I posted this on another forum:
In 2012, 7 QB's had 60+ carries. They combined for a 56-40 (.583) win-loss record.
In 2013, 9 QB's had 60+ carries. They combined for an exemplary 84-49 (.631) record.
Given the zero-sum nature of NFL games, any set of QB's who can post win percentages significantly higher than .500 likely bring significant competitive value to their teams. Is there a correlation between mobile QB's and win percentage? Is this simply a new trend with no historical basis?
Here, I will explore these possibilities by examining combined win-loss records of mobile QB's going back to the AFL-NFL merger of 1970. What I have discovered will shed light on the growing bias for mobile QB's among NFL coaches and general offices.
Definition: What is a mobile QB?
A mobile QB means different things to different people. Some define mobile QB's as run-first QB's with limited passing ability, such as Kordell Stewart, Tim Tebow, Tommie Frazier, and Pat White. Others define it as QB's with NFL-caliber passing ability but a tendency to run, such as Michael Vick and Randall Cunningham. Some exclude elite passers such as Steve Young, despite his 4200 career rushing yards, nearly as many as Cunningham. Others consider quick-footed but pass-first QB's like Aaron Rodgers as mobile while some don't.
For this study, I defined mobile QB in a number of ways. Each successive definition is more stringent and reduced the pool of QB's:
QB's with 60+ carries in a season. While this may seem low, only 9 QB's reached this measure in 2013 and all but one (Dalton) are assuredly mobile QB's (listed in order of carries):
Cam Newton, Russell Wilson, Colin Kaepernick, Robert Griffin III, Terrelle Pryor, Alex Smith, Geno Smith, Andrew Luck, Andy Dalton
QB's with 80+ carries in a season. This measure excludes Dalton, Luck, and both Geno and Alex Smith. This is a strict definition of mobile that tellingly excludes Geno Smith who is most definitely a mobile QB
QB's with 500+ rushing yards in a season. The most strict. By this measure, only Newton, Pryor, Wilson, and Kaepernick qualify as mobile during the 2013 season.
Methodology
My research looked at win-loss statistics for qualifying QB seasons using all three measurable definitions of mobile QB.
My sample included all QB seasons since the AFL-NFL merger in 1970. While many may believe mobile QB's are a recent phenomenon, the first qualifying QB using the strictest definition (500yd rushing season) was Detriot's Greg Landry who rushed for 530yds in 1971 in a 14-game season. The very next year, he ran for 524yds while Chicago's Bobby Douglass had a 968yd rushing season which stood as an NFL record for 34 years before being broken by Michael Vick in a 16 game season. The first mobile QB to have a winning season was New England's Steve Grogan in 1978 who rushed for 539yds on way to an 11-5 season.
Data was obtained from Pro Football Reference. Data sets are included below:
60+ rush attempt QB seasons since 1970 sorted by wins (112 QB seasons)
80+ rush attempt QB seasons since 1970 sorted by wins (40 QB seasons)
500+yd rushing QB seasons since 1970 sorted by wins (33 QB seasons)
Results
60+ rush attempt QB seasons since 1970
Combined record: 919-686-11 (.568)
80+ rush attempt QB seasons since 1970
Combined record: 349-253 (.579)
500yd rushing QB seasons since 1970
Combined record: 280-190-5 (.589)
Not only do mobile QB's under all three measurable definitions outperform the league average but the stricter the definition of mobile gets (and the smaller the sample), the better the win-loss percentage becomes. Progressively more mobile QB's appear to win progressively more games. Since 1970, QB's with 500yd rushing in a season, have won 58.9% of their games.
*************
Questions and Concerns
1. Why are QB seasons being examined and not individual games? What would the data look like if we looked at the outcome of individual games to see if QB's who ran for a specific yardage total (say 50yds in a game) were more likely to win the game?
Unfortunately, Pro Football Reference does not have that capability. Nevertheless, while such data would be interesting, it may be misleading for a variety of reasons. In any given game, QB's who are pocket passers could get 50yds but would not be accurately described as mobile. Likewise, clear running QB's have had games where they did not reach 50yds but the threat of their running affected the defensive gameplan. Using season rushing numbers avoids these problems and gives a clear picture of how a QB played over the course of an entire season.
2. How common are these definitions of mobile?
I included the number of qualifying QB seasons for each measure. 112 QB seasons qualified for 60+ rush attempts, 40 QB seasons qualified for 80+ rush attempts, and 33 QB seasons qualified for 500yd rushing. While 112 QB seasons may seem like a lot, it is important to remember that this number goes back to 1970. In 43 NFL seasons, only 112 QB seasons qualified. It is difficult to quantify how many QB seasons there have been since 1970, since it requires a definition of a qualifying season (such as a certain minimum of pass attempts), but I estimate that my mobile sample of 112 QB seasons is only about 10% of total QB seasons in this timeframe.
While 60+ rush attempts may seem like a low benchmark for a mobile QB (and it may arguably be, hence why this study included stricter measures), it is a comparatively rare feat for NFL QB's.
To help put this in perspective, 37 QB's threw at least 200 pass attempts in 2013. Even with the rising popularity of mobile QB's, only 9 QB's qualified to this lowest definition of 60+ carries in 2013.
3. How much does a .589 win percentage matter? That's not a large margin.
I've already seen this question from the less mathematically astute.
Imagine a set of the entire population of all NFL QB seasons since 1970. The average win-loss record of that entire sample will be .500. It has to be because football games are zero sum. For every winning QB in that entire population, there's a losing QB.
Now if I take out a portion of that population (mobile QB's with 60+ carries in a season), the remainder of that population are all the QB's with under 60 carries in a season.
Now remember that the average of the whole population is .500. If the average of the mobile QB's in the smaller sample is higher than .500 then necessarily, the remaining sample's average must be under .500. Otherwise, the two samples combined (mobile QB's and non-mobile QB's) would not produce an average of .500.
4.What about Brady/Manning/Montana/etc.? How many Super Bowls have mobile QB's won?
There are a lot more non-mobile QB's than mobile QB's. Given that mobile QB's are comparatively rare (my guess is under 10%), we can expect that close to 90% of the best QB's and 90% of SB winning QB's would be non-mobile, even if we accept that mobile QB's have some sort of regular season advantage, which is demonstrated by my research.
When you're talking Brady and others, you're talking outliers. My analysis is comparing mobile QB's across all talent levels (from Bobby Douglass and Tim Tebow to Steve McNair and Steve Young) and comparing it to a baseline average QB with a .500 win-loss %. Comparing to Hall of Famers would be meaningless for obvious reasons.
Of course I'd rather have Tom Brady than the AVERAGE mobile QB!
My research demonstrates it is better to have an average mobile QB than an average non-mobile QB. Mobile QB's win 56.8% more games than the average of the entire population of QB's since 1970.
As for Super Bowls, mobile QB's do not have many Super Bowl victories. In fact, Steve Young is the only SB winning mobile QB and his 1994 campaign did not include a 500yd rushing regular season. He rushed only 58 times for 293yds and 7TDs that year.
Is there a reason why mobile QB's outperform their non-mobile peers in the regular season but not in the playoffs/Super Bowl? I have no idea. Why would a set of QB's who have significantly higher regular season win percentages struggle in the playoffs? An individual QB might but it should average out over all of the QB's since 1970. In fact, 500yd rushing QB's are only 8-14 in the postseason with 4 Conference Championship appearances, all losses, despite their regular season track record.
To say that mobile QB's are all somehow not clutch or that defenses wait until the playoffs to figure them all out isn't meaningful. Such specific circumstances as 'choking' QB's or gimmicky QB's should average out over the entire population of mobile QB's over 43 years. Unless someone has a theory connecting mobile QB's and playoff losses in the face of their otherwise stellar win percentage, I will chalk the postseason record of mobile QB's to random chance.
In 2012, 7 QB's had 60+ carries. They combined for a 56-40 (.583) win-loss record.
In 2013, 9 QB's had 60+ carries. They combined for an exemplary 84-49 (.631) record.
Given the zero-sum nature of NFL games, any set of QB's who can post win percentages significantly higher than .500 likely bring significant competitive value to their teams. Is there a correlation between mobile QB's and win percentage? Is this simply a new trend with no historical basis?
Here, I will explore these possibilities by examining combined win-loss records of mobile QB's going back to the AFL-NFL merger of 1970. What I have discovered will shed light on the growing bias for mobile QB's among NFL coaches and general offices.
Definition: What is a mobile QB?
A mobile QB means different things to different people. Some define mobile QB's as run-first QB's with limited passing ability, such as Kordell Stewart, Tim Tebow, Tommie Frazier, and Pat White. Others define it as QB's with NFL-caliber passing ability but a tendency to run, such as Michael Vick and Randall Cunningham. Some exclude elite passers such as Steve Young, despite his 4200 career rushing yards, nearly as many as Cunningham. Others consider quick-footed but pass-first QB's like Aaron Rodgers as mobile while some don't.
For this study, I defined mobile QB in a number of ways. Each successive definition is more stringent and reduced the pool of QB's:
QB's with 60+ carries in a season. While this may seem low, only 9 QB's reached this measure in 2013 and all but one (Dalton) are assuredly mobile QB's (listed in order of carries):
Cam Newton, Russell Wilson, Colin Kaepernick, Robert Griffin III, Terrelle Pryor, Alex Smith, Geno Smith, Andrew Luck, Andy Dalton
QB's with 80+ carries in a season. This measure excludes Dalton, Luck, and both Geno and Alex Smith. This is a strict definition of mobile that tellingly excludes Geno Smith who is most definitely a mobile QB
QB's with 500+ rushing yards in a season. The most strict. By this measure, only Newton, Pryor, Wilson, and Kaepernick qualify as mobile during the 2013 season.
Methodology
My research looked at win-loss statistics for qualifying QB seasons using all three measurable definitions of mobile QB.
My sample included all QB seasons since the AFL-NFL merger in 1970. While many may believe mobile QB's are a recent phenomenon, the first qualifying QB using the strictest definition (500yd rushing season) was Detriot's Greg Landry who rushed for 530yds in 1971 in a 14-game season. The very next year, he ran for 524yds while Chicago's Bobby Douglass had a 968yd rushing season which stood as an NFL record for 34 years before being broken by Michael Vick in a 16 game season. The first mobile QB to have a winning season was New England's Steve Grogan in 1978 who rushed for 539yds on way to an 11-5 season.
Data was obtained from Pro Football Reference. Data sets are included below:
60+ rush attempt QB seasons since 1970 sorted by wins (112 QB seasons)
80+ rush attempt QB seasons since 1970 sorted by wins (40 QB seasons)
500+yd rushing QB seasons since 1970 sorted by wins (33 QB seasons)
Results
60+ rush attempt QB seasons since 1970
Combined record: 919-686-11 (.568)
80+ rush attempt QB seasons since 1970
Combined record: 349-253 (.579)
500yd rushing QB seasons since 1970
Combined record: 280-190-5 (.589)
Not only do mobile QB's under all three measurable definitions outperform the league average but the stricter the definition of mobile gets (and the smaller the sample), the better the win-loss percentage becomes. Progressively more mobile QB's appear to win progressively more games. Since 1970, QB's with 500yd rushing in a season, have won 58.9% of their games.
*************
Questions and Concerns
1. Why are QB seasons being examined and not individual games? What would the data look like if we looked at the outcome of individual games to see if QB's who ran for a specific yardage total (say 50yds in a game) were more likely to win the game?
Unfortunately, Pro Football Reference does not have that capability. Nevertheless, while such data would be interesting, it may be misleading for a variety of reasons. In any given game, QB's who are pocket passers could get 50yds but would not be accurately described as mobile. Likewise, clear running QB's have had games where they did not reach 50yds but the threat of their running affected the defensive gameplan. Using season rushing numbers avoids these problems and gives a clear picture of how a QB played over the course of an entire season.
2. How common are these definitions of mobile?
I included the number of qualifying QB seasons for each measure. 112 QB seasons qualified for 60+ rush attempts, 40 QB seasons qualified for 80+ rush attempts, and 33 QB seasons qualified for 500yd rushing. While 112 QB seasons may seem like a lot, it is important to remember that this number goes back to 1970. In 43 NFL seasons, only 112 QB seasons qualified. It is difficult to quantify how many QB seasons there have been since 1970, since it requires a definition of a qualifying season (such as a certain minimum of pass attempts), but I estimate that my mobile sample of 112 QB seasons is only about 10% of total QB seasons in this timeframe.
While 60+ rush attempts may seem like a low benchmark for a mobile QB (and it may arguably be, hence why this study included stricter measures), it is a comparatively rare feat for NFL QB's.
To help put this in perspective, 37 QB's threw at least 200 pass attempts in 2013. Even with the rising popularity of mobile QB's, only 9 QB's qualified to this lowest definition of 60+ carries in 2013.
3. How much does a .589 win percentage matter? That's not a large margin.
I've already seen this question from the less mathematically astute.
Imagine a set of the entire population of all NFL QB seasons since 1970. The average win-loss record of that entire sample will be .500. It has to be because football games are zero sum. For every winning QB in that entire population, there's a losing QB.
Now if I take out a portion of that population (mobile QB's with 60+ carries in a season), the remainder of that population are all the QB's with under 60 carries in a season.
Now remember that the average of the whole population is .500. If the average of the mobile QB's in the smaller sample is higher than .500 then necessarily, the remaining sample's average must be under .500. Otherwise, the two samples combined (mobile QB's and non-mobile QB's) would not produce an average of .500.
4.What about Brady/Manning/Montana/etc.? How many Super Bowls have mobile QB's won?
There are a lot more non-mobile QB's than mobile QB's. Given that mobile QB's are comparatively rare (my guess is under 10%), we can expect that close to 90% of the best QB's and 90% of SB winning QB's would be non-mobile, even if we accept that mobile QB's have some sort of regular season advantage, which is demonstrated by my research.
When you're talking Brady and others, you're talking outliers. My analysis is comparing mobile QB's across all talent levels (from Bobby Douglass and Tim Tebow to Steve McNair and Steve Young) and comparing it to a baseline average QB with a .500 win-loss %. Comparing to Hall of Famers would be meaningless for obvious reasons.
Of course I'd rather have Tom Brady than the AVERAGE mobile QB!
My research demonstrates it is better to have an average mobile QB than an average non-mobile QB. Mobile QB's win 56.8% more games than the average of the entire population of QB's since 1970.
As for Super Bowls, mobile QB's do not have many Super Bowl victories. In fact, Steve Young is the only SB winning mobile QB and his 1994 campaign did not include a 500yd rushing regular season. He rushed only 58 times for 293yds and 7TDs that year.
Is there a reason why mobile QB's outperform their non-mobile peers in the regular season but not in the playoffs/Super Bowl? I have no idea. Why would a set of QB's who have significantly higher regular season win percentages struggle in the playoffs? An individual QB might but it should average out over all of the QB's since 1970. In fact, 500yd rushing QB's are only 8-14 in the postseason with 4 Conference Championship appearances, all losses, despite their regular season track record.
To say that mobile QB's are all somehow not clutch or that defenses wait until the playoffs to figure them all out isn't meaningful. Such specific circumstances as 'choking' QB's or gimmicky QB's should average out over the entire population of mobile QB's over 43 years. Unless someone has a theory connecting mobile QB's and playoff losses in the face of their otherwise stellar win percentage, I will chalk the postseason record of mobile QB's to random chance.
Comment