Sponsored Content: Can Bishop Sankey Be A Breakout Rookie?

Heading into the 2014 NFL draft, there was not a ton of hype surrounding running back Bishop Sankey. Sure, he had a pretty good career at the University of Washington, but rookie running backs usually have to earn their position in the NFL. Thanks to a pretty good situation, Sankey might just be a breakout star in fantasy football in his 1st year.

When training camp started, many people thought that the starting running back position would come down to either Sankey or Shonn Greene. Now that Chris Johnson is no longer in town, Tennessee is looking for someone to be a reliable ball carrier.

Tennessee is not considered to be a playoff or Super Bowl threat by any means in 2014, but at the end of the day someone needs to put up numbers on offense. The Titans do not have a very successful passing game at this time, so Sankey has a chance to carry the ball quite a bit throughout the year. People in fantasy football like the opportunities he will receive, even if he might be a bit inconsistent at the beginning of the year.

Perhaps the biggest issue with Sankey at this point in time is his fumbling struggles. He had those issues at Washington, and he has already turned the ball over in the preseason this year. If Tennessee can’t trust him, they will quickly be pulling the plug on him as a starter.

Obviously, Sankey is not someone who is going to be taken extremely early on in fantasy football leagues. However, he is someone to keep an eye on in the middle rounds to add depth to the running back position. Things are really starting to open up for the rookie, and he might just make the most impact out of any rookie in fantasy football this year.

Pre Season Football Gambling Theory

I have always been of the opinion that pre-season football can be determined by two things; 1) A teams quarterback depth, and 2) the level of assholeness of the team’s coach.  In the case of the Hall of Fame Game, neither Doug Marone nor Tom Coughlin appear to be at the Belichick, Harbaugh, Harbaugh, or Tomlin level of asshole, and as such the game will likely come down to quarterback depth, wherein Buffalo is at a tremendous advantage.  The Bills are currently carrying four quarterbacks on their roster with significant playing experience.  This is crucial, as EJ Manuel and Eli Manning (my alleged look-a-like but I digress) are likely to see very limited playing time even by pre-season standards.

My pick is thus Buffalo minus 3, a good foreplay bet to start the season off.

Darren Rovell Doesn’t Understand Valuations

It’s easy to beat up on Darren Rovell, namely because everyone does it, but given that he’s ESPN’s “Sports Business” Reporter one could responsibly expect that he understand sports business. ESPN Article Namely companies of all sorts are valued on future earnings.  In essence, when you buy a company you are purchasing a stream of future cash flows, risk adjusted by the probably of realizing these earnings.  This is not dissimilar from stocks, wherein companies use a price to earnings multiple, earnings being future earnings not past earnings.  In other words, Steve Ballmer did not pay 12.1x revenues, but rather he offered either 6.2x projected earnings (assuming the $324.1 anticipated number) or 7.6x (assuming the $324.1 anticipated number).  Using these two numbers, Ballmer paid a price much more in line with precedent NBA transactions. Clipper Bid Book

 

Potential Not Polish

Image

A lot has been made about Teddy Bridgewater’s so called fall down the draft boards, and I hate to break it to you, but he was never valued very highly to begin with. Bridgewater played in what essentially was a mid-major, and he made his reputation by beating a talented Florida team in a bowl game Florida had no interest in partaking. He followed the bowl performance with an excellent statistical season. However, like many quarterbacks before him, Bridgewater is nothing but a college star with limited NFL potential. The NFL is a league that values exceptionalism, and quite frankly, Bridgewater doesn’t have an exceptional trait.

Bridgewater’s fall is not dissimilar to those of Geno Smith and Matt Barkley in 2013. Smith and Barkley, like Bridgewater were highly productive and polished college quarterbacks who lacked NFL potential. “Polish” can be taught and developed, however raw talent is much more limited. As such, EJ Manuel, a more raw quarterback, with significantly higher upside was valued much more highly than the polished Smith and Barkley.

Similarly, I anticipate raw quarterbacks such as Logan Thomas, Tom Savage, and Zach Mettenberger to be valued much more highly than the “polished” quarterbacks typically mocked in the same range. I would not be surprised, though I am not predicting, that one of the three aforementioned quarterbacks were to be selected before Teddy Bridgewater.

Capital in the 21st Century

Image

To preface this I haven’t yet had the time to read Piketty’s Communist Manisto in full detail, I have however been able to consume the most salient points of the book.  Most notably I think Piketty fails to realize how the confiscatory nature of social security hurts the poor and the middle class.  Over any extended period of time, the stock market in the capitalist world has generated a significant, positive total return.  However, with the passage of the social security act during the great depression, Franklin Roosevelt took away the ability for many lower income Americans to invest in the stock market for years to come.  This was of course an unforeseen externaliity, as when the act was passed only roughly 3 percent of all Americans owned stock.  However, as time has passed, the effect of this tax has been seen in full detail.  As time has passed, a significant, if hypothetical amount of the investment income of the poor or middle class has been forced into a zero return asset, versus the proven and substantial return that is the stock market.  Social security is notably capped so that only income to a certain level is subject to the payroll tax, and thus, they wealthy who already have a disproportionate amount of capital to invest in the stock market were given an additional advantage.  This advantage has over the last 80 years played out to its logical conclusion, and this further increased the wealth of those with the means to invest.

Two obvious solutions to this failure of government exist, and shockingly neither involves a confiscatory tax.  1) Convert social security into forced individual investment account or 2) convert the social security trust fund into a sovereign wealth fund.  Either solution does the obvious, it converts a substantial portion of the country’s wealth into a asset class that gains in value, a feature that has been lost in the public social security market since its inception.

Does NFL Combine Performance Predict to NFL Success?

Image

The 2012 Brigham Young University (BYU) football team was led by two pass rushers, Ezekiel Ansah and Kyle Van Noy.  Ansah, a Ghanaian student who had come to the university to run track was a freak athlete drooling with professional potential.  Van Noy was smaller but a much more productive football player.  Whereas Ansah recorded 4.5 sacks in his BYU career Van Noy recorded 4.5 sacks in the final two games of the 2012 BYU season to bring his season total to 13.  Following this final game, Ansah declared for the National Football League (NFL) Draft assuming that his athleticism would result in his being a high pick, and Van Noy choose to return to school for another year of football, as he lacked Ansah’s athleticism. In the last week of February the top NFL prospects, Ansah included were brought to the NFL Combine in Indianapolis for athletic, medical, and intelligence testing.  This data is used in conjunction with one’s performance on the field to determine their selection in the NFL Draft which occurs in February.  At the Combine, Ansah measured to be 6’5 271, ran forty yards in 4.63 seconds, bench pressed 225 pounds 21 times, recorded a vertical leap of 34.5 inches, a broad jump of 118 inches, ran the twenty yard shuttle in 4.26 seconds, and the three cone drill in 7.11 seconds, confirming what had previously been assumed that he is a freak athlete.  The slot a player is selected in the Draft largely determines the value of his rookie contract, the value of which can range from $22 million for the first pick, to $1.05 million for the last pick in the draft.  NFL teams award multimillion dollar contracts to players, based not only off of their performance as college athletes, but also based off of their testing at the combine which is thought to measure athleticism and potential.   The a forty yard dash is designed to measure “long speed,” the bench press measures upper body strength, the vertical jump jumping ability and lower body strength, the broad jump jumping ability and lower body explosiveness, the twenty yard shuttle change of direction ability and quickness, and the three cone drill again measures change of direction and quickness.  The NFL has created a readymade econometrics equation declaring NFL success to be a function of college performance plus athletic ability, health, and intelligence.  Ansah’s athletic ability was great enough to convince the Detroit Lions to use the fifth pick in the 2013 draft to select him, and he will sign a contract worth upwards of $18.5 million.  

Given the widespread intrigue of the NFL, data going back to 1999 measuring the performance of prospects at the Combine, sans medical and intelligence testing, is readily available on the internet.  This paper will focus on the Combine’s ability to predict NFL success for pass rushers, those who tested at the combine as defensive ends or outside linebackers, as sacks, out of all statistics compiled in football are the most individualistic, and attempt to measure the relationship between the combine performance of pass rushers and NFL success.  Success will be measured using two dependent variables, the number of sacks recorded per year in each player’s career, and the number of years played in the NFL.  By measuring sacks per year, one can quantify individual NFL success as a pass rusher, and the years variable will allow those players whose key skill is something other than rushing the quarterback, be it stopping the run, defending the pass, tackling, etc., to be more readily identified.  As NFL prospects have the ability to opt out of any drill at the combine, this study will only include those individuals who completed all drills between 1999 and 2012.  Those who choose to opt out typically do so for three reasons, one, they are considered an elite prospect and their performance will only drop them in the draft, two, they are injured and unable to perform the drill, and three, they simply are not good at the drill, and their performance at the drill will cause them to fall in the draft.  Furthermore, this study restricts players who failed to record a sack in their NFL career, assuming that they either played a role in the NFL that did not require them to rush the passer, or lacked the talent to make an NFL team.  

The NFL’s popularity has resulted in numerous journalists and bloggers undertaking significant studies examining the correlation between combine performance and NFL success, but rarely has this examination occurred in a scholarly manner.  Articles appearing in The Journal of Strength and Conditioning Research entitled “The NFL Combine: Does It Predict Performance in the National Football League?,” “The National Football League Combine: A Reliable Predictor of Draft Status?,” and “The National Football League (NFL) Combine: Does Normalized Data Better Predict Performance in the NFL Draft?,” have examined the relationship between combine success and NFL performance, but none of these studies have focused on pass rushers, and considered data over such a long ranging time period.  This paper differs from past research in two unique manners, one it focuses on only NFL skill, pass rushing, and two it uses a significantly larger dataset than prior studies, when controlling for position.    

Data Used

The data from used in this study is readily available on a wide variety of websites, but can be found most easily on NFLCombineResults.com[1] as it is already complied into spreadsheet form.  This data is compiled by the NFL scouts who administer the combine and then made readily available to the public.  In recent years, the Combine itself has been broadcast on the NFL Network.  The dependent variables used in the paper will be twofold, sacks per year, a readily available metric of pass rusher success, and years played in the NFL, a metric designed to measure NFL success more broadly.  With respect to dependent variables those used will include all those measured at the combine that are made readily available to the public: height, weight, the forty yard dash, bench press, twenty yard shuttle, three cone drill, vertical leap, broad jump, and explosive power, a popular sabermetric which claims to predict pass rusher success if the following formula is great than 1.05: (vertical leap (in) +3.5*broad jump (in))*((weight/height[ft.])/3000).

To capture NFL success two metrics were used, both taken from http://www.pro-football-reference.com, sacks, and years played in the NFL.  A sack is when a defensive player tackles the pass thrower (almost always the quarterback) behind the line of scrimmage.  These two data points were then assigned to the prospect combine date, and used to create a third metrics, sacks per year.  As many of the players in this data set are still active NFL players, their career sack totals are still increasing and thus sacks alone would not accurately express NFL success for many of the recent combine performers.  By measuring sacks per year, an average, success can be measured on a per year basis and thus NFL success for all players, not thus those whose careers have concluded can be measured.  This dataset was limited to 227 defensive ends and outside linebackers who recorded at least half a sack between 1999 and 2012 and participated fully in the NFL Combine.  Career sacks in this data set range from ½ to 117, with a mean of 40.472 and a median of 33.  In short, this paper will attempt to measure a relationship between combine performance and NFL performance.  Given that the average player of the 227 measured in the data set played five years in the NFL with a mean number of sacks of 14.6 with a median number of 8, the average is large enough to capture NFL success over a typical season for a typical player, though this number may be skewed for the players who have entered the league more recently as they are likely to still be developing as players, and thus are unlikely to have reached their full potential as players.  In addition to sacks per year, years in the NFL were used as a measure of performance.  This measure assumes that NFL coaches and personal men understand NFL success better than statisticians, and thus it is a measure of NFL success that can be used to explain the longevity of careers for those who did record a high number of sacks.  Thus, years in the NFL can be used to explain a correlation between combine performance and overall NFL success, rather than success specifically as a pass rusher for which there exits far better data.    

The table below shows the descriptive statistics used for each variable measured in this paper, sacks both on a per year and per year basis, years played in the NFL, height in both feet and inches, weight, the forty yard dash time, number of repetitions record on the 225 pound bench press, the twenty yard shuttle time, three cone drill time, vertical leap, broad jump in both feet and inches, and explosive power calculation.

Descriptive Statistics

 

Mean

Median

Standard Deviation

Minimum

Maximum

Sacks (Career)

14.621

8.00

18.095

0.50

117.00

Sacks Per Year

2.7272

1.80

2.7694

0.100

16.750

Years in NFL

5.3612

5.00

2.8036

1

14

Height (inches)

75.251

75.00

1.6249

70.00

80.00

Height (feet)

6.2709

6.25

0.13541

5.8333

6.6667

Weight (pounds)

257.77

258

18.545

221

304

40 Time (seconds)

4.7418

4.73

0.14708

4.41

5.20

Bench Press (repetitions)

23.379

23.00

4.6798

11

36

20 Yard Shuttle (seconds)

4.3441

4.35

0.17904

3.830

5.04

Three Cone Drill (seconds)

7.2279

7.2100

0.25387

6.690

7.880

Vertical Leap (inches)

34.154

34.00

3.4324

25.00

45.50

Broad Jump (inches)

115.85

115.00

6.1080

102.00

134.00

Broad Jump (feet)

9.6538

9.5833

0.50900

8.50

11.167

Explosive Power

0.92880

0.92179

0.065538

0.74714

1.1277

 

The limitations of this date set are obvious, NFL prospects are not required to participate in the combine, and often those who are considered to be the best prospects, typically the best athletes choose not to participate, feeling that their performance can hinder but not help their draft stock.  These prospects, typically the biggest, fastest, strongest, and quickest are likely to be the most successful in the NFL.  Secondly, prospects have the ability to opt out of specific drills, and for the sake of ease those players who did not fully participate in the combine were removed from the data set.  Thirdly, this data does not take into account football ability, rather it is a measure of pure athleticism, the data operates under the assumption that an invitation to the NFL Combines means that an individual possess the requisite ability to succeed as a professional football player; this is not always the case.  Further not all of these players were drafted into roles that fit their skillset, and thus their lack of NFL success is not entirely their own fault.  Finally, not all of these players were drafted into pass rushing roles.  Rather the play in a defensive scheme where they are asked to tackle and cover rather than blitz the quarterback.  

 

Empirical Results

Using two different dependent variables for performance, both sacks per year, and years in the NFL, I attempted to measure whether combine performance predicted NFL success.  When first regressing using sacks in the NFL as the dependent variable it was found that bench press reps and vertical leap were not significant at any level, and were thus removed from the equation.  Given the collinearity that exists between vertical leap, broad jump, and explosive power by the inherent nature of the explosive power calculation, explosive power was not included in this regression.  First, the model was then narrowed down again and a series of T and F tests were performed.  Given a null hypothesis that each of these variables had no effect on an individual’s NFL performance a series of two sided T Tests were conducted, with six regressors. Thus (N-K-1) was equal to 220.  All regressors accounted for were found to be significant at the 10 percent levels, but forty yard dash was found to be insignificant at the 5 percent level, and the twenty yard shuttle was soundly rejected at the 2 percent level.  The remaining regressors (height in inches, weight, three cone drill time, and broad jump in feet) were found to be significant at the 1 percent level.

 

R-squared            0.264119   Adjusted R-squared   0.244050

Given that the null hypothesis could not be rejected at the two percent level for both the forty yard dash and the twenty yard shuttle, the model was again revised, and forty yard dash and twenty yard shuttle were removed.  The model now equaled the followed SacksPerYear = -30.4259 + 0.0482784 ((Weight_Lbs) – 2.76023 (ThreeCone) + 1.51352 (BroadJump_feet) + 4.15375 (Height-feet).

 

Moreover, feet replaced inches as the measure of height, for the sake of simplicity regarding the regressed equation.  T tests were again conducted, assuming a two trailed null hypothesis that these variables had no impact on sacks per year, and all four variables were found to be significant at the 1 percent level, thus the null hypothesis was rejected, and F tests were conducted to measure the effect of these variables on sacks per year.  It was found after performing 14 F tests that the null hypothesis could be soundly rejected in all 14 cases, thus height, weight, broad jump, and three cone drill all soundly effect sacks per year. 

F Value

Restrictors

Reject / Fail to Reject Null

7.862152

Height

Reject at 1 %

16.27738

Broad Jump

Reject at 1 %

13.7174

Three Cone

Reject at 1 %

15.68722

Weight

Reject at 1 %

18.63666

Height, Broad Jump, Three Cone

Reject at 1 %

14.62851

Weight, Broad Jump, Three Cone

Reject at 1 %

19.67792

Height, Weight, Broad Jump

Reject at 1 %

18.04975

Weight, Height, Three Cone

Reject at 1 %

25.72575

Height, Weight

Reject at 1 %

14.38257

Height, Broad Jump

Reject at 1 %

10.37237

Height, Three Cone

Reject at 1 %

11.93594

Weight, Three Cone

Reject at 1 %

11.97756

Weight, Broad Jump

Reject at 1 %

21.5293

Three Cone, Broad Jump

Reject at 1 %

   

In addition to T and F Tests, White’s test for heteroskedacity was conducted, and the data in the model was found to be soundly heteroskadistic, with a P value of 0.010100, and thus robust standard errors have been used for all data in this paper. 

Next, Ramsey’s Reset Test for omitted variables was conducted.  With a null that the model is well specified, the resulting P value was 0.00446 and thus the null was rejected and the model determined to not be well specified, meaning the model is lacking in variables.  Unfortunately, given the data collected at the combine itself, no amount of squares, and added variables could create a well specified model.  The regression performed at the combine is correlated to NFL performance, but it does not fully explain NFL performance.  In the case of this paper, the unexplained variable, U, is too largely.  Namely, this data set does not account for football ability, as collegiate sack statistics have only recently become official. 

Following the Ramsey’s Test confirmation that combine performance along does not predict NFL success amongst pass rusher, I next tested whether or not sabermetrics (sports analytics) when applied to the combine can better predict NFL performance.  The popular metric, explosive power claims to predict ones pass rushing ability using an equation based on combine performance.  Should ones explosive power be greater than or equal to 1.05, success as a pass rusher is predicted.  Whereas success is predicted, it is not defined as a number, and thus for the purpose of this paper a total of 5 or more sacks per year shall be considered a success.  Before regressing, the correlation between explosive power and sacks per year was measured.  Given an unrestricted sample, a positive correlation of 0.26939814 between sacks per year and explosive power was found, this correlation decreased significantly to 0.06952094 when the sample was restricted to those who averaged 5 or more sacks per year. 

Given this small, but positive correlation, Explosive Power replaced broad jump and vertical leap in the regression model.  Given that the first model was found to be heteroskedastic, White’s Test was immediately performed and this data, with a P value of 0.002131 was found to be heteroskedasitc.  Thus robust stand errors were applied.  Next, T tests were conducted; again assuming a two tailed null hypothesis that explosive power, three cone drill time, and height in feet played no role in ones NFL performance.  This hypothesis was rejected at the 1 percent level.

 

Next six F tests were conducted, and with one and two restrictions, at the 1 percent level, the null hypothesize claiming that explosive power, the three cone drill, and height have no effect on sacks per year were soundly reject.

F Values

Restrictors

Reject / Fail to Reject

27.79085

Height

Reject at 1 %

15.91423

Three Cone

Reject at 1%

12.70688

Explosive Power

Reject at 1%

17.70293

Height, Three Cone

Reject at 1%

15.86291

Three Cone, Explosive Power

Reject at 1%

23.09064

Explosive Power, Height

Reject at 1%

 

Given that this data was consistently found to be significant at the 1 percent level, a Ramsey’s Reset Test was then performed.  With a P value of 0.137 the model was found to be well specified.

 

Thus the following equation was determined to be an accurate predictor of sacks per year for NFL prospects tested at the combine: SacksPerYear = -28.3865 + 9.14075(ExplosivePower) – 2.70408 (ThreeCone) + 6.72447 (Height_Feet).

In addition to testing the relationship between success (sacks) and combine performance, I also tested the relationship between career longevity and combine performance.  Using the dependent variable years in NFL, I regressed the observed data from the NFL combine, and the explosive power metric.  Initial T testing found the forty yard dash, bench press, and height to be insignificant at any level.  They were thus removed from the model.  The recalculated model found broad jump to only be significant at the 10 percent level, whereas the remaining dependent variables, three cone drill, vertical leap, shuttle, weight and explosive power were all significant at the 1 percent level.  Broad jump was thus removed as a variable, and the final model read as follows: YearsinNFL = 2.53133 + 2.72657(ThreeCone) + 0.05558435(weight) + 0.365367(VerticalLeap) – 6.10958(Shuttle) – 18.5302 (ExplosivePower).  However when conducting a Ramses Reset Test with this model it was found to not be well specified.  Thus broad jump was added back into the final model, which with a P Value of 0.0655 is well specified.  The final model thus reads as follows: YearsInNFL = -10.6250 + 2.88837(ThreeCone) + 0.0916198(Weight_lbs) + 0.477216(VerticalLeap_in) – 6.21480(shuttle) – 29.9054(ExplosivePower) + 0.0860357(BroadJump_in).  T tests conducted found that the three cone drill, weight, vertical leap, twenty yard shuttle, and explosive power variables are all significant at the 1 percent level.  Broad jump however is a weak measure, only significant at the 20 percent level.  White’s Test was conducted, and the model was found to be homoscedastic, with a P value of 0.061255.

 

A series of F tests were conducted to measure the joint significance of the independent variables regressed.  With the exception of broad jump which could not be rejected at the 1% significance level all we can reject the null hypothesis that the three cone drill time, weight, vertical leap, twenty yard shuttle, explosive power, and broad jump have no effect on salary.

F Values

Restrictors

Reject / Fail to Reject

2.60655

Broad Jump

Reject at 2%, Fail to Reject at 1%

7.829734

Explosive Power

Reject at 1%

20.66821

Shuttle

Reject at 1%

8.532994

Vertical Leap

Reject at 1%

7.577941

Weight

Reject at 1%

9.781593

Three Cone

Reject at 1%

3.961732

Broad Jump and Explosive Power

Reject at 1%

11.31309

Broad Jump and Shuttle

Reject at 1%

4.352487

Broad Jump and Vertical

Reject at 1%

3.831788

Broad Jump and Weight

Reject at 1%

5.71335

Broad Jump and 3 Cone

Reject at 1%

13.22994

Explosive Power and Shuttle

Reject at 1%

4.334664

Explosive Power and Vertical

Reject at 1%

3.980407

Explosive Power and Weight

Reject at 1%

9.043681

Explosive Power and Three Cone

Reject at 1%

14.72068

Shuttle and Vertical

Reject at 1%

12.43742

Shuttle and Weight

Reject at 1%

10.96884

Shuttle and Three Cone

Reject at 1%

4.384796

Vertical and Weight

Reject at 1%

9.443739

Vertical and Three Cone

Reject at 1%

9.373583

Weight and Three Cone

Reject at 1%

It can thus be concluded that years played in the NFL can be predicted by the following equation.  YearsInNFL = -10.6250 + 2.88837(ThreeCone) + 0.0916198(weight_lbs) + 0.477216(VerticalLeap_in) – 6.21480(Shuttle) – 29.9054(ExplosivePower) + 0.0860357(BroadJump_in).

Conclusion

Regressions preformed in this paper have confirmed the commonly held but never proven belief regarding the combine.  It is a predictor of NFL performance, but it is only a partial predictor of NFL performance.  A pass rushers’ average sacks can be explained by their weight, three cone drill time, broad jump distance and height. The unexplained variable U, is too large to allowed combine performance to solely explain NFL performance.  Ezekiel Ansah’s freakish athleticism does not guarantee that he will become a successful pass rusher, but it increases the probability that he will do so.  The obvious correlation between college performance and NFL performance is unaccounted for, and this paper set out only to explain the connection between combine performance and NFL success.  College football players like Kyle Van Noy who were highly productive players in college but lacked Ezekiel Ansah’s freakish athleticism are not guaranteed to fail in the NFL.  The model generated using just variables observed at the combine, thus is not well specified enough to accurately predict NFL performance.  Given the failure of observed data to predict NFL performance, the regression model added the popular sabremetric explosive power.  Sabremetrics, (sports analytics) use data to predict athletic results.  When explosive power was added to the model it alone could not predict NFL success, but in conjunction with other data measured at the combine, a model combining explosive power, three cone drill time, and height was found to be a sufficient and significant predictor of average sacks per year.  Success was also measured via career longevity, as a means to account for those in the data set who despite working out at pass rushing positions at the combine, are not in fact truly pass rushers. Again the model passed solely on combine data was found to be insufficient.  It was determined that years in the NFL could be predicted via three cone drill time, weight, vertical leap height, twenty yard shuttle time, explosive power, and broad jump distance. 

The implications of this paper are twofold, one that the combine does measure relevant data regarding pass rushers, and two that this data alone is insufficient to predict NFL among pass rushers.  When combined with analytical data, in this case explosive power, this combine data becomes a far better, but still imperfect predictor of both NFL performance and longevity amongst outside linebackers and defensive ends.  Future research in this subject should consider variables that accurately explain NFL success and regress these variables against combine performance at other positions.  In addition these studies should attempt to lower the unexplained data U, by measuring collegiate success, and controlling for level of competition.  Given the importance of the NFL Draft in building an NFL team, an NFL team with a complete model would be at a significant strategic advantage.  When evaluating pass rushers, NFL teams should focus on specific attributes: height, explosive power, and the three cone drill,  and ignore irrelevant data such as forty time and bench press repetitions.  This does not mean that on field ability can be ignored, rather that certain variables can when collegiate performance is accounted for better predict NFL success.  More broadly given the success of the sabermetric explosive power, it is clear that NFL teams should invest in analytics, the way basketball and football teams have to better build their roster.

The conclusion drawn in this paper is simple.  Certain variables measure at the NFL combine can be used to predict the NFL success of pass rushers.  However any equation created solely using these variables is insufficient as the unaccounted for variables, namely on field performance are too great.  This U can be lowered when using the sabermetric variable explosive power, and can be further lowered when collegiate performance is observed.  When drafting and thus investing in players teams should value combine performance, sabermetric data, and collegiate performance.  The combine itself cannot predict NFL success on its own, but it is a crucial part of the draft process.

The Unpredictable Beauty of House of Cards

 

Image

Most TV shows and movies are simple.  The lay out in the preview, or in the prior episodes what is going to happen, and then anyone with half a brain can predict what is going to happen.  What makes House of Cards exceptional is the manner in which the plot pivots.  I told a friend the easiest way not to spoil House of Cards season two was pretty simple: tell someone exactly what happens in the first episode, they won’t believe you.  Simply put most TV shows do not have primary character A push primary character B in front of a movie train, without making primary character A suffer the consequences.

The conclusion of season one of House of Cards created the belief amongst viewers that season two would be dedicated to the staff of Slugline taking down Frank Underwood.  Within two episodes of season two, this belief was shatter.  The show pivoted.  No longer was it about the intersection of of politics and media, rather it became a show about the dynamics of power.  In many respects, House of Cards transformed itself from the final season of The Wire, to a modernized Game of Thrones, set in Washington D.C.

Like Game of Thrones, House of Cards has no qualms about killing or eliminating beloved of primary characters, and it does so rapidly, without the audience expectations.  What makes House of Cards great is that it is unpredictable, it doesn’t cater to the expectations or or desires of the audience, but at the same time it is not so tragic that the viewer via reverse psychology can know what to expect.

The Seahawks Discovered the Formula to Beating Peyton Manning

Image

Lots has been made about Seattle’s performance in the Super Bowl.  Quite frankly they were dominant.  However, It seems to me that most of the media has failed to pick up on the simple brilliance of why they were able to dominate.  Most teams approach Peyton Manning by diversifying their defense, by showing him as many looks as possible.  Quite frankly this is a terrible approach.  Manning isn’t an elite athlete, he doesn’t have an elite arm, what he has is an elite brain.  He is able to adapt to, and defeat exotic looks that defenses show, better than said defenses are able to defend from these uncomfortable, and often new positions.  The simple brilliance of Seattle’s gameplan was that they showed Denver very few looks.  In short, Seattle played their cover three man shell the vast majority of the game.  This defense is extraordinarily fundamentally sound, especially with the Seahawks personal.  As such, Seattle forced Denver to try and beat them on skill rather than scheme, and the Broncos were outmatched, and thus defeated handily.

In Time For The Second Quarter of Texas A&M Alabama: Turnover Adjusted Efficiencies Week 1

Below you will see my week one power rankings, the methodology is described in my earlier post, but in essence in quantifies how a team plays on a game by game basis by adjusting total yards for turnovers.  The Dan Williams pick six caused a massive outlier in the Arizona-St. Louis game.  So Ignore Arizona and St. Louis.  The team of the week was Denver, and my Ravens had the worst performance of any team.  It will require roughly six weeks, for this to provide a worthwhile sample size, but I still find weekly rankings interesting.

Yards Required to Score Adjusted Yards Gained Turnover Adjusted Offensive Efficiency Turnover Adjusted Defensive Efficiency Team Rating
Arizona 890.00 370.00 40.50 131.31 85.90
Denver 894.00 578.00 78.66 72.35 75.51
Houston 803.00 329.40 42.85 102.94 72.90
Green Bay 1041.00 351.20 36.47 105.34 70.90
Kansas City 831.00 218.70 35.77 105.51 70.64
Dallas 795.00 226.50 22.57 102.80 62.69
Chicago 690.00 316.00 41.97 82.92 62.45
Detroit 844.00 396.50 44.57 76.71 60.64
Miami 765.00 260.00 32.61 88.16 60.39
San Francisco 811.00 446.00 55.31 63.53 59.42
Tennessee 654.00 175.60 33.54 81.55 57.55
Philadelphia 807.00 246.60 40.42 73.40 56.91
New Orleans 747.00 295.40 42.23 66.00 54.11
New England 955.00 241.40 29.26 78.70 53.98
New York Jets 978.00 301.90 34.92 71.79 53.36
Oakland 572.00 337.00 58.97 45.36 52.16
Seattle 671.00 290.00 44.10 59.01 51.56
Carolina 555.00 230.80 40.99 55.90 48.44
Indianapolis 558.00 304.00 54.64 41.03 47.84
Tampa Bay 843.00 163.70 28.21 67.37 47.79
Buffalo 996.00 195.00 22.94 70.74 46.84
Atlanta 822.00 257.40 34.00 57.77 45.89
Pittsburgh 815.00 161.20 20.29 69.51 44.90
Washington 1044.00 221.00 26.60 59.58 43.09
Cincinnati 790.00 195.00 17.08 63.83 40.46
Minnesota 1015.00 208.60 23.99 55.43 39.71
Cleveland 894.00 228.00 11.84 67.39 39.61
New York Giants 881.00 190.50 -2.80 80.90 39.05
Jacksonville 1218.00 41.00 3.93 64.69 34.31
San Diego 719.00 203.20 -2.94 57.15 27.10
Baltimore 1307.00 335.40 26.47 21.34 23.90
St. Louis 692.00 318.00 -34.15 59.50 12.67
Average 840.53 269.78 30.81 71.86             51.33

Turnover Adjusted Efficiency Week One: Ravens and Broncos Posted

bm

As I described in one of my earliest post, I am experimenting with a power rankings system that measures all that happens on the football field in terms of yardage, an explanation of which can be found here Turnover Adjusted Offensive Efficiency.  Below you will find the Turnover Adjusted Offensive and Defensive Efficiency Ratings for the Ravens in Broncos, following Thursday Night’s game.  Given the inherent nature of a blowout and the small sample size, the number are skewed, but it should give a rough idea as to how the ratings will play out.

 

Team Yards Required to Score Adjusted Yards Gained Turnover Adjusted Offensive Efficiency Turnover Adjusted Defensive Efficiency Team Rating
Denver 894 578 78.66215047 72.35376467 51.01591514
Baltimore 1307 335.4 26.46522062 21.33784953 -52.19692985

Georgia, Alabama, LSU, & Stanford Are All Run Heavy Teams. Coincidence, I Think Not.

Image

Whereas the vast majority of college football teams have moved toward spread offenses, three of the best teams over the past few years have adapt power running.  By my estimation roughly 100 teams in Division I football now run a spread offense.  While the number of teams running some variation of the spread offense has increased, the number of high school prospects best suited to run the spread offense likely has not increased.  The talented pool for spread offense has thus been diluted.  Conversely, the dominant SEC schools of Georgia, LSU, and Alabama, in addition to the Stanford Cardinals have either stuck with or adopted traditional power running based schemes.  Whereas the number of players suited to play in a power running offense has remained constant, the number of teams running such an offense has decreased significantly.  These teams, many of whom already have the pick of the litter talent wise,  are competing against fewer teams which has further consolidated this talent.  

Similarly this trend occurred in the early 2000s in the NFL, wherein three of the best defense were the Baltimore Ravens, New England Patriots, and the Pittsburgh Steelers.  It was not a coincidence that these teams also happened to be the only three running 34 defenses.  The 34 defense is not inherently better than the 43 defense, rather the defense succeed because while 29 teams were selecting and signing defensive personnel and the basis of their ability to function in a 43 defense, 3 teams were selecting and signing defensive personnel on the basis of their ability to perform in a 34 defense.  The 34 defensive talent was the consolidated into significantly fewer teams, whereas the 43 talent was diluted across the league.  Eventually the other teams in the league recognized this trend, and in an attempt to capitalize on this talent dilution, they adopted 34 defenses, so that in 2013 half the league will run 34 defenses, and I anticipate 43 defenses to thus outperform their 34 counterparts in the near future. 

I foresee a similar trend occurring in college football.  Given the dislocation of power running resources, teams will once again turn towards these schemes, because it will best enable them to acquire talent.

My Dog Heavy Week One Picks

Ravens + 7.5, Rationale: John Harbaugh has yet to lose a game with multiple weeks to prepare.

Bills +10.5, Rationale: The Patriots have a lot of public money on them and they are breaking in a new quarterback.

Titans +7, Rationale: The Steelers are an aging defense and lack a running back.  The Titans have the talent to make things interesting week 1.

Saints -3: Rationale: The Falcons struggle on the road and New Orleans will be much improved after a lost season.

Bucs – 4.5 Rationale:  The Bucs are loaded with talent, if everything clicks this is a blowout, plus Geno Smith just isn’t very good.

Jaguars +4.5, Rationale: Kansas City might be the most overrated team coming off a 2-14 season in history.  Jacksonville will be competitive especially if Henne starts.

Cincinnati +3, Rationale:  The Bengals tend to start fast, and Chicago still lacks an adequate line to block Cincinnati’s pass rush.

Miami +1, Rationale: I don’t trust Cleveland.  They still lack talent on offense.

Carolina +4.5, Rationale: Seattle’s struggles on the East Coast are enough to keep the game to a field goal.

Lions -6, Rationale: Minnesota just isn’t a good football team.

Colts – 10, Rationale: Have you ever seen Terrelle Pryor try and throw a football?

Cardinals +4.5, Rationale: Patrick Petersen finally provides the complement to Larry Fitzgerald that Arizona has missed since Boldin, and Tyrann Mathieu is a stud.

Green Bay +5, Rationale: San Francisco will struggle on offense without Crabtree, and Frank Gore is a year older.

Dallas -3.5, Rationale: The Giants lack depth at key positions.

Eagles +4.5, Rationale: Chip Kelly’s offense will catch Washington off guard, and Mike Shanahan will have a conservative game plan as RG3 returns from injury.

Texans -5.5, Rationale: San Diego is really bad.

 

Just For Fun: I Picked Every NFL Game of the 2013 NFL Season

Though it is a largely futile, below you will find my NFL season picks.  I have chosen a winner for every game in the regular season, and bracketed and predicted the playoffs as well.  Have some fun with it.  A few of my picks do not conform to group think.

Away Home Winner
Baltimore Denver Ravens
Sunday, Sept. 8
New England Buffalo Patriots
Tennessee Pittsburgh Steelers
Atlanta New Orleans Saints
Tampa Bay Jets Bucs
Kansas City Jacksonville Chiefs
Seattle Carolina Seahawks
Cincinnati Chicago Bengals
Miami Cleveland Dolphins
Minnesota Detroit Lions
Oakland Indianapolis Colts
Green Bay San Francisco Packers
Arizona St. Louis Rams
New York Giants Dallas Cowboys
Monday, Sept. 9
Philadelphia Washington Eagles
Houston San Diego Texans
Week 2
Thursday, Sept. 12
New York Jets New England Patriots
Sunday, Sept. 15
St. Louis Atlanta Falcons
San Diego Phildelphia Eagles
Dallas Kansas City Cowboys
Miami Indianapolis Colts
Tennessee Houston Texans
Washington  Gren Bay Packers
Cleveland Baltimore Ravens
Carolina Buffalo Panthers
Minnesota Chicago Bears
New Orleans Tampa Bay Saints
Detroit Arizona Cardinals
Jacksonville Oakland Raiders
Denver New York Giants Broncos
San Francisco Seattle Seahawks
Monday, Sept. 16
Pittsburgh Cincinnati Bengals
Week 3
Thursday, Sept. 19
Kansas City Philadelphia Eagles
Sunday, Sept. 22
Houston Baltimore Ravens
New York Giants Carolina Panthers
Detroit Washington Redskins
San Diego Tennessee Titans
Arizona New Orleans Saints
Tampa Bay New Orleans Patriots
Green Bay Cincinnati Bengals
St. Louis Dallas Cowboys
Cleveland Minnesota Browns
Atlanta Miami Dolphins
Buffalo New York Jets Bills
Indianapolis San Francisco 49ers
Jacksonville Seattle Seahawks
Chicago Pittsburgh Steelers
Monday, Sept. 23
Oakland Denver Broncos
Week 4
Thursday, Sept. 26
San Francisco St. Louis Rams
Sunday, Sept. 29
Pittsburgh Minnesota Steelers
New York Giants Kansas City Chiefs
Indianapolis Jacksonville Colts
Baltimore Buffalo Ravens
Seattle Houston Texans
Cincinnati Cleveland - Bengals
Chicago Detroit Lions
New York Jets Tennessee Titans
Washington Oakland Redskins
Philadelphia Denver Broncos
Tampa Bay Arizona Cardinals
Dallas San Diego Cowboys
New England Atlanta Falcons
Monday, Sept. 30
Miami New Orleans Saints
Week 5
Thursday, Oct. 3
Buffalo Cleveland Browns
Sunday, Oct. 6
New England Cincinnati Patriots
Detroit Green Bay Packers
Seattle Indianapolis Seahawks
Baltimore Miami Ravens
New Orleans Chicago Bears
Philadelphia New York Giants Eagles
Kansas City Tennessee Titans
Jacksonville St. Louis Rams
Carolina Arizona Cardinals
Denver Dallas Broncos
San Diego Oakland Raiders
Houston San Francisco 49ers
Monday, Oct. 7
New York Jets Atlanta Falcons
Week 6
Thursday, Oct. 10
New York Giants Chicago Bears
Sunday, Oct. 13
Cincinnati Buffalo Bengals
Detroit Cleveland Browns
Oakland Kansas City Chiefs
Carolina Minnesota Panthers
Philadelphia Tampa Bay Bucs
Green Bay Baltimore Packers
St. Louis Houston Texans
Pittsburgh New York Jets Steelers
Jacksonville Denver Broncos
Tennessee Seattle Seahawks
New Orleans New England Patriots
Arizona San Francisco 49ers
Washington Dallas Redskins
Monday, Oct. 14
Indianapolis San Diego Colts
Week 7
Thursday, Oct. 17
Seattle Arizona Seahawks
Sunday, Oct. 20
Tampa Bay Atlanta Falcons
Cincinnati Detroit Lions
New England New York Jets Patriots
Houston Kansas City Texans
Buffalo Miami Dolphins
Dallas Philadelphia Eagles
Chicago Washington Redskins
St. Louis Carolina Panthers
San Diego Jacksonville Jaguars
San Francisco Tennessee 49ers
Baltimore Pittsburgh Steelers
Cleveland Green Bay Packers
Denver Indianapolis Colts
Monday, Oct. 21
Minnesota New York Giants Giants
Week 8
Thursday, Oct. 24
Carolina Tampa bay Bucs
Sunday, Oct. 27
Dallas Detroit Lions
Cleveland Kansas City Chiefs
Miami New England Patriots
Buffalo New Orleans Bills
San Francisco Jacksonville 49ers
New York Giants Philadelphia Eagles
Pittsburgh Oakland Raiders
New York Jets Cincinnati Bengals
Washington Denver Redskins
Atlanta Arizona Cardinals
Green Bay Minnesota Packers
Monday, Oct. 28
Seattle St. Louis Seahawks
Week 9
Thursday, Oct. 31
Cincinnati Miami Dolphins
Sunday, Nov. 3
Kansas City Buffalo Chiefs
San Diego Washington Redskins
Atlanta Carolina Panhers
Minnesota Dallas Cowboys
Tennessee St. Louis Rams
New Orleans New York Jets Saints
Tampa Bay Seattle Seahawks
Philadelphia Oakland Eagles
Pittsbrgh New England Patriots
Baltimore Cleveland Ravens
Indianapolis Houston Texans
Monday, Nov. 4
Chicago Green Bay Packers
Week 10
Thursday, Nov. 7
Washington Minnesota Redskins
Sunday, Nov. 10
Seattle Atlanta Falcons
Detroit Chicago Bears
Philadelphia Green Bay Packers
Jacksonville Tennessee Titans
St. Louis Indianapolis Colts
Oakland New York Giants Giants
Buffalo Pittsburgh Bills
Cincinnati Baltimore Ravens
Carolina San Francisco 49ers
Denver San Diego Broncos
Houston Arizona Cardinals
Dallas New Orleans Saints
Monday, Nov. 11
Miami Tampa Bay Bucs
Week 11
Thursday, Nov. 14
Indianapolis Tennessee Titans
Sunday, Nov. 17
New York Jets Buffalo Bills
Baltimore Chicago Ravens
Cleveland Cincinnati Bengals
Atlanta Tampa Bay Bucs
San Diego Diego at Miami Dolphins
Arizona Jacksonville Jaguars
Oakland Houston Texans
Washington Philadelphia Eagles
Detroit Pittsburgh Steelers
Kansas City Denver Broncos
Minnesota Seattle Seahawks
San Francisco New Orleans Saints
Green Bay New York Giants Giants
Monday, Nov. 18
New England Carolina Panthers
Week 12
Thursday, Nov. 21
New Orleans Atlanta Falcons
Sunday, Nov. 24
Pittsburgh Cleveland Browns
Tampa Bay Detroit Lions
Minnesota Green Bay Packers
San Diego Kansas City Chiefs
Chicago St. Louis Rams
Carolina Miami Panthers
New York Jets Baltimore Ravens
Jacksonville Houston Texans
Indianapolis Arizona Cardinals
Tennessee Oakland Raiders
Dallas New York Giants Giants
Denver New England Patriots
Monday, Nov. 25
San Francisco Washington Redskins
Week 13
Thursday, Nov. 28
Green Bay Detroit Lions
Oakland Dallas Cowboys
Pittsburgh Baltimore Ravens
Sunday, Dec. 1
Tampa Bay Carolina Panthers
Jacksonville Cleveland Browns
Tennessee Indianapolis Colts
Denver Kansas City Chiefs
Chicago Minnesota Vikings
Miami New York Jets Dolphins
Arizona Philadelphia Eagles
Atlanta Buffalo Atlanta
St. Louis San Francisco Rams
New England Houston Texans
Cincinnati San Diego Chargers
New York Giants Washington Redskins
Monday, Dec. 2
New Orleans Seattle Seahawks
Week 14
Thursday, Dec. 5
Houston Jacksonville Texans
Sunday, Dec. 8
Indianapolis Cincinnati Bengals
Buffalo Tampa Bay Bucs
Kansas City Kansas City Redskins
Minnesota Baltimore Ravens
Cleveland New England Browns
Carolina New Orleans Panthers
Oakland New York Jets Jets
Detroit Philadelphia Eagles
Miami Pittsburgh Steelers
Tennessee Denver Broncos
New York Giants San Diego Chargers
Seattle San Francisco 49ers
St. Louis Arizona Cardinals
Atlanta Green Bay Packers
Monday, Dec. 9
Dallas Chicago Bears
Week 15
Thursday, Dec. 12
San Diego Denver Broncos
Sunday, Dec. 15
Washington Atlanta Falcons
Chicago Cleveland Browns
Arizona Tennessee Titans
Houston Indianapolis Colts
New Orleans St. Louis Rams
New England Miami Patriots
Philadelphia Minnesota Vikings
Seattle New York Giants Giants
Buffalo Jacksonville Bills
San Francisco Tampa Bay Bucs
New York Jets Carolina Panthers
Kansas City Oakland Raiders
Green Bay Dallas Packers
Cincinnati Pittsburgh Bengals
Monday, Dec. 16
Baltimore Detroit Ravens
Week 16
Sunday, Dec. 22
Miami Buffalo Bills
Minnesota Cincinnati Bengals
Indianapolis Kansas City Chiefs
Tampa Bay St. Louis Rams
Cleveland New York Jets Browns
Chicago Philadelphia Eagles
Dallas Washington Redskins
New Orleans Carolina Panthers
Tennessee Jacksonville Jaguars
Arizona Seattle Seahawks
Denver Houston Texans
New York Giants Detroit Giants
Oakland San Diego Chargers
Pittsburgh Green Bay Packers
New England Baltimore Patriots
Monday, Dec. 23
Atlanta San Francisco 49ers
Week 17
Sunday, Dec. 29
Carolina Atlanta Falcons
Green Bay Chicago Bears
Houston Tennessee Titans
Cleveland Pittsburgh Steelers
Washington New York Giants Giants
Baltimore Cincinnati Ravens
Philadelphia Dallas Cowboys
Jacksonville Indianapolis Colts
New York Jets Miami Dolphins
Detroit Minnesota Vikings
Buffalo New England Bills
Tampa Bay New Orleans Bucs
Denver Oakland Broncos
San Francisco Arizona Cardinals
Kansas City San Diego Chargers
St. Louis Seattle Rams
 AFC Wins Losses    NFC Wins Losses
Patriots 11 5 Eagles 11 5
Bills 7 9 Redskins 11 5
Dolphins 7 9 Giants 7 9
Jets 1 15 Cowboys 7 9
Ravens 13 3 Packers 12 4
Bengals 10 6 Lions 6 10
Steelers 8 8 Bears 6 10
Browns 8 8 Vikings 3 13
Texans 11 5 Panthers 11 5
Colts 9 7 Falcons 9 7
Titans 7 9 Bucs 8 8
Jaguars 3 13 Saints 7 9
Broncos 10 6 Seahawks 11 5
Chiefs 7 9 49ers 8 8
Raiders 5 11 Rams 8 8
Chargers 4 12 Cardinals 8 8
Playoff Seed AFC NFC
1 Ravens Packers
2 Texans Seahawks
3 Patriots Eagles
4 Broncos Panthers
5 Bengals Redskins
6 Colts Falcons
Wild Card
Away Home Winner   Away Home Winner
Colts Patriots Patriots Falcons Eagles Eagles
Bengals Broncos Bengals Redskins Panthers Redskins
Divisional
Away Home Winner   Away Home Winner
Bengals Ravens Ravens Redskins Packers Packers
Patriots Texans Patriots Eagles Seahawks Seahawks
Conference
Away Home Winner   Away Home Winner
Seahawks Packers Seahawks Patriots Ravens Ravens
Super Bowl
NFC AFC Champion
Seahawks Ravens Ravens

2014 NFL Draft Order

Jets 1
Jaguars 2
Vikings 3
Chargers 4
Raiders 5
Lions 6
Bears 7
Bills 8
Dolphins 9
Titans 10
Chiefs 11
Giamts 12
Cowboys 13
Saints 14
Steelers 15
Browns 16
Bucs 17
49ers 18
Rams 19
Cardinals 20
Colts 21
Falcons 22
Broncos 23
Panthers 24
Bengals 25
Redskins 26
Eagles 27
Texans 28
Patriots 29
Packers 30
Seahawks 31
Ravens 32

Mortgaging the Future: My Brief Theory on the Jets (And Every Other Team That Has Overachieved Under a Rookie Coach)

Image

The Jets Rex Ryan inherited were not a particularly talented football team.  I think in short that Rex Ryan through his bravado as a young coach, (and his defensive schemes) was able to convince the Jets that they were better than they were.  The team overachieved, they produced over capacity in economic terms.  In essence they were a 6-10 or 7-9 team that went 9-7.  This in turn lowered the draft picks awarded to the Jets, and diminished their future talent, as they did not get the influx of top tier talent that is required for a formerly bad team to achieve sustained success.  Eventually Rex’s message of confidence wore off, and the veteran free agent stop gaps the Jets had signed early in his tenure aged, and their skills diminished.  The combination forced the Jets to return to their natural capacity, which owing to their lack of early first round draft picks typical of a bad team, was far worse than that of a team that experience gradual improvement under a young head coach.  The Jets over-achievement in Ryan’s first two years cost the franchise the essential building blocks of sustainable development, and thus when the team was forced to fall back on its homegrown talent, their true, and not particularly talent self was revealed.

Some Thoughts On Heads & Knees

Following DJ Swearinger’s hit on Dustin Keller, there has been a lot of discussion regarding whether or not the NFL should protect players knees in addition to their head.  In essence this is a discussion of life versus livelihood.  While some suspect that concussions can cause long term neurological issues, Prevalence and Characterization of Mild Cognitive Impairment in Retired National Football League PlayersChronic Traumatic Encephalopathy in Sport: a Systematic Review, the link is by no means scientifically proven.  Years of qualitative evidence will however suggests that tearing a ligament in one’s knee will adversely effect ones ability to play football.  Though medical technology has  advanced significantly, a torn ACL is still a 6-12 month injury, with no guarantee of a return to past form.  A tear of ones knee, though by no means life threatening, is a direct threat to ones livelihood.  Given that professional athletes have a limited earning period, it should by no means be surprising that the vast majority of athletes are more concerned with the health of their knee than the health of their head, and would rather suffer a concussion than a debilitating knee injury.  Swearinger’s hit was by no means cheap or dirty, but regardless, the hit to Keller’s knee will significantly alter his earning trajectory, confirming why a shredded knee is the worst fear of many players.  Given the Player’s Union’s inherent mandate to ensure maximum revenue for its members, it is in the best interest of the Union to advocate for rules that protect the knees.  This protection must however be reconciled with the fact that football is an inherently violent game, and this violence is part of what makes the game popular.  Football fans do not want to see the sport adopt a lacrosseesque strike zone.  Every business sells a product, and the NFL is no different, it sells football to fans, who have chosen to consume NFL football from a wide variety of entertainment options.  Simply put it is in the best interest of the players to limit their exposure to violence, particularly violence that limits the longevity of their careers, however this violence can only be limited to a point before it drives away fans.  This violence should thus not be limited via legislation (which will drive away fans), but rather by investment in new technology designed to protect players from injury.

The Pseudo Science of Pre-Season Gambling

I’d call myself a gambling addict but I hit winners at too good of a rate.  To me it’s a thrill similar to options trading, except it does require you to be eyeing the market 24/7.  Gambling is fun, and a great way to supplement your income.  The obvious issue is that most consider the preseason hard to predict.  This is simply not true.  The issue is that most bet preseason games like regular season games and bet accordingly.  The key to the preseason is depth, particularly depth at the backup quarterback position.  Though it is more an art than a science, but if you can accurately evaluate which team has a better backup quarterback the spreads are close enough that you can hit at a winning clip.  If you follow me on Twitter, @WallStFootball, you will know I had an extraordinarily good week, (7-0-1,) and I used the aforementioned method.  It’s not complicated, and only applies to the preseason, but it’s an excellent way to make some easy money.

week 2 1

week 2 1.5

week 2 2

Towards a Better Boxscore Part II: Tackles, Not as Clear Cut As You May Think

To the casual observer tackles might be the simplest statistic used to evaluate defensive players.  Simply put, tackles are a means used to quantify a defensive players ability to stop an offensive player.  Good defenders should be able to tackle well, and thus those who led the league in tackles are good defenders.  Why is it then that the league leader in tackles tends to be the middle linebacker on a bad team?  To begin with, the NFL is not an Oklahoma Drill wherein the only variables are a blocker, a runner, and a defender.  Rather football is an 11 on 11 wherein four or five defensive players, the linebackers, and the strong safety specialize in tackling.  While the defensive line specializes in occupying blockers and rushing the passer, and the secondary specializes in coverage.  With these positions, being a skilled tackler is an added benefit but by no means a prerequisite.

Now that I have stated the obvious, the League’s leading tacklers almost always play positions that specialize in tackling, the focus of the article will now turn to the less obvious.  Why are the leading tacklers almost always on bad defenses.  The reason is obvious when you think about opportunity.  Defense is fundamentally about preventing an offense from scoring, and this is best done by returning the ball to your team’s offense as quickly as possible, either via turnovers, or short possessions.  The best defenses spend the least amount of time on the field, as they force the most three and outs, and the most turnovers.  Bad defenses struggle to get off the field, and are most prone to giving up long drives.  These bad defenses thus place their tackling specialists in the opportunity to make the most tackles.  The fundamental flaws with tackles as a statistic are twofold, one tackles do not consider the number of plays run against a defense, and two tackles do not consider where on the field a player is tackled.  Take Jerod Mayo in 2010 for example, he lead the league in tackles, but the Patriots finished 25th in total defense.  The casual observer may assume that Mayo offered stellar linebacker play on a terrible defense but the truth is much less romantic, Mayo plays a linebacker which inherently put him in a position to make a lot of tackles, and given the nature of the Patriots defense, he had the opportunity to make a lot of tackles, the combination of which, not his particular skill allowed him to be the league leader in tackles.

To combat the skewing of the tackling statistic, I propose that tackle be viewed as a percentage, (tackles made / plays on the field)*100.  This simple equation would allow the best tacklers to emerge, be it someone who makes 10 tackles on 60 plays on the field (a tackle percentage of 16.66), or 12 tackles on 48 plays on the field (a tackle percentage of 25).  The numerator in this equation considers individual performance, tackling, and the denominator considers team performance, plays run against.  Though this statistic will still “punish” players who play on defenses with multiple strong tacklers, it will no longer reward those tackling specialists who play on bad defenses.

The Pistol Is The Future & The Future is Now

Though it is an alignment not an offensive scheme, I have been convinced since I first saw Colin Kaepernick run his offense at Nevada out of the pistol that it is the formation of the future.  Traditional under center alignment’s limit the quarterbacks down field vision and release time as he is required to drop back, and traditional shotgun formations limit the running back, as he cannot accelerate downhill until the quarterback has handed off the ball.  In essence, under center formations hinder the quarterback and shotgun formations hinder the running back.  The Pistol in its most basic nature marries the benefits of under center and shotgun formations, while divorcing most of the downside.  By aligning the quarterback 4-5 yards behind the center and the running back 3-4 yards behind him, Chris Ault at Nevada was able to devise an offense wherein his quarterback could easily see the field and release the ball, and his running backs could accelerate prior to taking the hand off.

 

In the NFL, the Pistol has largely been married to a zone read scheme, or what NFL types have taken to call the “Read Option.”  This marriage was necessary in order to bring the scheme into the league, but it is by no means a prerequisite for running the Pistol.  In fact the first adapter was the Kansas City Chiefs under quarterback Tyler Thigpen, who while functionally mobile, would never be confused with a running quarterback.  The Pistol inherently lends itself to a play action scheme combined with a short timing and rhythm passing game.  Last season the Washington Redskins ran a variant of this scheme, wherein they combined a zone blocking running scheme, with timing and rhythm passing, frequent deep balls off of play action, and Robert Griffin III’s running ability, which was employed as a means of keeping the defense honest.  As is always true in football, a mobile quarterback benefits this scheme, but is not a requisite to succeed in the pistol.  Griffin’s running ability was the cheese on top of the nachos.  A team lacking a mobile quarterback can frequently use wide receiver screens to a similar effect.

Given the benefits of the Pistol in the run and passing games, and the fact that one does not require a mobile quarterback to run such an offense, I thoroughly expect the majority of the league to use this formation in the upcoming season.  The Pistol is not a fluke offense like the Wildcat that limits ones scheme, rather it allows for an expansion of the playbook.  The Pistol is the future and the future is now.

Top Five Advanced Statistics To Know For Fantasy Football

ImageTop Five Advanced Statistics To Know For Fantasy Football 

Another season of fantasy football is quickly approaching, and many people are trying to do as much preparation as possible. There are thousands of preview articles to read online, but sometimes the best way to prepare is to simply look at the numbers (or more importantly, the right numbers). Here are five football advanced statistics every person should know and understand to improve their chances of success in Fantasy Football 2013

Total Quarterback Rating 

As the name suggests, this is a composite number that attempts to judge the value of every single signal caller in the NFL. The standard quarterback rating has been around for a while, but that focused only on passing statistics. TQR considers passing, running and situational statistics to see who is the best. While this might not come in handy if you already prefer one of the elite quarterbacks, it could help you land a solid backup instead of just guessing. 

 Fantasy Points Against 

Instead of simply looking at how good a defense is, most fantasy owners want to go a bit deeper than that. This statistic breaks it down by position, meaning that a person could see how a defense handles the average running back, quarterback, wide receiver or tight end. 

 Specific Red Zone Statistics 

 Whether you are looking at touches, catches, targets or goal-to-go plays, red zone statistics are essential when doing research. Scoring a touchdown or two can really help a person rack up the fantasy football points in a hurry. Some players excel down in that area, while others struggle. 

Win Probability Added 

If you are a fan of the Wins Above Replacement (WAR) stat in basketball, this is the NFL answer to it. The goal of the stat is to quantify the value of every single player on the field. Obviously, quarterbacks are generally going to have the highest value since they always have the ball in their hands, but it is a nice tool to compare players who are lining up at the same position to get their overall value. 

Elusive Rating 

This might be one of the newest statistics, and it is really only useful for those trying to figure out their running game. Pro Football Focus developed the statistic, which tries to illustrate which runner is the hardest to bring down. The formula is [(missed tackles forced/(carries+receptions)]*[(yards after contact per attempt)*(100)].

 

Warning Baseball Post: Misaligned Incentives & The Calculus of Steroids

I am not a baseball guy.  Quite frankly, the game bores the hell out of me.  I grew up playing lacrosse, and I find baseball to be dull, slow, and lacking athleticism.  Despite this, I have always been intrigued by the statistical aspects of baseball, as in many regards in is closer to a numbers game like poker than to other professional sports.  Given that every play is determined by only two participants, who typically have track records long enough that the law of large numbers will apply, baseball is far easier to model over the long run than any other sport.  Further this dependence on statistics has made players individual performance the golden cow of baseball.  Though Joe Morgan can preach timely hitting all he wants, deep down in this post post Moneyball era, we all know that’s horseshit.  The ease at which baseball statistics can be understood (outside of wins and losses there isn’t a skewed stat I can think of off the top of my head) has directly translated into player compensation, it may sound obvious that players are paid on performance, but in true team sports there are other factors to be considered.  Thus any increase in performance will result in an increase in salary.  Further, there is no salary cap in professional baseball, teams must only spend less than they take in, and thus, even the best players can be compensated for playing at an even higher level.  Further, baseball contracts, unlike those in football, are fully guaranteed, meaning that if caught using performance enhancing drugs, a player only risks the money he stands to make over the period of the suspension.

The basic calculus of steroid use for an established baseball player is as follows, if the increase in salary a baseball player stands to make by using steroids is greater than the amount of money he would stand to lose over the period of time he is suspended multiplied by the chance of not getting caught, then he should take steroids.  Mathematically if S > LP where in S = Salary Gained by Using Steroids, L = Salary Lost if Suspended, and P = the Probability of Not Being Suspended.

A similar calculus exists for non established baseball players who are attempting to make the major leagues.  If the increase in ones expected baseball salary, i.e. the expected gain in salary brought about by making a level of professional baseball not attainable sans performance enhancing drugs, multiplied by the odds of not getting caught minus the loss in salary over the period he is suspended multiplied by the probability of not getting caught is greater than the individuals expected income given his intelligence and socio-economic background than he should take steroids.  Mathematically if SP-LP > I wherein S is the Expected Gain In Salary, P is the Probability of Not Getting Caught, L is the Salary Lost While Suspended, and I is the Expected Income if the Individual did not play baseball.

This basic calculus jives well with the Bio-Genesis suspensions that have recently occurred.  The vast majority of those suspended have been Latin American, with the exception of Alex Rodriguez and Ryan Braun.  The World Bank Ranks every country by Per Capita GDP, the closest thing we can easily find to substitute for expected income if not playing baseball.  American GDP per Capita is equal to $49,965, Mexican GDP per Capita is equal to $16,676, Venezuelan GDP per Capita is equal to $13,475, and the GDP per Capita in the Dominican Republic is $10,204.  By plugging these GDP per Capita figures into I for the later equation, one can see why it is more common for Latin American baseball players to be caught using steroids, their break even point is far lower than their American counter parts.  It isn’t that these Latin American baseball players play by different rules than their American counterparts, it’s that the gains required for steroid use to be a worthwhile investment are far lower for Latin American baseball players.

Comparative Advantage in Football

Image

One of the most basic topics in economics is comparative advantage.  This post will explore its application on the football field.  In economics comparative advantage is typically used as an argument for free trade.  “It is an economic theory first advanced by Robert Torrens and David Ricardo that analyzes international trade in terms of differences in relative opportunity cost. The theory suggests that countries should specialize in the goods they can produce most efficiently rather than trying for self-sufficiency and argues strongly in favor of free international trade.” Comparative advantage – Definition  In essence say the United States, produce ten units of corn per hour worked and two units of rice for each hour worked, and China can produce nine units of rice per hour worked and three units of corn per each hour worked, then assuming no tariffs, transportation costs, etc, it is most efficient for the United States to stop its production of rice, and for China to stop its production of corn, and for the United States to thus trade its surplus corn to China in exchange for rice.  Similarly this concept can be applied to football, most notably at positions that come in pairs, safety, wide receiver, and defensive end.  Take last years Baltimore Ravens for example.  At safety they started Ed Reed, and Bernard Pollard, and at wide receiver they started Torrey Smith and Anquan Boldin.

By 2012 injuries had robbed Ed Reed of his ability to tackle, though he maintained his elite coverage ability.  Conversely Bernard Pollard is in essence a glorified outside linebacker, standing 6’0 230 pounds, Pollard is a tremendous hitter, and had more tackles than any other Raven, however, Pollard’s coverage ability left something to be desired.  Baltimore thus unknowingly (I assumed) deployed its comparative advantage during the playoff run.  In economic terms, on a scale of 1-5, 5 being best, Ed Reed could produce one unit of run support , and five units of pass coverage, whereas Bernard Pollard could produce two units of pass coverage and five units of run support. Ed Reed played deep, specializing in pass coverage, and Pollard was routinely found in the box, specializing in run defense.  Together, they formed one of the top safety combinations in the NFL, as this deployment of comparative advantage was able to largely mask the glaring weaknesses in their games.  However, when Baltimore realized that they would be unable to afford to resign Ed Reed, they were forced to cut Bernard Pollard, because it would be near impossible to find another safety who could so well mask Pollard’s weakness in coverage.

Similarly, the Ravens starting wide receivers were Torrey Smith, and Anquan Boldin.  Smith is one of the fastest receivers in the NFL, and Boldin, who can a 4.71 forty yard dash at the combine was likely the slowest starting receiver in the league.  Again the Ravens deployed economic theory to their advantage.  Smith specialized in deep routes, whereas Boldin specialized in routes that required body positioning, short routes, patterns over the middle, and end zone throws.  Though neither receiver exclusively ran these patterns, this specialization allowed the Ravens to thrive.

Image

Comparative advantage is in simple terms about allow people to do what they do best.  Good coaches recognize players strengths, and weakness, and deploy their players in such a manner.  Bad coaches fail to do this, and force their players to play outside of their skill set.  Weaknesses can be masked in football, if players are allowed to specialize at a specific niche in which they excel.  Another example of this would be running back by committee.  A team like the Saints this past season had an ideal pass catcher in Darren Sproles, a power runner in Chris Ivory, a balanced back in Pierre Thomas, and a pounder in Mark Ingram.  Though individually none of these backs are anything special, together, they formed one of the best combinations in the league.  To conclude most teams lack players that thrive in all aspects of the game, and by specializing and deploying combinations, teams can mask their players weaknesses and form elite units without elite talent.

2014 Draft Pre Season Top Five By Position

Below you will find the first of my 2014 NFL draft coverage. Unlike most of what I do, the draft will not have an analytic basis, rather I will rely on my knowledge of college football, and my ability to determine which skills best translate to the NFL. I am more than prepare to take my licks in the comment section. The one thing I will say is that non-conformity is not inherently wrong, and that this Top 5 ranking will evolve as the season progresses.  I intend to add individual write ups that will correspond to 2013 game tape.

Quarterbacks
1) Johnny Manziel
2) Teddy Bridgewater
3) Brett Hundley
4) Kevin Hogan
5) AJ McCarron

Dislike Tajh Boyd

Running Backs
1) Lache Seastrunk
2) De’Anthony Thomas
3) Dri Archer
4) Charles Sims
5) James Wilder, Jr.

Dislike Ka’Deem Carey


FB
1) J.C. Copeland
2) Jalston Fowler

Wide Receiver
1) Sammy Watkins
2) Marqise Lee
3) Donte Moncrief
4) Mike Evans
5) Sean Price

HM
Malcolm Mitchell, Kelvin Benjamin, Tevin ReeseTy Montgomery

(Not high on Jordan Matthews, Cody Hoffman, or Allen Robinson. I question their speed.)

Tight End

1) Austin Seferian-Jenkins
2) Colt Lyerla
3) C.J. Fiedorowicz
4) Arthur Lynch
5) Eric Ebron

Offensive Tackle
1) Cyrus Kouandjio
2) Jake Matthews
3) Taylor Lewan
4) Cornelius Lucas
5) Antonio Richardson

Interior Offensive Line
1) Gabe Jackson
2) David Yankey
3) Xavier Su’a-Filo
4) Bryan Stork
5) Cyril Richardson

34 Defensive End (Five Technique)
1) Stephon Tuitt
2) Ra’Shede Hageman
3) Ed Stinson
4) Garrison Smith
5) Taylor Hart

43 Defensive End
1) Jadeveon Clowney
2) Aaron Lynch
3) Leonard Williams (Not actually eligible but we can all dream of an unrestricted labor market).
4) Frank Clark
5) James Gayle

43 Defensive Tackle (Three Technique)
1) Will Sutton
2) Timmy Jernigan
3) Deandre Coleman
4) Dominique Easley
5) Travis Raciti

Nose Tackle (One or Zero Technique)
1) Louis Nix
2) Anthony Johnson
3) Viliami Moala
4) Daniel McCullers
5) Beau Allen

34 Outside Linebacker
1) Anthony Barr
2) Adrian Hubbard
3) Carl Bradford
4) Kyle Van Noy
5) Morgan Breslin

HM
Trent Murphy

43 Outside Linebacker
1) Christian Jones
2) C.J. Mosley
3) Ryan Shazier
4) Justin Jackson
5) Shayne Skov

Middle Linebacker
1) Yawin Smallwood
2) Trey DePriest
3) Chris Borland
4) A.J. Johnson
5) Andrew Jackson

Cornerback
1) Loucheiz Purifoy
2) Aaron Colvin
3) Lamarcus Joyner
4) Bradley Roby
5) Jason Verrett

HM
Andre Hal

Strong Safety
1) Karlos Williams
2) Chris Young
3) Dion Bailey
4) Vinnie Sunseri
5) Jordan Richards

Free Safety
1) Ha Ha Clinton-Dix
2) Tre Boston
3) Craig Loston
4) Ed Reynolds
5) Ty Zimmerman

Fantasy Advice From a Different Perspective

e5a53__0930_oag_panthersEvery sports publication worth its salt publishes a fantasy football issues, and I can neither compete with nor do I expect to compete with these massive sites.  I will thus take a different approach in my fantasy advice pieces.  As any employee at JP Morgan will tell you, significant arbitrage opportunities exist if you are will to sell your soul to the devil, and the same applies regarding fantasy football.  Assuming your league is entirely local, the easiest way to succeed is to draft players from your rival team, as they will like clockwork be undervalued, as few are willing to make the sacrifice of rooting for a rival, or alums of a rival college, to succeed at fantasy football.  Say for example your league primarily made up of Ravens fans, I can guarantee you that like clockwork Steelers and Patriots players will fall in the draft.  Outside of star NFL players, the majority of your fantasy football league will be unwilling to draft members of their hometown team’s rival.  This provides an opportunity that can be easily exploited.  Take Patriots running backs Steven Ridley and Shane Vereen like a 3rd and 5th round draft choices respectively in a 12 team league.  In a Baltimore based league, it is highly possible that players such as these will fall at least a round or two, at which time, you in your quest for the fantasy championship, can grab them up.  Employing such a strategy throughout the draft will give you several of what Mel Kiper would term value picks.  Players deserving to go a round or two higher, because of the misaligned incentives of fantasy football, and place you on the fast track to a championship.  My second piece of advice is to draft a mobile quarterback.  Every 10 rushing yards a quarterback records is equal to 30 passing yards.  Many fail to notice these added points.  Cam Newton and Joe Flacco had very similar pure passing statistics, which would in theory equate to a very similar fantasy performance, with Flacco doing slightly better.  See below.

Cam Newton, QB CAR 280 485 57.7 3,869 7.98 82 19 12 36 86.2 242
Joe Flacco, QB BAL 317 531 59.7 3,817 7.19 61 22 10 35 87.7 239

However, the above statistics only consider rushing yards, not passing yards, when rushing yards are factored in a Quarterback like Newton performs at an elite fantasy level.

2012 16 127 741 5.8 72 8 49 4 1
2012 16 32 22 0.7 16 3 11 3 1

Running quarterbacks such as Cam Newton, Colin Karpernick, Robert Griffin III, and Russell Wilson, can thus perform at a fantasy level comparable to Tom Brady, Peyton Manning, and Drew Brees, despite not yet being on their level as a passer, and thus comparable value can be found a round or two after the run on the top tier quarterbacks is completed.

Note: Aaron Rogers is an above average runner in addition to being an elite passer, and should always be drafted with the first pick in the draft.

The Salary Cap Should Be Calculated After Taxes

Though the following link is slightly dated, I know California for one has raised their highest rate, State Income Tax Rates – Highest Tax Rates for Each State, it gives a fairly accurate listing of the highest personal income tax rate in every state.  As you can see from looking through the list, the states of Florida, Texas, Washington, and Tennessee with regards to attracting potential free agents, as they in sports with salary caps, Football, Basketball, and Hockey, can in effect offer significantly more income.  Tennessee mitigates its advantage by its bizarre jock tax, see this Grantland article for details Paying to Play in Memphis? The Strange Case of Tennessee’s Jock Tax – The Triangle Blog, but Florida, Texas, and Washington, have no personal income tax whatsoever.  In this regard, the Seahawks, Texans, Cowboys, Dolphins, Jaguars, and Buccaneers are at an inherent advantage when it comes to attracting potential free agents.  As Forbes demonstrated in its breaking down of the Dwight Howard saga, Could State Taxes Cause Dwight Howard To Flee L.A. For Houston? teams in these markets can offer players less money before taxes, and the athletes will still come out ahead.  If one assumes that athletes are one rational individuals, and two, smart enough to max a free agency decision on the basis of post tax earnings, one has to assume that we will see a flood of professional athletes to income tax free states.  As a proponent of the free market, I love the fact that the governors of these states, such as Rick Perry, have placed their states in a position to thrive, by lowering their tax rates.  However, sports are not a traditional business, an an oligopoly is generally bad for professional professional sports, as it results in a decrease in fan interest.  Thus, to increase competition in the professional athletics free agent market I propose that the salary cap be calculated as net income rather than as income paid.  In this regard, teams in states governed by idiots such as California would not be placed at an inherent advantage.  Such a proposal comes with the obvious limitation that owners have budgets, and increasing the money available to spend in theory does not necessarily increase the money available to spend in actuality.  However, it eliminates the competitive disadvantage facing those owners willing to spend money.  No longer will we see the narrative of players such as Lebron James “taking less money” to sign with the Miami Heat.

Considering A NFL Summer League

brady

Given the gym etiquette dictates that only sports or news may be watched on the gym’s television, and that with its slow pace of play, baseball has never appealed to me,  I found myself flipping between a Major League Lacrosse game, and the NBA Summer League while on the bike.  Never had I more desired a football game to be on television than at that very moment, or at least substantive analysis regarding football.  I realized that baseball, a game whose fan base is rapidly ageing, has been given roughly two full months without an real or consistent competition on the athletic calender.  Furthermore, Baseball has positioned itself as a leisurely activity, a pastime played by athletes rather than a fast paced team sport.  Thus from mid June through the end of August, the American sports calendar lacks a televised fast paced team sport.    This is a niche that the National Football League could and in my opinion should exploit.  Though there are a plethora of success stories, (Tom Brady, Arian Forster, etc) the league by and large struggles to develop its late round draft picks, as these players are rarely exposed to live game action, a necessity in player development.  The need for a developmental league has already been recognized in various past forms, NFL Europe, the UFL, and the XFL.  These three leagues failed, as, lacking a built in fan base, they were unable to attract fans.  The NFL possess an inherent advantage should they chose to launch an NFL branded summer league, built in fan bases.  My proposal is as follows: the NFL should create a branded NFL Summer League, wherein the bottom 50 players on each teams roster are afforded the chance to develop, essentially rosters would be made up of practice squaders, special teamers, and camp bodies.  Akin to the NBA, each NFL team would have their own summer league team, which would allow the league to leverage its existing fan base to place butts in seats and draw eyeballs to the television.  Such a league would run from the start of June through the second week of July, allowing these players to transition immediately from the summer league to training camp.  As Malcolm Gladwell argues in his work Outliers  it takes 10,000 hours to master a task, and thus giving these fringe players an extra six weeks of games and practice, they will be afforded a greater opportunity to gain the necessary experience to succeed in the NFL.  The NFL would be afforded the opportunity to sell an additional six weeks of live programming to its broadcast partners.  Given the nature of the competition, the broadcast rights for the NFL Summer League would obviously be less than those of the regular season games, or conversely the NFL could televise it in house on the NFL Network which is perpetually in need of live programming.  Similarly, each team would be able to sell three more games of (discounted) tickets, again raising revenues.

Under the current collective bargaining agreement (CBA) such a league would be disallowed, as the agreement strictly limits the amount of practice time, games, and contact that a player may be exposed to.  The CBA would thus need to be amended to allow the creation of such a summer league.  Given that such a league would bring six additional paychecks to fringe players, many of whom are essentially unemployed, and hoping for a chance to stick in the NFL, I believe such a change would be passable, so long as sufficient protections are put into place, namely limitations on which players would be eligible for the summer league.

Now that we have effectively established the NFL Summer League, only one question is left.  Would you watch?

Night Time at Death Valley

The Closest Thing To Legal Sports Betting

mag_odog_bands_239__MG_3505

As we all know it is effectively illegal to bet on sports, either with a bookie or via the internet.  This illegality keeps the most straight laced of us (those society has deemed pussies) away from gambling.  Regardless of one’s political position, we can all agree that gambling is fun; it may be risky, and questionably moral, but it is nonetheless quite entertaining.  Sports betting is particularly fun as it gives the gambler a rooting interest in the game.  There is one site that allows the individual to obtain such a rooting interest and bet on the game without placing a “bet.”  Stubhub.  Ticket prices are a result of numerous factors, the economy, general fan interest, weather, game experience, the fan base of the away team, but the most important factor of all is team performance.  In essence, the better a team performs over the course of the season, the more the tickets to a team’s game will cost.  Conversely the worse a team performs the lower the ticket price will be.  However shorting a ticket is a far more complicated transaction, involving selling the ticket soon, and repurchasing it at a later date for profit.  Take for example two college football teams I am extremely bullish on, Texas A&M and LSU.  Tickets for the Texas A&M – Alabama game are currently running for a minimum of $424 in the nosebleeds.  The game is at Kyle Field.  Nosebleeds in Death Valley are currently running as low as $279 for the Texas A&M LSU game.  Buying a ticket for such a game is essentially a bet on LSU and Texas A&M, the margins are lower than a traditional bet, but higher than almost any other possible investment.  If LSU and Texas A&M, both play well, which I expect they will do I expect that the ticket prices will appreciate to near Texas A&M Alabama values.  Obviously this requires a much higher capital investment than what a casual gambler may possess, but for those with means, I would recommend investing in the following games, in addition to Texas A&m LSU: Baylor – Texas, South Carolina – Florida, USC – Stanford, Ole Miss – Texas A&M, and Northwestern – Michigan.  In the NFL, I would recommend: Bears – Eagles, Saints – Panthers (In Carolina), Packers – Steelers, 49ers – Falcons, and Patriots – Ravens.  Obviously, this is a risker, and less lucrative bet than true sports betting, but attractive returns can nonetheless be found.

Towards a Better Box Score Conclusion: Turnover Adjusted Offensive Efficiency

romo2

In the three previous posts, I have discussed my philosophy on how offense should be evaluated.  This post will synthesize my past discussions and introduce what I have deemed Turnover Adjusted Offensive Efficiency (TAOE).  TAOE is a calculation of all that can happen on a football field presented in the currency of yardage.  Simply put TAEO adds penalty yards and turnovers to the box score to create what I feel is the most accurate measure of a team’s offensive performance.  It also allows for the calculation of Turnover Adjusted Defensive Efficiency (TADE) which is calculated by taking 100-TAOE.  TAOE like yards can be calculated game by game, the game totals can be summed to create a season long state.  Again, we start with Yards Required to Score, (the distance from the end zone that the offense starts with the ball) and yards gained plus penalty yards gained or lost, essentially the endpoint of any given drive.  This later calculation is divided by the former to give the raw offensive efficiency calculation on a given drive, and for the aggregate game itself.  Next turnovers are considered.  Using my previously mentioned turnover to yardage equivalency calculation (-Net Punting Average – Yards Beyond Original Line of Scrimmage Returned) (If the return does not pass the original line of scrimmage, the calculation is simply (-Net Punting Average + Distance to Original Line of Scrimmage) the yards gained plus penalty yards figure is adjusted for each drive.  If there are no turnovers in a given game, then the figure is made the same, if not, the turnover adjustment is added or subtracted to the calculation.  This new is then divided by the original yards required to score a touchdown number, giving us the Turnover Adjusted Offensive Efficiency.

We turn one last time to the Cowboys Bears Box score:

Yards Required to score Yards gained + penalty yards Offensive Efficiency Turnover Yardage Conversion Adjusted Yards Gained Turnover Adjusted Offensive Efficiency
80 33 41.25 0 33 41.25
84 25 29.76190476 0 25 29.76190476
74 12 16.21621622 0 12 16.21621622
80 1 1.25 -61 -60 -75
80 80 100 0 80 100
80 68 85 -26 42 52.5
27 0 0 -113 -113 -418.5185185
80 68 85 0 68 85
85 35 41.17647059 -23 12 14.11764706
83 3 3.614457831 -17 -14 -16.86746988
89 89 100 0 89 100
842 414 49.16864608 -240 174 20.66508314

As you can see, the Cowboys gained 414 yards in the game, a nominally good outcome.  This outcome shifts from good to average when Offensive Efficiency is considered, as the Cowboys gained less than half the yards necessary to score a touchdown every given drive.  Next we consider turnovers, in this game Tony Romo threw five interceptions, so the adjustment is rather extreme, but it illustrates the point well.  When turnovers are factored in the Cowboys performed as if they gained only 174 yards of total offense, which as any casual football fan can tell you is a putrid performance.  The Adjust Yards Gained is then divided by the Yards Required to Score, giving a Turnover Adjusted Offensive Efficiency rating of 20.66508314 to Dallas.  For a point of reference Baltimore in their season opener against Cincinnati had an TAOE of 74 and in their blowout loss to Houston had a TAOE of just under 10.  By measuring TAOE against (1-TAOE) TADE can be calculated in this case Chicago had a TADE of 79.33 (the maximum being 100) a number that far better reflects their defensive performance than yards allowed, the stat the NFL uses to measure total defense.  Further a Turnover Adjusted Efficiency (TAE) rating can be calculated by taking (TAOE – (100-TADE)).  Throughout the course of the upcoming NFL season I will calculate TAE, and use it to create my power rankings.  Note that it like any mathematical calculation is not effective until the sample size is large enough, and is not an effective gambling tool until all teams have played six games.

The Expected Value of a Turnover

Beyond the obvious shifts in momentum, which may or may not exist in actuality.  The value of a turnover has rarely been quantified.  In this post, I will argue that a turnover is roughly akin to a team taking a sack equal to the value of its net punting average, minus any additional yards beyond the line of scrimmage that the turnover is returned, mathematically (-net punting average – yards past the line of scrimmage the ball is returned).  For example the Dallas Cowboys net punting average was 40 yards, apparently the math and football gods aligned on this one, also Tony Romo’s proclivity to throw interceptions in bunches will come in handy later in the post.  In this case, any turnover in which the ball traveled forty yards down field was roughly equal to a punt, the exception being that downs were lost. From this point forward, the negative consequences of a turnover beyond worse.  Should the ball be intercepted 30 yards down field, and returned 10 yards, this interception was in essence a shanked punt, or an explosive punt return.  When the original line of scrimmage is passed on the return that the effect of a turnover is greatest.  Not only is the net punting yardage lost for the yards effectively gained, but also, added to this loss of yardage is every yard beyond the line of scrimmage that the turnover is returned.

romo

To use a game example, we will stick with the Cowboys because one their net punting average was the nice round 40 yards, and two Tony Romo is a turnover machine.  Say the Cowboys start a drive on their own 35 yard line.  They gain 15 yards to the 50 yard line, and then Tony Romo throws an interception caught on the opponent’s 10 yard line and returned 0 yards.  When the equation I derived above is applied it is shown that this interception has no effect on the game beyond the loss of downs which are not accounted for in the equation, (-40+40) = 0.  With each yard closer to the 50 yard line that the ball is returned the effect of Romo’s interception becomes greater.

To accurately judge the effect of interceptions one must add this effective yardage lost to the yards gained in any drive.

Take Tony Romo’s Turnover Fest Against the Bears last Year

Yards Required to Score Yards Gained + Penalty Yardage Turnover Yardage Conversion Adjusted Yards Gained
80 33 0 33
84 25 0 25
74 12 0 12
80 1 -61 -60
80 80 0 80
80 68 -26 42
27 0 -113 -113
80 68 0 68
85 35 -23 12
83 3 -17 -14
89 89 0 0
842 414 -240 174

The Cowboys’ offensive performance appears to be good on paper, the team gained 414 yards when penalties are counted, but this yardage statistic, what we call total offense fails to account for turnovers.  By applying the turnover adjustment, we see that the Cowboys had the equivalent of a 174 yard offensive performance, a much lesser result, and a yardage total that more accurately reflects their offensive performance.

Hidden Yards: Are Penalties the Walks of Football?

It has widely been assumed that penalty yards are flukes, but any astute observer of football can tell you that this is simply not sure.  Some receivers have a knack for drawing pass interference penalties, some lineman hold or get away with holding, former Ravens great Frank Walker literally committed pass interference every time he stepped onto the football field.  The time has come for football coaches to recognize that drawing penalties or getting away with penalties is a skill that should be valued.  Moneyball explored how the Oakland Athletics were able to exploit the fact that walks, despite their basic equivalency to singles were chronically undervalued in baseball.  The walk lacks the excitement of a base hit, and similarly gaining 10 yards via a holding penalty simply is not as exciting as gaining ten yards running off tackle; there is no potential to score via penalty.  (The exception being holding in the end zone, but the point stands nonetheless.)  Football has consistently recognized that committing penalties is bad, but the sport has not fully recognized the benefit of drawing a penalty.  10 yards is 10 yards regardless of how those ten yards are gained.  Thus, as part of my “Towards a Better Box Score” series, I propose that yards gained or lost via penalty be calculated into the box score to better reflect total offense.  We turn again to our trusty Cowboys Bears box score, and see that I calculated into the yardage statistic not only the yards gained or lost by the offense, but also the penalty yardage gained or lost, and derive my raw offensive efficiency statistic from the following equation (yards gained plus penalty yards) / (yards required to score a touchdown) the latter being the distance from the end zone at the beginning of a drive.  I equate yards gained plus penalty yards to on-base percentage which though not as sexy a statistic as batting average is far more effective in evaluating a baseball player, and thus yards gained plus penalty yards is a far more effective statistic at evaluating a team’s offensive performance.

Yards Required to score Yards gained + penalty yards Offensive Efficiency
80 33 41.25
84 25 29.76190476
74 12 16.21621622
80 1 1.25
80 80 100
80 68 85
27 0 0
80 68 85
85 35 41.17647059
83 3 3.614457831
89 89 100
842 414 49.16864608

Towards A Better Box Score

Chris Brown of Smart Football recently published an article describing what he would like to see in a box score, to better increase the understanding of what occurred during a football game.  To me what is required is a simple and easy fix.  The box score should at some point measure team offensive performance in addition to individual offensive performance.  The box score should not only list the yards gained and the penalty yards gained by a team on a given drive, but it should also list the yards required to score a touchdown at the start of the drive.  A forty-five yard drive beginning on your opponent’s 45 yard line is ultimately much more effective than a forty-five yard drive beginning on your 10 yard line, and a box score fails to capture this difference.  It is not so much how many yards one gains in a football game, but rather how efficient one is at gaining yardage.  The adage bend don’t break on defense is surprisingly accurate, in that so long as a team does not score, it is largely irrelevant how many yards one gives up.  The ultimate goal of a football game is not to gain yards, but rather to score a touchdown, and the box score should capture this.  If a box score is ultimately supposed to be a means of capturing how well a team performed on a given day.  Below is an example of my proposed box score, using the Cowboys Bears game from 2012, as it will come in handy regarding a later post.

Yards Required to score Yards gained + penalty yards Offensive Efficiency
80 33 41.25
84 25 29.76190476
74 12 16.21621622
80 1 1.25
80 80 100
80 68 85
27 0 0
80 68 85
85 35 41.17647059
83 3 3.614457831
89 89 100
842 414 49.16864608

In essence, Dallas achieved 49% of the yardage required to score on every drive.  This calculation is not yet adjusted for turnovers, and is obviously skewed because of their garbage time touchdown, but it calculates offensive performance in a manner better than the simple raw numbers of today’s box score.