NFL once again published a huge amount of data on Kaggle – free for Data Scientists to analyse and create insights. I wanted to know, what we actually can learn from the record of 20 000 plays of the 2018 season.
Here we go…
The dataset contains the plays of 253 games between Sep. 6th 2018 and Dec. 30th 2018. Per game, there were ~76 plays, the minimum number of plays was 41 whereas the maximium was at 117 plays in one game.
Looking at the distribution of the plays, the majority of them happen – as we would expect – in course of the 4 quarters and some plays happen during Overtime (5th). Interestingly, the second and fourth quarter have significantly more plays than the first and the third…
1st down plays: 39%; 2nd down plays: 33%; 3rd down plays: 27%; 4th down plays: 1.8%
Looking at the Defense variables in the data set, we can clearly see below, that the most frequently used defense setup contains four Pass Rushers (60% of the plays). The achieved Yards by the Offense lies in the median at 5. Interestingly, the analysis also shows, that if the defense adds one more Pass Rusher, the offense effectivity decreases down to 3 Yards (in the median). I am ignoring the other Defense Formations as the respective amount of plays with Pass Rushers below 4 and above 5 are too low for any interpretations.
The offense variables show, that the first and second down do have the highest chance of gaining more Yards compared to the third and fourth down.
The following density chart shows that roughly 50% of the plays have 10 Yards to Go.
Looking now at the effectiveness of the offense plays, we can see that the most played offense formation is the “Shotgun” formation, resulting in roughly 75 000 Yards achieved in all recorded games.
6.58% (1260) of the plays result in a Quarterback Sack
Looking at the quick analysis, it clearly shows why Football is such an exciting sport. Gaining Yards for the Offense is obviously fairly hard and that makes it so exciting to watch.