Introduction
My name is Kyle Martin, and I am a senior Business Analytics major from Kansas City, Kansas. I have always been interested in watching and playing sports, and more specifically football. That being said, NFL data is what I will be doing analysis on in this post. I have found a data set on the website Kaggle, that contains NFL team data from 2003-2023. Twenty years of data will give a nice clean perspective of what drives winning. That is the question I will be setting out to answer, what stats, if any, specifically drive winning in the National Football League.
Data
[1] "year" "team" "wins"
[4] "losses" "win_loss_perc" "points"
[7] "points_opp" "points_diff" "mov"
[10] "g" "total_yards" "plays_offense"
[13] "yds_per_play_offense" "turnovers" "fumbles_lost"
[16] "first_down" "pass_cmp" "pass_att"
[19] "pass_yds" "pass_td" "pass_int"
[22] "pass_net_yds_per_att" "pass_fd" "rush_att"
[25] "rush_yds" "rush_td" "rush_yds_per_att"
[28] "rush_fd" "penalties" "penalties_yds"
[31] "pen_fd" "score_pct" "turnover_pct"
[34] "exp_pts_tot" "ties"
Data Dictionary
# A tibble: 35 × 2
variable description
<chr> <chr>
1 year year
2 team team name
3 wins wins
4 losses losses
5 win_loss_perc winning percentage
6 points points scored
7 points_opp points against
8 point_diff point differential
9 mov avg margin of victory
10 g games played
# ℹ 25 more rows
Correlation With Wins
The first visualization I will look at is to make a chart that shows which stats correlate positively with wins, and which stats correlate negatively with wins.
Takeaway: The main take away from this chart is that penalties, turnovers, and points allowed negatively effect a teams chance of winning, while scoring points is the best way to positively effect your chance of winning. The most controllable metric outside of scoring (either against you or for you) is passing touchdowns and yards per play.
Points Scored vs. Wins
The next metric I will take a look at is the amount of points scored vs. wins. Since we just learned that points for is the best way to positively effect wins, I want to take a closer look at the direct correlation between the two.
Takeaway: Almost every team with 400+ points finishes above .500; whereas every team below 280 finishes under it. Exceptions tend to be teams with catastrophic turnover problems or weak defense. Scoring alone isn’t sufficient, but it is necessary.
Turnovers vs. Wins
Now we will take a look at the other side. I want to see if turnovers negatively effect wins the same way that points for positively effects wins.
Takeaway: The negative correlation between turnovers and wins has existed throughout the entirety of the 21st century, even as teams and players have gotten better throughout the years. There has been only one team over this 20 year period who was able to reach double digit wins while having 40+ turnovers.
Elite vs. Struggling Teams
Takeaway: The biggest separations between elite and struggling teams are in points scored, points allowed, and turnovers. Yards per play gives a minimal edge to elite teams, but over the course of a 16 game season (now 17) those few yards per play add up. Especially when struggling teams can’t get first downs.
Does the Pattern Hold in 2025
Next I will look at some of the same comparisons, but using data from the 2025 season. I scraped the data from ESPN.com, with the same variables. This will give us a look at whether these trends that took place over a 20 year period, hold up in a one year sample size.
Takeaway: The historical trend seems to hold up. Teams that scored more points, won more games. Almost every team that scored 400+ points had a winning season, whereas teams that scored less than that seemed to struggle
Conclusion
There are many factors that go into teams being successful, or falling flat, but we can determine that there are certain factors that matter more. Scoring, or being scored on, drastically changes whether or not that team will be able to win football games at a high rate. Whether or not a team can hold onto the ball also plays a big role in the success of that team’s season.
