Principal Component Analysis of NBA Shooting Success in R
The goal of this was to find the factors which have the highest correlation and level of variance on NBA three-point shooting success — seeing as so much of NBA scouting now revolves around the ability to make NBA threes at a reliable mark.
The dataset used is from www.basketball-reference.com/ and www.sports-reference.com/cbb/. The dataset is comprised of the Top 50 NBA Players by 3 Point Attempts in the 2018–19 regular season. From these players, data of their 2018–19 NBA 3P%, Career College 3P%, Career College FT%, Wingspan, and Height were used. The ‘y data’ I was looking to find the principal components for was NBA 3P%, while the other four ‘x’ factors would be used to find which had the highest impact on three-point success.
Using R statistical processing software, I can select which data I would like to comprise the ‘x-factors’, and which data to be the ‘y factor’. Inputting the above data for the top 50 players by makes gave me the following results:
From this data, we can come to some conclusions. If we look at the right column we can see the t-values of the factors. A t-value is statistically significant at less than 0.05. As we can see, both Career College FT% and Career College 3P% are statistically significant, with FT% having a slightly higher correlation with NBA volume three-point shooting success. The other factors of Wingspan and Height had little impact on success. It should be noted however that because of the low Multiple R-Squared value of 0.2905, this data will not project terribly well in models.
Although this conclusion may seem obvious, NBA teams continue to draft college prospects with high draft picks with the assumption that their success from the college three-point line will translate to the NBA, despite their immense struggles shooting college free throws. A look back to the 2017 NBA Draft makes this issue increasingly clear.
All of these prospects were heralded not only as versatile players but players who could shoot threes well in the NBA due to their college success. Of these players, Jayson Tatum, who had the lowest College 3P% but the highest College FT% has thus far translated into far and away the best NBA shooter. Meanwhile, the other three top picks, who all struggled to hit college free throws, have struggled immensely shooting the ball at the pro level and have each been traded just two years removed from being drafted. The point of this is not to say that college three-point success doesn’t matter (because it does), but rather to point out that what has a statistically higher rate of correlation to NBA success is college free throw percentage.
