When I watch a basketball game, I often find ​myself looking at the total points a player has ​scored or what they are averaging in a ​particular season. Nowadays, players in the ​NBA can easily score high amounts of points ​without even trying. To test out my beliefs, I ​decided to form my own hypothesis test to ​find out whether or not a NBA player can ​average more than 15.0 PPG.

Historically, the typical NBA player is found to average ​around 8 to 12 points per game. Players that are viewed to ​be the “elite” or “stars” of the NBA typically average around ​20 to 30 points per game. Using advanced basketball ​metrics/analytics, similar studies have been conducted ​before to evaluate a player’s point per game average. Using ​ESPN’s website of the player statistics for the 2023-2024 ​NBA season, I was able to use their dataset of the player’s ​PPG average and conduct a study of my own as to whether ​a random sample of 30 NBA players has a mean PPG ​average higher than 15.0.

The first graph I chose ​to display was a ​distribution plot. Since ​my hypothesis test ​was dealing with a ​mean greater than 15.0 ​PPG, I shaded in the ​region underneath the ​curve to demonstrate ​that. The graph also ​shows a bell-shaped ​curved and follows a ​normal distribution.

My histogram is skewed to ​the right and displays one ​gap from 28-32 PPG. The ​numerical values on the ​graph range from less than 0 ​PPG to about 37 PPG. The ​highest peak on the graph is ​around 5 PPG.

-Random: MET To conduct my test, I took a random sample of ​30 NBA players from ESPN’s website that displayed a list of ​players in order based on their season statistics. The better ​their stats were, the closer they were to being #1. Since there ​were 573 total NBA players used on the website, which is my ​population, I used a random number generator on Google to ​generate a random sample of 30 players. I pressed the ​generator button 30 times to come up with my sample size. My ​minimum value was set to 1 and my maximum value to 573.

-10% Condition: MET My sample ​size of n=30 is less than 10 % of the ​entire population of NBA players in ​the league, so the condition is met.

-Normal/Large Sample: MET ​My sample is equal to 30, so ​the condition is satisfied.

Sample Mean (x̄): 8.61

Standard Deviation of Sample (Sx): 7.95

T Test on Minitab:

t= 4.40

P= 1.000

Since my P value of 1.00 is greater than the ​significance level of α= 0.05, we would fail to ​reject Ho. We do not have enough evidence ​that the true mean PPG (Points Per Game) in a ​random sample of 30 NBA players in the 2023-​2024 season is greater than 15.0 PPG.

After conducting my hypothesis test, a weakness that I noticed was how I got a ​duplicate PPG average on one of my random samples. As a result, the accuracy ​of my test is a little misleading because I wanted to have different point values. ​Another weakness in the study was how the ESPN website listed out the players. ​Instead of listing out the NBA players by their PPG, they chose to list every ​player based on every stat. This shows a weakness in the study because a player ​could be ranked higher on the list because their all-around stats were better ​than someone else even though that player could be averaging more points. A ​final weakness in the study was how the website only showed the PPG for all ​players in just the NBA. Per NBA rules, teams are allowed to hold players on ​roster spots that are referred to as “Two-Way Contracts.” This means that a ​player can play on both the NBA team and the G-League affiliate for that team. ​This will affect my study because a player could be averaging more points ​against weaker competition in the G-League, but the ESPN site only accounts for ​the stats they have put out in the NBA.

To generate more accurate results from my hypothesis test, I would suggest to ​the ESPN website to include the stats that a player in the G-League is averaging ​to gain a better sense of their playing abilities. Next to the stats though, I would ​suggest that an asterisk should be added to inform people that those players ​stats are combined with their time in the G-League and in the NBA. By doing this, ​the players stats are more representative of their playing abilities and can allow ​for this test to be more accurate. Another suggestion I would make to the ESPN ​list is to make different lists of a specific stat so that we don’t have to look ​through a list that takes into account all stats when ranking players from best to ​worst. This change in the list would make it easier for the readers to find PPG ​averages and conduct hypothesis tests more efficiently.