Great article! Cousins was 2nd in the league in yards per attempt but threw the 6th fewest passes. My solution?? Jefferson targets go up. Cooks targets go up and his carries go down. None of that requires a scheme change. Yes turnovers hurt, but 40 yard TD passes to Justin Jefferson and Adam Theilen can also change the game as well.
It's very surprising to me that they don't want to send Cook out on routes, ever. Who remembers this video that highlighted his abilities as a route runner and ball catcher?
I wonder if the lack of quick game has to do with how often we’re under center. I could be wrong, but I feel like quick game passes usually come from the shotgun, and MN has led the league in % of snaps from under center the last 2 years. So while I would love to see more quick game out of the gun, I imagine they’re skeptical of that because it takes away the threat of the play action boots, etc.
Bingo. On a related note, the few times Dalvin motions out of the backfield and they’re left in empty formation, and Kirk stays under center... my blood pressure rises EVERY time. That makes no sense to me lol.
But yeah, I feel like they’re in a “damned if you do, damned if you don’t” situation when it comes to the amount of quick game ran versus focusing on under center play action drop backs.
Please stop running on 2nd and long! Be more aggressive early! We only get aggressive when it’s too late! Of course, we won’t have to worry as much about how many points our offense has to put up if we can get some improvement on the defensive side of the ball. Good read!
I agree with the concept of running less on early downs for lots of reasons, but also I do feel like the overall conversation gets away from the facts sometimes (not necessarily here, but in general).
On one hand, it frankly does not make a lot of analytic sense to run on early downs with long yardage. The stats show that running is consistently less efficient, and as Coller pointed out the two teams in the super bowl were both top three when it comes to passing rates on 2nd-and-long. While I want us to be a team that runs plenty given that we have Dalvin Cook and given how elite our OL is at run-blocking (e.g., such that it would be dumb to have our OL ignore the only thing that it has consistently been good-to-great at, which to be clear is not what I think Coller is suggesting), statistically the most consistent path to championships is being built to go through the air for explosive gains (which the Vikings are great at) and also intermediate and short gains (which the Vikings are not-so-great at), and to make defenses fear you going through the air in any situation. For all of these reasons, my typical response to seeing the Vikings call a run on 2nd and 8 is a long groan.
So that must mean that the top 10 teams in pass rate on 2nd and long are all of the best teams with all of the best success, and the bottom 10 teams are all middling teams with bad offenses, right? Nope. The top 10 teams on passing on 2nd and long (outside of TB and KC) are the Jags, Cards, Panthers, Niners, Raiders, Cowboys, Chargers, and Falcons. Put differently, outside of the two super bowl participants, no team that was in the top 10 in passing rate on 2nd and long made it into the playoffs, and the best team by Offensive DVOA (the best overall offensive metric in my book) of those 8 teams was the Raiders at 14th overall. That's not fantastic!
Conversely, the bottom 10 teams in pass rate on 2nd and long are the Vikes, Broncos, Ravens, Titans, Pats, Jets, Saints, Bills, Packers, and Colts. Clearly a lot of dysfunctional teams there, but 5 out of 10 are playoff teams, two of the 10 were the championship game losers, and that list includes 5 of the top 8 teams in offensive DVOA.
(This isn't an example of 2020 being flukey, either - in 2019 only 2 of the top 10 teams in pass rate on 2nd and long made it into the playoffs - both of which lost in the wild card round - and 5 of the bottom 10 teams in pass rate made it in, including 3 of the 4 teams in the championship games.)
Part of the problem is that we need to only look at times when the game is close, so that you can avoid situations where, e.g., a team runs because they up by 17 points with 3 minutes left on 2nd and 7 (which, to be clear, analytics does not always love, but certainly that is not a typical and/or useful data point). If you only look at times when the score margin is +/-10 points, the top team teams in pass rate don't change much as it now includes 3 playoff teams (as the Packers jump into the mix). HOWEVER, if you only look at times when the game is close the bottom ten teams in pass rate only includes 4 playoff teams (Baltimore, Washington, Tennessee, Chicago), none of which advanced past the divisional round and only one of which was in the top 10 offensively (Titans). (This also holds true for past years - in 2019 only 1 of the 4 championship teams was bottom 10 in pass rate on 2nd and long in close games, neither of the superbowl participants were bottom 10, but also only 3 of the top 10 teams got to the playoffs, with KC sneaking in there at #9).
What in the world are we to make of this? Obviously, if you have a QB that might be the GOAT, you put the ball in his hands as often as practical. If you do not have a potential GOAT, seemingly this indicates that you might not want to put the ball into his hands at a league-leading rate, whether the game is close or not. That said, while you may want to avoid topping the charts when it comes to 3rd down rate if you don't have a potential GOAT, you also seemingly want to avoid being in the BOTTOM of the charts, but rather the best option for a good-not-transcendent QB might be to be average when it comes to your rate of throwing (which in the NFL means passing around 69% of the time, nice).
What we're to make of this is that a lot of teams are doing a thing that's not the most efficient and one way to get ahead of them is to do something more efficient. And absolutely score effects are a big deal when it comes to raw rates. The Vikings were playing from behind a lot and still ran all the time on 2nd-and-8, as opposed to the Packers who were running at the end of games. The 4 teams in the playoffs you mentioned that were at the bottom, 3 of them are bad at passing the ball. The Vikings had a passing game with a good success rate and very high Y/A and lots of weapons to use. So it's not really about which teams made the playoffs and how that correlates with 2nd and long runs because making the playoffs has a ton of factors (like division ahem Washington) but it's about whether it's a potential area for statistical advantage... which I think it clearly is...and I don't think Kirk has to be Brady or Mahomes for that to be used.
TL:DR - I agree that the Vikings should avoid being near the bottom of the league, but I do not think that the numbers unambiguously declare that the Vikings should be near Mahomes or Brady rates of throwing, even though I think that it is analytically advisable for Mahomes and Brady to keep throwing at these extreme elevated rates.
Long story long, let me clarify - I agree that if you have Mahomes or Brady or Rodgers or the like that the best idea is to throw well above league average, and I agree that the best chances of advancing far into the playoffs is having one of those guys and putting the ball in his hands at a very very high rate. For example, I think that it is extremely dumb that the Texans were not in that top 10 list of throwing rates, because I think that Watson is a Mahomes or Brady level talent, and the best way for any team with Watson to win is to make sure that Watson decides where the ball goes in the overwhelming majority of plays.
Additionally, I agree that it is not good to be well below average when it comes to how often you throw, as, even if you get into the playoffs, that means that you will probably not go far as you indicated, because this probably means that your offense wasn't that good (or at least was very likely to be game-script dependent).
All that said, it may certainly be that every team outside of the teams that have Brady and Mahomes and Rodgers are exclusively and consistently doing the wrong thing, and that it is nothing but bizarrely repeatable noise that all of the teams that pass well above average on passing downs in close games with average (or merely good) QBs are not efficient offensively, and that it is just a random happy coincidence that the only teams that pass far above average in close games with efficient offenses also have a sure-fire first-ballot HOF QB.
However, at some point it feels silly to just ignore the numbers that say that throwing far above average with a good-not-elite QB does not work out, but rather makes you less efficient than other teams with good-not-elite QBs that are merely throwing it at average rates (or perhaps slightly more than average rates). Basically, the numbers seem to indicate that there is a ceiling at which you can pass and have success if your QB is something less than a generational talent. To be clear, that ceiling is FAR above what the Vikings are doing (the average rate of passing on 2nd and long when the game is close across the NFL is 69% compared to 58% for the Vikes), such that I think the Vikings should pass far far more on downs like 2nd and long. But that ceiling still seems to exist, and it seems to exist at a level below what some NFL teams are currently doing, and well below what the ceiling is for QBs like Mahomes and Brady. Basically, I don't want the Vikings to pass at a rate at which Mahomes or Brady do, cuz I am alarmed of what the numbers show about teams that throw that often without that type of QB.
Of course there's a connection with efficiency and throwing less because they're only trying to throw in the most advantageous situations like 1st and 10 play-action but it's not like Cousins is so far below good QBs that he couldn't throw more. In 2016 Washington leaned into Cousins and threw 600 times and had the 2nd best net yards per attempt in the NFL and they had a top 5 scoring % while ranking 27th in rushing attempts. Not saying that I'd have them rank 27th with Cook but the question is whether there's room in the margins to make him better when it comes to something as straightforward as play selection and that seems to be yes.
Okay, I feel like we are both basically saying the same thing now. I agree that of course Kirk Cousins should throw more, and of course there is room in the margins to make him better. I would be willing to bet good money that the Vikings will never get to (much less win) the Superbowl if they continue to be bottom 10 in pass rate on downs like 2nd and long (much less 32nd like they are right now). If I had my way they would immediately jump from 58% to around 70%, and be around the 11th-16th rate, which includes a lot of great and smart teams and offenses (Seattle, Rams, Browns, Colts).
I always struggle with how to compare the Vikings to the 2016 Washington team. To be clear, it is always interesting and worthwhile to compare the two, obviously, but it is hard to know how many lessons we can pull that apply in a one-to-one manner. To start, it is important to note that you are correct in that that 2016 Kirk offense was amazing and extremely efficient (#5 in offensive DVOA), though it is noteworthy that that team was not particularly consistent in their efficiency but was instead "only" 21st in offensive variance (such that the offense had a tendency to either explode or implode). But more importantly, I am just not positive that we will ever get the version of Kirk from 2016 when he was in McVay's system and had McVay speaking in his ear telling him what to look for before each snap. We'll see what Goff does this year, but I have a hunch that moving out of an extremely QB-friendly system like McVay's where McVay is literally telling you exactly where to throw it on regularly schemed-wide-open throws will result in Goff's efficiency plummeting. If your point is that therefore that we should have something like McVay's system and playcalling because that would make our offense more efficient, I would emphatically agree with you, but I also think that is about as likely as trading for Deshaun Watson. There are maybe 3 or 4 OCs in the league that do that year in and year out (frankly it is starting to be slightly debatable whether McVay is starting to lose his fastball, though having Stafford may change that in a hurry). The annually innovative OCs are unicorns, IMO, such that I am not positive that it is always useful to compare offenses that have them to offenses that don't. Every team should be consistently stealing ideas from the Shanahans and Reids of the world, obviously, but the second that you try to do exactly what they do for a full game (much less a full season) is often the second that you will fall flat on your face.
Basically, I don't think that the numbers indicate that there is literally no ceiling as to how often a very-good-but-not-transcendent QBs like Kirk can throw regardless of scheme while making the offensive efficiency increase. Instead, I think that there is a ceiling for those Kirk-esque QBs as to how often they can throw and increase efficiency in any given scheme, and my personal guess would be that for the Vikings/Kirk/Klint that ceiling is somewhere around 70% (and therefore far higher than their current 58 but also comfortably less than the 79% as shared by the Bucs and Chiefs) on early downs.
Great article! Cousins was 2nd in the league in yards per attempt but threw the 6th fewest passes. My solution?? Jefferson targets go up. Cooks targets go up and his carries go down. None of that requires a scheme change. Yes turnovers hurt, but 40 yard TD passes to Justin Jefferson and Adam Theilen can also change the game as well.
It's very surprising to me that they don't want to send Cook out on routes, ever. Who remembers this video that highlighted his abilities as a route runner and ball catcher?
https://youtu.be/oal9T61thY0?t=79
I wonder if the lack of quick game has to do with how often we’re under center. I could be wrong, but I feel like quick game passes usually come from the shotgun, and MN has led the league in % of snaps from under center the last 2 years. So while I would love to see more quick game out of the gun, I imagine they’re skeptical of that because it takes away the threat of the play action boots, etc.
I think that's def part of it. That might also be part of the cause of their interior pressure issue
Bingo. On a related note, the few times Dalvin motions out of the backfield and they’re left in empty formation, and Kirk stays under center... my blood pressure rises EVERY time. That makes no sense to me lol.
But yeah, I feel like they’re in a “damned if you do, damned if you don’t” situation when it comes to the amount of quick game ran versus focusing on under center play action drop backs.
Great article Mathew they need to use Dalvin Cook into the passing game
Please stop running on 2nd and long! Be more aggressive early! We only get aggressive when it’s too late! Of course, we won’t have to worry as much about how many points our offense has to put up if we can get some improvement on the defensive side of the ball. Good read!
I agree with the concept of running less on early downs for lots of reasons, but also I do feel like the overall conversation gets away from the facts sometimes (not necessarily here, but in general).
On one hand, it frankly does not make a lot of analytic sense to run on early downs with long yardage. The stats show that running is consistently less efficient, and as Coller pointed out the two teams in the super bowl were both top three when it comes to passing rates on 2nd-and-long. While I want us to be a team that runs plenty given that we have Dalvin Cook and given how elite our OL is at run-blocking (e.g., such that it would be dumb to have our OL ignore the only thing that it has consistently been good-to-great at, which to be clear is not what I think Coller is suggesting), statistically the most consistent path to championships is being built to go through the air for explosive gains (which the Vikings are great at) and also intermediate and short gains (which the Vikings are not-so-great at), and to make defenses fear you going through the air in any situation. For all of these reasons, my typical response to seeing the Vikings call a run on 2nd and 8 is a long groan.
So that must mean that the top 10 teams in pass rate on 2nd and long are all of the best teams with all of the best success, and the bottom 10 teams are all middling teams with bad offenses, right? Nope. The top 10 teams on passing on 2nd and long (outside of TB and KC) are the Jags, Cards, Panthers, Niners, Raiders, Cowboys, Chargers, and Falcons. Put differently, outside of the two super bowl participants, no team that was in the top 10 in passing rate on 2nd and long made it into the playoffs, and the best team by Offensive DVOA (the best overall offensive metric in my book) of those 8 teams was the Raiders at 14th overall. That's not fantastic!
Conversely, the bottom 10 teams in pass rate on 2nd and long are the Vikes, Broncos, Ravens, Titans, Pats, Jets, Saints, Bills, Packers, and Colts. Clearly a lot of dysfunctional teams there, but 5 out of 10 are playoff teams, two of the 10 were the championship game losers, and that list includes 5 of the top 8 teams in offensive DVOA.
(This isn't an example of 2020 being flukey, either - in 2019 only 2 of the top 10 teams in pass rate on 2nd and long made it into the playoffs - both of which lost in the wild card round - and 5 of the bottom 10 teams in pass rate made it in, including 3 of the 4 teams in the championship games.)
Part of the problem is that we need to only look at times when the game is close, so that you can avoid situations where, e.g., a team runs because they up by 17 points with 3 minutes left on 2nd and 7 (which, to be clear, analytics does not always love, but certainly that is not a typical and/or useful data point). If you only look at times when the score margin is +/-10 points, the top team teams in pass rate don't change much as it now includes 3 playoff teams (as the Packers jump into the mix). HOWEVER, if you only look at times when the game is close the bottom ten teams in pass rate only includes 4 playoff teams (Baltimore, Washington, Tennessee, Chicago), none of which advanced past the divisional round and only one of which was in the top 10 offensively (Titans). (This also holds true for past years - in 2019 only 1 of the 4 championship teams was bottom 10 in pass rate on 2nd and long in close games, neither of the superbowl participants were bottom 10, but also only 3 of the top 10 teams got to the playoffs, with KC sneaking in there at #9).
What in the world are we to make of this? Obviously, if you have a QB that might be the GOAT, you put the ball in his hands as often as practical. If you do not have a potential GOAT, seemingly this indicates that you might not want to put the ball into his hands at a league-leading rate, whether the game is close or not. That said, while you may want to avoid topping the charts when it comes to 3rd down rate if you don't have a potential GOAT, you also seemingly want to avoid being in the BOTTOM of the charts, but rather the best option for a good-not-transcendent QB might be to be average when it comes to your rate of throwing (which in the NFL means passing around 69% of the time, nice).
What we're to make of this is that a lot of teams are doing a thing that's not the most efficient and one way to get ahead of them is to do something more efficient. And absolutely score effects are a big deal when it comes to raw rates. The Vikings were playing from behind a lot and still ran all the time on 2nd-and-8, as opposed to the Packers who were running at the end of games. The 4 teams in the playoffs you mentioned that were at the bottom, 3 of them are bad at passing the ball. The Vikings had a passing game with a good success rate and very high Y/A and lots of weapons to use. So it's not really about which teams made the playoffs and how that correlates with 2nd and long runs because making the playoffs has a ton of factors (like division ahem Washington) but it's about whether it's a potential area for statistical advantage... which I think it clearly is...and I don't think Kirk has to be Brady or Mahomes for that to be used.
TL:DR - I agree that the Vikings should avoid being near the bottom of the league, but I do not think that the numbers unambiguously declare that the Vikings should be near Mahomes or Brady rates of throwing, even though I think that it is analytically advisable for Mahomes and Brady to keep throwing at these extreme elevated rates.
Long story long, let me clarify - I agree that if you have Mahomes or Brady or Rodgers or the like that the best idea is to throw well above league average, and I agree that the best chances of advancing far into the playoffs is having one of those guys and putting the ball in his hands at a very very high rate. For example, I think that it is extremely dumb that the Texans were not in that top 10 list of throwing rates, because I think that Watson is a Mahomes or Brady level talent, and the best way for any team with Watson to win is to make sure that Watson decides where the ball goes in the overwhelming majority of plays.
Additionally, I agree that it is not good to be well below average when it comes to how often you throw, as, even if you get into the playoffs, that means that you will probably not go far as you indicated, because this probably means that your offense wasn't that good (or at least was very likely to be game-script dependent).
All that said, it may certainly be that every team outside of the teams that have Brady and Mahomes and Rodgers are exclusively and consistently doing the wrong thing, and that it is nothing but bizarrely repeatable noise that all of the teams that pass well above average on passing downs in close games with average (or merely good) QBs are not efficient offensively, and that it is just a random happy coincidence that the only teams that pass far above average in close games with efficient offenses also have a sure-fire first-ballot HOF QB.
However, at some point it feels silly to just ignore the numbers that say that throwing far above average with a good-not-elite QB does not work out, but rather makes you less efficient than other teams with good-not-elite QBs that are merely throwing it at average rates (or perhaps slightly more than average rates). Basically, the numbers seem to indicate that there is a ceiling at which you can pass and have success if your QB is something less than a generational talent. To be clear, that ceiling is FAR above what the Vikings are doing (the average rate of passing on 2nd and long when the game is close across the NFL is 69% compared to 58% for the Vikes), such that I think the Vikings should pass far far more on downs like 2nd and long. But that ceiling still seems to exist, and it seems to exist at a level below what some NFL teams are currently doing, and well below what the ceiling is for QBs like Mahomes and Brady. Basically, I don't want the Vikings to pass at a rate at which Mahomes or Brady do, cuz I am alarmed of what the numbers show about teams that throw that often without that type of QB.
Of course there's a connection with efficiency and throwing less because they're only trying to throw in the most advantageous situations like 1st and 10 play-action but it's not like Cousins is so far below good QBs that he couldn't throw more. In 2016 Washington leaned into Cousins and threw 600 times and had the 2nd best net yards per attempt in the NFL and they had a top 5 scoring % while ranking 27th in rushing attempts. Not saying that I'd have them rank 27th with Cook but the question is whether there's room in the margins to make him better when it comes to something as straightforward as play selection and that seems to be yes.
Okay, I feel like we are both basically saying the same thing now. I agree that of course Kirk Cousins should throw more, and of course there is room in the margins to make him better. I would be willing to bet good money that the Vikings will never get to (much less win) the Superbowl if they continue to be bottom 10 in pass rate on downs like 2nd and long (much less 32nd like they are right now). If I had my way they would immediately jump from 58% to around 70%, and be around the 11th-16th rate, which includes a lot of great and smart teams and offenses (Seattle, Rams, Browns, Colts).
I always struggle with how to compare the Vikings to the 2016 Washington team. To be clear, it is always interesting and worthwhile to compare the two, obviously, but it is hard to know how many lessons we can pull that apply in a one-to-one manner. To start, it is important to note that you are correct in that that 2016 Kirk offense was amazing and extremely efficient (#5 in offensive DVOA), though it is noteworthy that that team was not particularly consistent in their efficiency but was instead "only" 21st in offensive variance (such that the offense had a tendency to either explode or implode). But more importantly, I am just not positive that we will ever get the version of Kirk from 2016 when he was in McVay's system and had McVay speaking in his ear telling him what to look for before each snap. We'll see what Goff does this year, but I have a hunch that moving out of an extremely QB-friendly system like McVay's where McVay is literally telling you exactly where to throw it on regularly schemed-wide-open throws will result in Goff's efficiency plummeting. If your point is that therefore that we should have something like McVay's system and playcalling because that would make our offense more efficient, I would emphatically agree with you, but I also think that is about as likely as trading for Deshaun Watson. There are maybe 3 or 4 OCs in the league that do that year in and year out (frankly it is starting to be slightly debatable whether McVay is starting to lose his fastball, though having Stafford may change that in a hurry). The annually innovative OCs are unicorns, IMO, such that I am not positive that it is always useful to compare offenses that have them to offenses that don't. Every team should be consistently stealing ideas from the Shanahans and Reids of the world, obviously, but the second that you try to do exactly what they do for a full game (much less a full season) is often the second that you will fall flat on your face.
Basically, I don't think that the numbers indicate that there is literally no ceiling as to how often a very-good-but-not-transcendent QBs like Kirk can throw regardless of scheme while making the offensive efficiency increase. Instead, I think that there is a ceiling for those Kirk-esque QBs as to how often they can throw and increase efficiency in any given scheme, and my personal guess would be that for the Vikings/Kirk/Klint that ceiling is somewhere around 70% (and therefore far higher than their current 58 but also comfortably less than the 79% as shared by the Bucs and Chiefs) on early downs.