top of page

Exploratory Analysis

nfl_edited.png

Client: NFL

Role: Data Analyst

Duration: Nov 2023 (4 weeks)

Process Deck 🚧

Contributer(s): Tyler Gustafson

Explore more of Abhishek Shetty's Home Field Advantage Analysis and Bikram Khaira's deep dive into specific NFL team offensive strategies here!

DISCLAIMER: Please note this analysis was conducted for a graduate level project

An Analysis of NFL Offenses (2009 - 2017)

Overview

Explore the intricacies of NFL strategies with the "NFL Offensive Play Analysis 2009 - 2017." This comprehensive report dives into eight years of play-by-play data, examining how offensive plays evolve over games and impact the final score! Determine how can analyzing NFL play data from the 2009 - 2017 seasons inform the development of effective offensive and defensive coaching strategies by highlighting key aspects and risks?

research_question.jpg

My Focus!

Initial Exploration

At first glance, this snapshot of the season averages almost appears that the NFL hasn’t changed much:

  • Average Total Points are about the same at ~42 points

  • Total yards actually go down

  • And penalties are up!

Does this mean the NFL has gotten more boring or stale?

inital_exploration.jpg

No. First, the lower average total yards can be accounted for the rule change that occurred in 2016 where touchbacks start 5 yards closer (5 yards X 5 average number of TDS per game = an already 25 yards per game boost) and then if we account for the increased penalty yards we are at approximately the same number of yards gained a game.

Now if we look closer at those total points that have remained stagnant, we can see a couple of trends emerge. The above two graphs look at two things:

1) # of big losses (defined as losses by more than 3 scores) and

2) # of Exciting Games (defined as a combined score of > 49 pts (7 TDs) and a final score decided <3)

First, we can see a clear pattern trending down of these less interesting blowout games (~15-20 games less a season) and then an uptick in high-scoring close decided games.

The takeaway?

A more competitive interesting league to watch.

Now that we understand these dynamics let's spend some more time digging into what these offenses look like in games.

Impact of General Factors

Understanding the distribution of what plays are typically executed and where can provide valuable insights to any coach on both the offensive and defensive sides of the ball. When exploring the influence of field position on offensive strategies, we created a detailed chart that categorizes play types based on the distance remaining for a 1st down separated by the specific down. This chart offers insightful data on how teams adapt their play-calling in varying field positions:

play_type_highlevel.jpg

We can quickly see a handful of insights by Down:

  • On 1st Down: the data shows a balanced distribution between different play types, indicating a balanced strategy and that defenses should play conservatively

  • During 2nd Down: a notable shift in play selection occurs if more than five yards are gained on 1st down, with a higher inclination towards running plays

  • On 3rd and 4th Downs: the trend strongly favors passing plays, except in scenarios where only one yard is needed for a first down

play_type.jpg

All of these insights can inform an offensive and defensive coach’s strategy. We can also break this down even further by what direction the play is actually executed:

In addition to the insights we already highlighted, several more patterns emerge by play type:

  • Pass Plays: During 1st and 2nd downs, there's a tendency to steer clear of passes to the middle, with teams generally opting for passes to the left or right side of the field. On 3rd and 4th downs, though there's a slight increase in passes directed towards the middle. Finally, it appears the closer the quarterback is to the 1st down the more likely they are to throw a right-sided pass.

  • Run Plays: On 1st down, the direction of running plays is balanced, showing no strong preference for any side. However, in later downs, the direction for running plays varies more, influenced by the distance remaining to a first down. Left or right-sided runs are preferred until only a yard remains.

These are all helpful insights as play moves down the field, but let's do a double click into the redzone and see what happens to play distribution when the team is only ten yards away from the endzone.

play_type_endzone.jpg

When we look at the chart above we can see that passing plays dominate near the end zone until the 1-yard line, where preferences shift. Right-sided passes are favored, while middle passes are rare. Now that we have this understanding let's take a closer look at what areas of the field might be high risk areas for turnovers (fumbles and interceptions).

turnover_yardline.jpg

According to the data, running backs need to be extra cautious protecting the ball across mid field and running on the left side or middle. Quarterbacks, however, should be extra cautious about throwing the ball across the middle. One interesting link from the previous chart is we see the lowest interception rate is the right pass within the goal line which was also the highest play type from our heatmap on the previous slide so coaches are coaching to minimize mistakes near the endzone.

The final lens we want to understand is what is the impact of time pressure on critical errors:

time_pressure_criticalerrors.jpg

We see that time pressure increases turnovers in the late 2nd and 4th quarters and drives up 2nd-quarter penalties. This is a good indication that coaches should ensure their teams are extra disciplined going into the end of the first half.

Future Investigations

This is only the foundation for a great analysis. There's so much more to explore. What happens if the quarterback is left-handed? How about following a coach's career to better prepare against their strategies? We've got plenty of analysis ahead that can help teams get ready for their next big game! 🏈🏆

Check out the code and more on my GitHub

GitHub-logo.png
bottom of page