The Statistic That Wins NBA Games
- Bruin Sports Analytics
- Apr 2
- 5 min read
By: Anika Soitkar and Elly Goan

Introduction
The National Basketball Association, more commonly known as the NBA, consists of 20 teams across the continent of North America, featuring arguably the most competitive basketball league in the world. Each team plays through a regular season, in which they play 82 games against various opponents. At the end of the season, 16 of the best teams (8 from each of the two conferences) qualify for the playoff, where they battle in a bracket-style format to determine a champion.
One of the biggest debates among analysts, coaches, players, and fans is: Which statistic is the best predictor of winning? Sports fans may hear that a basketball team can only win games by scoring more three-pointers. Others may argue that making their free throws is the most important statistic.
In this article, we’ll explore the 2023-2024 NBA season and analyze which NBA statistic has the strongest impact on a team's success.
Definitions
Field Goal Percentage (FG%): The percentage of shots made from the field (two-pointers and three-pointers) out of all attempted shots.
Three-Point Percentage (3P%): The percentage of three-point shots made out of all attempted three-point shots.
Free Throw Percentage (FT%): The percentage of free throws made out of all attempted free throws. Free throws are unopposed one-point shots, awarded after certain fouls by the opponent.
Rebound (REB): A player gains possession of the ball after a missed shot.
Steal (STL): A defensive player takes the ball away from an offensive player (without committing a foul).
Turnover (TOV): The offensive team loses possession of the ball before attempting a shot (includes steals, bad passes, and rule violations.)
Personal Foul (PF): A violation that occurs when a player makes illegal physical contact with an opponent (includes hitting, pushing, or holding).
Methodology
To visualize the distribution of various NBA statistics in winning and losing games, we analyzed game data using box plots. Each statistic is represented by two box plots—one for games that resulted in a win and one for losses—highlighting differences in medians, variability, and outliers. Box plots provide a clear and concise visual comparison, making it easier to identify different factors and trends that may influence a team's success.
The Data
Field Goal Percentage

Shooting efficiency, rather than raw scoring totals, reflects a team’s ability to maximize possessions. As demonstrated by the boxplot model, there is a large discrepancy in the field goal percentage between games with a win and games with a loss. For a losing game, the median for a field goal percentage is 44.9% while the median for a winning game jumps up to 50%. This indicates a trend that favors a correlation between having a higher field goal percentage rate and winning.
Three-Point Percentage

Efficiency of making three-pointers also plays a role in a team’s ability to maximize scoring opportunities. The boxplot model shows that for losing games, the median three-point percentage is 33.9%, while for winning games, it increases to 39.0%. This difference also indicates a correlation between higher three-point shooting efficiency and winning.
Free Throw Percentage

While free throws provide easy scoring opportunities, their overall impact on winning appears to be less significant than shooting efficiency from the field. As seen in the boxplot model, the medians for the free throw percentages are close, with 77.8% as the median for game losses and 80% as the median for game wins. This small difference suggests that free throw percentage is not a major determinant of game outcomes. The presence of lower outliers for free throw percentage in winning games also shows that teams still win despite poor free throw percentages.
Rebounds

Total rebounds can also play a role in teams’ success rates, impacting their possession and second-chance opportunities. The boxplot model reveals that the median for total rebounds in losing games is 41, while the median for winning games rises to 45.This difference suggests a relationship between securing more rebounds and winning games.
Steals

Steals contribute to a team’s defensive efficiency and help limit the opponent’s chances to score. The boxplot model shows that the median number of steals in losing games is 7, while the median in winning games increases slightly to 8. Although the difference in the medians is small, there are more high outliers in the amount of steals in winning games, indicating a bit of correlation. This slight trend could be because steals can lead to more scoring opportunities.
Turnovers by Team

Turnovers can lead to missed opportunities and more chances for the opponent to score. As seen in the boxplot model, the median number of turnovers in losing games is 14, while in winning games, it is slightly lower at 13. Though turnovers are slightly more common in games that end in losses, the small gap indicates that they are not a major factor in differentiating wins from losses.
Personal Fouls

Personal fouls can indicate a team's defensive aggression, but they do not appear to have a significant effect on game outcomes. With the boxplot showing that the median number of fouls in games that result in a loss is 19 and the median in games that result in a win is 18. Although the median of the number of fouls in losing games is slightly higher than that in winning games, the insignificant amount of disparity suggests that their overall impact on the outcome of the game is minimal.
Correlation Coefficients

To further assess the strength of the relationship between different NBA stats and winning a game, we can look at the correlation coefficients. Correlation values range from -1 to 1, where 1 indicates a perfect positive correlation, 0 means no correlation, and -1 indicates a perfect negative correlation. As the magnitude of the coefficient gets closer to 0, the strength of the relationship decreases.
This model shows that field goal percentage is most strongly correlated to winning, with a correlation coefficient of 0.4963. This is followed by three-point percentage, rebounds, steals, and then free throw percentage. Personal fouls and turnovers have negative correlation coefficients, indicating that teams that commit fewer fouls or turnovers tend to have a slightly better chance of winning, but the impact is minimal.
Conclusion
In analyzing various statistics and their correlation with winning, the data suggests that an NBA team's indication of success, or winning a game, is most reliant on their field goal percentage. Although three-point percentage and rebounding also contribute to winning, their impact is less pronounced.
While these results suggest that teams prioritizing high-percentage shot opportunities are more likely to win, there are several limitations to consider. Since correlation does not imply causation, improving field goal percentage alone does not guarantee success. Additionally, this study does not account for the teams’ strength of schedule, which can cause statistics to look better or worse depending on the difficulty of opponents. Individual play contributions can also have a large impact on winning that is not captured by team averages.
Overall, this research emphasizes the significance of field goal percentage in shaping game outcomes but also highlights the need for additional advanced statistics in order to understand the many factors that contribute to an NBA team’s success.
Comments