The NFL today is considered a "passing" league. The aerial attack is more important than ever. Thus, it gives rise to the importance of skill positions like quarterback and wide receiver. How do wide receivers factor into an NFL offense nowadays?
via www.soraspy.com
This is a statistical analysis of data on the NFL, specifically each team’s run to pass play ratio, their usage of various amounts of wide receivers, and the comparison of these data to their divisions, either NFC or AFC. The "Who" we are studying are the NFL teams. The "What" we are studying is the play tendencies of each team in the NFL, whether they are run dominant, pass dominant, or balanced, and their frequency of various wide receiver usages. The categorical variables are team, division/conference, and offensive philosophy. The quantitative variables are run-pass ratio and the various WR usages. The "Why" is because we want to see if there are any correlations between any of these variables or if any conclusions can be made from them.
The data sets will be broken down and displayed in various ways. The quantitative data will be displayed with histograms in conjunction with box plots. There will be box plot comparison charts that break down run-pass ratios in terms of division and frequencies of WR usages by amount of receivers. The categorical data will feature three things: a contingency table that compares division against offensive philosophy, a pie chart that breaks down the NFL by offensive philosophy, and a comparative bar graph that breaks down the offensive philosophies by division.
The key data set is the WR usage frequencies. How often do teams use 0-1 WRs? What about 2 WRs? 3 WRs? 4-5 WRs? By using the data for WR usage, it will be easier to compare the WR usage against the offensive philosophies in the NFL.
Make the jump for the fun stuff!
Original Data Sets
Team |
Ratio |
Run/Pass Dominant |
Division |
Arizona Cardinals |
0.57 |
Pass |
NFC |
Atlanta Falcons |
0.86 |
Balanced |
NFC |
Baltimore Ravens |
0.99 |
Balanced |
AFC |
Buffalo Bills |
0.77 |
Pass |
AFC |
Carolina Panthers |
0.88 |
Balanced |
NFC |
Chicago Bears |
0.89 |
Balanced |
NFC |
Cincinnati Bengals |
0.73 |
Pass |
AFC |
Cleveland Browns |
0.86 |
Balanced |
AFC |
Dallas Cowboys |
0.74 |
Pass |
NFC |
Denver Broncos |
0.69 |
Pass |
AFC |
Detroit Lions |
0.62 |
Pass |
NFC |
Green Bay Packers |
0.78 |
Pass |
NFC |
Houston Texans |
0.74 |
Pass |
AFC |
Indianapolis Colts |
0.58 |
Pass |
AFC |
Jacksonville Jaguars |
1.1 |
Balanced |
AFC |
Kansas City Chiefs |
1.17 |
Balanced |
AFC |
Miami Dolphins |
0.8 |
Balanced |
AFC |
Minnesota Vikings |
0.87 |
Balanced |
NFC |
New England Patriots |
0.9 |
Balanced |
AFC |
New Orleans Saints |
0.57 |
Pass |
NFC |
New York Giants |
0.89 |
Balanced |
NFC |
New York Jets |
1.02 |
Balanced |
AFC |
Oakland Raiders |
1.03 |
Balanced |
AFC |
Philadelphia Eagles |
0.76 |
Pass |
NFC |
Pittsburgh Steelers |
0.98 |
Balanced |
AFC |
San Diego Chargers |
0.84 |
Balanced |
AFC |
San Francisco 49ers |
0.8 |
Balanced |
NFC |
Seattle Seahawks |
0.71 |
Pass |
NFC |
St. Louis Rams |
0.73 |
Pass |
NFC |
Tampa Bay Buccaneers |
0.87 |
Balanced |
NFC |
Tennessee Titans |
0.86 |
Balanced |
AFC |
Washington Redskins |
0.58 |
Pass |
NFC |
TEAM |
0-1 WR Freq. |
2 WR Freq. |
3 WR Freq. |
4-5 WR Freq. |
ARI |
6% |
34% |
29% |
31% |
ATL |
26% |
32% |
36% |
6% |
BAL |
10% |
45% |
41% |
5% |
BUF |
3% |
26% |
46% |
25% |
CAR |
16% |
47% |
31% |
7% |
CHI |
11% |
29% |
50% |
10% |
CIN |
6% |
31% |
57% |
6% |
CLE |
13% |
45% |
40% |
2% |
DAL |
20% |
34% |
43% |
3% |
DEN |
13% |
28% |
53% |
6% |
DET |
2% |
39% |
55% |
3% |
GB |
13% |
27% |
41% |
19% |
HOU |
11% |
46% |
40% |
4% |
IND |
2% |
19% |
72% |
7% |
JAC |
16% |
38% |
43% |
2% |
KC |
15% |
47% |
34% |
4% |
OFF |
0-1 WR Freq. |
2 WR Freq. |
3 WR Freq. |
4-5 WR Freq. |
MIA |
16% |
36% |
42% |
7% |
MIN |
9% |
39% |
46% |
6% |
NE |
15% |
39% |
40% |
6% |
NO |
13% |
34% |
38% |
16% |
NYG |
12% |
44% |
40% |
4% |
NYJ |
12% |
46% |
39% |
4% |
OAK |
7% |
50% |
42% |
1% |
PHI |
3% |
34% |
51% |
12% |
PIT |
16% |
30% |
41% |
13% |
SD |
13% |
53% |
29% |
5% |
SEA |
9% |
40% |
47% |
5% |
SF |
16% |
41% |
37% |
6% |
STL |
5% |
35% |
49% |
10% |
TB |
11% |
46% |
41% |
2% |
TEN |
6% |
49% |
44% |
1% |
WAS |
12% |
41% |
42% |
5% |
NFL AVG |
11% |
38% |
43% |
8% |
*All Data are courtesy of Yahoo! Sports and Football Outsiders.
-----------------------------
The following analyses are heavy on statistical jargon, so here is a glossary of sorts to prepare.
Mean, Median, Mode: The mean is an average of the data. The median is the center-most point of the data set. The mode is the most-occurring value in the data set. Terms like uni- or bi-modal describe how many modes, or "peaks" show up in the histograms.
IQR/Standard Deviation: The IQR is a tool to help analyze box plots. The range of values within the "box" is the IQR. Standard Deviation is used in conjunction with the mean. It helps to analyze how much variability exists in our data set. A low standard deviation shows less variability and more consistency.
Outliers: Outliers are any data values that do not represent the data set well. On the box plots, they are represented by circles or stars (super outliers).
Skewness: This just tells whether data values are stuffed to one side or the other (imbalance).
If that Statistics lesson was not enough to whet your appetites, then keep going! The fun is just starting!
----------------------------
In regards to the histogram and box-plot of the frequency of NFL run-pass ratios, if a team has a ratio of less than .8, they are considered a pass dominant team, between .8 and 1.2, a balanced team, and greater than 1.2, a run dominant team. Looking at the histogram and box-plot, we can see the distribution is roughly symmetric, unimodal, and no obvious skewness occurs. However, it should be noted that there is one outlier of 1.17, as seen in the Kansas City Chiefs. This does not skew the data much however, because the high frequency of balanced ratios, centering around .7 to .9, narrowed the IQR significantly, making the not very distant value of 1.17 appear as an outlier. Because the maximum value of 1.17 is lower than 1.2, it is noted that no NFL teams are predominantly run based. This may be because of the heightened physicality of today’s game, a renewed focus in the passing game, or both. The minimum value is .57, as seen in the New Orleans Saints and Arizona Cardinals, indicating that they pass almost twice as must as they run. The mean of the distribution is .818 and the median is .82, reinforcing the lack of skewness in the distribution. In terms of the spread, the IQR is .16 and the population standard deviation is approximately .15.
Analyzing the same data but now in terms of each team’s respective division, more conclusions may be made. The AFC’s distribution is unimodal at its center, roughly symmetric, and with a slight bit more of frequencies occurring on the right then on the left, as seen in the higher ratios of the Jaguars and Chiefs. The mean, median, IQR, and population standard deviation of the distribution are approximately .88, .86, .25, and .16 respectively, showing the symmetry and lack of significant skewness. The NFC’s distribution is bimodal and skewed to the right, with all values occurring from .5 to .9 and from .8 to .9 being the most frequent. The mean, median, IQR, and population standard deviation of the distribution are approximately .75, .77, .205, and .12 respectively. Because there are no outliers that can skew the mean, the mean and median are both close in value and accurate indicators of the center.
In comparing the AFC to the NFC, one sees that the AFC’s distribution is more symmetric and evenly distributed, from .58 to 1.17, where as the NFC is highly concentrated, skewed to the right, and exists only from .57 to .89. None of the NFC’s teams goes beyond the ratio of .9, meaning that all of their teams are passing-inclined balanced or pass dominant. However, the AFC has multiple running-inclined teams (ratio greater than 1.00) and passing-inclined teams (less than 1.00). The AFC’s range is .59 and almost double that of the NFC’s range, .32. The NFC’s mean and median are both lower by almost .1 than the AFC’s, showing that both centers show the NFC’s larger tendency to pass more. The conclusion that can be made from these two graphs is that the AFC is largely balanced, with both running and passing-inclined teams, where as the NFC is holistically a passing-inclined or dominant conference with neither conference sporting a run dominant team.
Continuing on with the data that summarizes the offensive philosophies, it can be seen that in the entire NFL, there is almost an even distribution of Pass Heavy and Balanced teams, balanced teams outnumbering Pass Heavy by only 12 percent or 4 teams. The NFC has an almost even distribution of pass dominant to balanced teams, where as the AFC’s amount of balanced teams exceeds the amount of pass dominant teams by 120%. The NFC has 80% or 4 more pass dominant teams than the AFC, and the AFC has 4 more balanced teams than the NFC in return. This helps verify our previous conclusion that the NFC is highly passing-inclined while the AFC is far more evenly distributed and balanced.
Starting with the first wide receiver usage bar chart, we can see that the distribution is roughly symmetric, unimodal, and has no obvious skewness to note. The maximum value is seen with the Atlanta Falcons, using the 0-1 WR setup 26% of the time, while there is a tie for the minimum between the Detroit Lions and Indianapolis Colts at 2%. This makes the range 24%. The mean, median, IQR, and population standard deviation of the distribution are 11.19%, 12%, 8.5%, and 5.34% respectively. It may be noted that though the Lions and Colts both use this setup only 2% of the time, a correlation should not be made because while the Lions are considered one of the best offensive teams in today’s game while the Colts were one of the season's train wrecks.
Moving on to the second WR usage bar chart, it can be seen that the distribution is skewed to the right and is roughly uniform and symmetric about its center. There is a possible bimodality from 30 to 35% and 45 to 50%, and there is no outlier though there is only one team that uses this setup only 19% of the time, the Colts. This fact is significant considering that the mean is twice that value, 38.25%. The maximum value occurs with the San Diego Chargers, another contending (pretending?) playoff team, at 53%. Using these values, we see that the range is 34%. Furthermore, the median, IQR, and population standard deviation are 39%, 12.5%, and 7.97% The same analysis can be made that symmetry allows for the median and mean to be close in value as well.
Continuing on to the 3 WR setup, the distribution is roughly symmetric, unimodal with 14 teams using this set 40 to 45% of the time, and no obvious skewness is to be noted. There is one super outlier at 72%, held once again by the Colts. This may be attributed to their strength and success in passing and moving the ball to multiple different receivers, making them one of the biggest offensive threats in the NFL to date. The minimum value occurs with the Arizona Cardinals and San Diego Chargers at 29%, a very high minimum in general, occurring over one-fourth of the time. This makes the range an unusually high 43%. The mean, median, IQR, and population standard deviation of the distribution are 43.41%, 42%, 8.5%, and 8.5% respectively. Despite the Colts’ super outlier, the otherwise almost perfect symmetry of the distribution left the mean relatively unskewed, leaving the median and mean close in value.
Analyzing the final possible set of WR, the 4-5 setup, a skewness to the right, unimodality, and lack of symmetry can be noted in this distribution. There are many outliers, because of the high concentration of frequencies about the distribution’s center. This narrowed the IQR significantly, leaving the New Orleans Saints and Green Bay Packers as outliers, and the Arizona Cardinals and Buffalo Bills as super outliers. It should be noted that excluding the Bills, who have a losing track record as of the most recent seasons, all of these teams have either won or participated in the Super Bowl in the past three years. The minimum value is achieved by the Tennessee Titans and Oakland Raiders, at a mere 1%, the lowest value of all WR sets. The Arizona Cardinals experienced the maximum value of 31%. The range is therefore 30%. The mean, median, IQR, and population standard deviation are 7.59%, 6%, 4.5%, and 6.68%. The various discrepancies in the values are due to the strong skewness to the right and high number of outliers.
In conclusion with the final graph, the 3 and 2 WR sets were the most frequently used by the NFL teams with the means being 43% and 38% respectively. The 0-1 and 4-5 sets were used far less, with means of 11% and 8%. The 4-5 WR set is the most unusual distribution, being the only significantly skewed distribution and having the most amount outliers as well. The other distributions were relatively symmetric and not skewed for the most part. The 3 WR set has the highest range of 43% because of the super outlier of 72% seen in the Colts.
---------------------------------------------------
In conclusion, the 3 and 2 WR sets were the most frequently used by the NFL teams with the means being 43% and 38% respectively. The 0-1 and 4-5 sets were used far less, with means of 11% and 8%. The 4-5 WR set is the most unusual distribution, being the only significantly skewed distribution and having the most amount outliers as well. The other distributions were relatively symmetric and not skewed for the most part. The 3 WR set has the highest range of 43% because of the super outlier of 72% seen in the Colts.
Now comparing this against the offensive philosophies in today’s NFL, it is notable that 2 and 3 WR sets tend be the most common. With the virtual abolishment of the NFL as a run-heavy league, 0-1 WR sets that feature lots of blockers up front are unnecessary. Only balanced offensive teams like Atlanta, which uses the set on 26% of their plays, involve the 0-1 WR formation. With the proliferation of the NFL as a passing league, where quarterbacks are the dominant scorers, 2 and 3 WR sets become a lot more common. In fact, the Indianapolis Colts, who have one of the lowest run-pass ratios at .58, run the 3 WR formation an astounding 72% of the time. 4-5 WR sets are still not dominant but are becoming more involved in offenses like Green Bay and New Orleans, who use the set 19% and 16% of the time, respectively. These high-powered offenses not only have elite quarterbacks, but also a lot of depth at WR to make the usage of such sets possible. This "spread" offense should only continue to grow in usage.