Modern professional basketball relies heavily on advanced statistics to evaluate team performance. This project examines whether regular-season data can be used to estimate a team’s probability of winning the NBA championship. Using team-level data from the 1995-1996 season through the present, collected from Basketball-Reference.com, we apply regression modeling to analyze championship outcomes.

Using IBM SPSS, multiple linear regression is first used to identify the statistical factors that contribute most strongly to overall team strength. These factors include offensive and defensive efficiency metrics such as shooting percentage, turnover rate, and free-throw rate. Building on this analysis, logistic regression is then used to estimate the probability that a team wins the championship based on its regular-season performance and playoff seeding.

This study demonstrates how classical statistical methods can be applied to modern sports analytics, illustrating how mathematical modeling helps explain and predict competitive success in professional basketball.

References

Logistic Regression - https://en.wikipedia.org/wiki/Logistic_regression

Linear Regression - https://en.wikipedia.org/wiki/Linear_regression

Least Squares - https://en.wikipedia.org/wiki/Least_squares

Ordinary Least Squares - https://en.wikipedia.org/wiki/Ordinary_least_squares

Where all the data is from - https://www.basketball-reference.com/