Exploring Performance Metrics: NFL Combine Analysis

Background

In the National Football League (NFL), athletes exhibit a remarkable range of physical attributes tailored to the demands of their positions. Every year, the NFL Combine collects physical and performance data from new athletes declaring for the NFL draft. Scouts use this data to further evaluate player prospects beyond their on-field performances.

In this post, I will investigate the effectiveness of K-means clustering in identifying player positions using NFL Combine data from 2000 to 2018. Additionally, I will conduct several tests to investigate annual trends in the Cone Drill.

Exploratory Analysis

The following code will be run using a Python 3 kernel in Jupyter Notebook. We’ll start by loading in the dataset and libraries below!

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Heading

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Heading

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Heading

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla euismod condimentum felis vitae efficitur. Sed vel dictum quam, at blandit leo.

Heading

Cluster 1 Results:

LR Slope: -0.007284025968538165

R-squared: 0.27525153202504027

t-statistic: -2.5409474310335876

p-value: 0.021099052456581946

Similarly for clusters 2 and 3…

Cluster 2 Results:

LR Slope: -0.0077117141767901794

R-squared: 0.36953339540415586

t-statistic: -3.156605168807558

p-value: 0.00576068502698579

Cluster 3 Results:

LR Slope: -0.01193051964700224

R-squared: 0.48944498115367974

t-statistic: -4.036966314785228

p-value: 0.0008558581836797252

 

 

Cluster 1: The average cone drill time for Power Players has been decreasing by approximately 0.0073 seconds each year. The R-squared value of 0.2752 suggests that about 28% of the variance in the average cone drill speed for this cluster can be explained by the year. The negative t-statistic and p-value suggest that there is a significant negative linear relationship between the year of the NFL Combine and the average cone drill speed. Therefore, we reject the null hypothesis and conclude that the average cone drill speed for Power Players has been decreasing over the years.

Cluster 2: Agile Players have also seen cone drill performance improve with their average completion time decreasing by approximately 0.0077 seconds each year. An R-squared value of 0.3695 suggests that about 37% of the variance in the average cone drill speed for this cluster can be explained by the year. The negative t-statistic and p-value suggest that there is a significant negative linear relationship between the year of the NFL Combine and the average cone drill speed. Therefore, we reject the null hypothesis and conclude that the average cone drill speed for Agile Players has been decreasing over the years.

Cluster 3: Hybrid Players have seen the greatest improvement of all with their average cone drill time decreasing by 0.0119 seconds each year. The R-squared value of 0.4894 suggests that about 49% of the variance in the average cone drill speed for this cluster can be explained by the year. The negative t-statistic and p-value suggest that there is a significant negative linear relationship between the year of the NFL Combine and the average cone drill speed. We again reject the null hypothesis and conclude that the average cone drill speed for Hybrid Players has been decreasing over the years.

To investigate these results further, we will plot all three clusters alongside their regression models for comparison.

Amin Abbasi

© Copyright. All rights reserved.

We need your consent to load the translations

We use a third-party service to translate the website content that may collect data about your activity. Please review the details in the privacy policy and accept the service to view the translations.