Evaluating the Worst Basketball Players of the Past Decade

By looking at the worst rated NBA players over the past decade, I will reveal the players that most frequently contribute to poor individual and team performance, and then determine what basketball related statistics they have in common that could help show why they produce so much losing..These factors could possibly include poor shooting percentages, subpar defensive ratings, a reluctance to pass the ball, or maybe just the fact that a player is just not very good at the game of basketball.Data ScrapingThe first decision I had to face when it came to data scraping seems simple but in fact is a much more complicated problem: what data should I go after?.Below is a screenshot of data shown on the website when looking at the 2017–18 season:With that being the case, I then took the 100 players for each season from 2009–2018 (I didn’t include data from the current season because there isn’t enough of it yet) with the lowest win share totals and exported this data into a CSV file..This data not only include the player names and their respective win share totals but it also include other advanced metrics that will be used later to help see what causes these low win share numbers.Finally, in order to analyze this data, I decided to utilize R for this project..Therefore, I decided to develop a frequency table that showed how often a players name showed up..Here is a snap shot of the created table:This table is now very useful, as I can focus in on the players that are consistently bad over a long period of time.Automatic Variable Selection to identify why these players are so bad:Now that I have a more concise set of data, I can begin to perhaps identify what makes these players so bad when it comes to producing wins..With that being said, the following table marks the three variables (ignoring Intercept) that make up the high R² model with “TRUE” underneath it:As shown above that the following three variables have the highest effect on win share totals then all of the other listed variables: player efficiency rating, 3 point attempt rate, and usage percentage.Conclusions and Recommendations:In order to provide useful information about the above findings, it is vital to understand what the three automatically selected variables represent.. More details

Leave a Reply