Fellow NBA enthusiasts, we are back again with a post dedicated to analyzing the impact of rest in the NBA’s grueling schedule. First and foremost, we hope our readers are safe and sound from the COVID pandemic. Due to how the pandemic has shaped the landscape of sports and how the NBA implemented a prolonged rest before the Disney bubble, I was curious how rest influences NBA athletes’ play.
Every year, there is a discussion to shorten the NBA schedule due to the rigorous effect that it has on players’ bodies. Here’s the interesting resulting find from this study: strictly from a stats perspective, the amount of rest during a regular NBA schedule i.e. back to back games versus games played after 2 days of rest actually DOESN’T impact a player’s play.
Because the 2019-2020 season was shortened and due to a lack of data, I looked at 2018-2019 data. The source was from NBA.com and I analyzed splits based on 0-day vs 2-day rest. The sample size of the data was about 90 players sorted from the top 90 highest usage % players for the 2018-2019 season.
The output (or dependent) variables I examined were differences in points per game, three point attempts, free throw attempts, rebounds per game and overall plus/minus between 0-day rest games and 2-day rest games. My logic behind the three point attempts, free throw attempts and rebounds per game is simple: one can intuitively assume that if a player is rested, the player would settle for less outside shots, drive more and have more energy to collect boards. I used a simple absolute percentage difference (found here: https://www.calculatorsoup.com/calculators/algebra/percent-difference-calculator.php) and then applied a special twist: if 2-day rest > 0-day rest, the absolute percentage difference would be a positive output while if 2-day rest < 0-day rest, the absolute percentage difference would be a negative output.
The input (or independent) variables were as such: age, total games play, and individual dummy variables of big and prior major injuries. [For those unfamiliar with the term “dummy variable” in regression, it just insinuates, for instance, whether a player had a major injury or not] Noteworthy is how I defined major injuries: if a player had any form of surgery that forced that player to be out for greater than a month, the player was deemed as having a prior major injury.
The resulting regression was a bit anti-climatic insofar as the model didn’t show any trends. However, that is a learning lesson in-and-of itself. The p-values (exhibited in the visualization below) all exhibited alpha coefficients of above 5% which means that none of the independent variables were significant contributors to the player’s stats with and without rest (Note: as a reminder from some of the prior work, anything less than 5% is a significant variable to be considered, which means that the specific variable plays a role into the output). I do want to highlight two findings that came close to being “significant” variables:
The first is age and free throw attempts. The alpha coefficient for age in regards to free throw attempts was 8%, which is close to our 5% threshold. The coefficient was also found to be positive. This means that as players get older, the more free throw attempts we are likely to see on 2-day rest games than back-to-back. This makes some intuitive sense: the more rest that an older player gets, the likelier it is that the player is to be aggressive and drive.
The second is the dummy variable of big and three points attempted. The p-value here was shown as 6%, which is very close to our 5% threshold. The coefficient was shown to be negative. Directionally speaking, this means that a power forward or a center is likely to shoot and settle for more threes on back-to-back games than on games with 2-day rests. I don’t want to nitpick the data here, but a couple examples of this are Karl-Anthony Towns and Jaren Jackson Jr. who both averaged 0.5 more 3 point attempts on back to back games than games played after 2-days of rest.
However, in conclusion, none of these variables exhibited a p-value of less than 5%, so it’s safe to say that a player doesn’t play much differently on back to back games than they do on games with 2-day rests.
I hope you found this article to be insightful and I’m always happy to hear feedbacks and answer any questions. If you have any, please send me an email at DillonKang@gmail.com Otherwise, I’m excited for another NBA season to start in a month! Stay safe everyone!
The Baseline Jam 