff having that kind of bull-headed "we're going to do it this way, end of discussion' approach to things like a specific offensive formation since taking over in 2000.

 

[Deus Irae]

After the Chiefs game adjustments were made for a Matt Cassel-led offense. Now the team is back to a Tom Brady-led offense, which is what this very long thread has been going back and forth about. There are a couple of forum posters who very strongly believe the Pats should line up in the shotgun less often, and use four or even three WR formations less often as well. I was subtlely attempting to suggest that the Pats have probably already looked back at the pros and cons of those (and many other) ideas for the upcoming season - with a whole lot more data available at their disposal for analysis than what we have.

For some reason there seems to be a lot of assumption by people on both sides of the debate that the Pats absolutely will do this or won't do that. I don't recall Belichick and his staff having that kind of bull-headed "we're going to do it this way, end of discussion' approach to things like a specific offensive formation since taking over in 2000.

I also don't recall the team refusing to use the shotgun out of fear of injury, and the increased use of it during Cassel's tenure seems to pretty much put the kibosh on such a notion.