NBA Draft Analysis
1. Introduction
Within the National Basketball Association, the draft has always been seen as an off-season event where teams may make or break their upcoming season. Highly anticipated by organizations, fans and most importantly the elite up and coming players from around the globe, the NBA Draft is an event where all 30 teams in the most elite league in the world select 2 (more or less depending on if their pick has been traded away prior to the draft) highly regarded players from around the world to join their roster for the upcoming NBA season.
The NBA draft is designed to provide a leverage to the teams with the lowest winning record in the league from the previous season. The lower the team’s ranking in the season prior, the higher chance they will have at attaining the number 1 pick in the draft that following off-season. Looking through the last 32 drafts, we take a look at all the players picked from the drafts between 1989 and 2021 and their NBA career stats as long as draft day variables to analyze a variety of different relationships within the data set that’s been chosen for this assignment.
2. Biggest Busts in the Draft (Picks 1–3)
Achieving a top 3 pick in the NBA draft is widely considered a successful draft lottery outcome for almost any team. This is due to the fact that at the very minimum you are almost always getting a valuable asset to join your roster, if not a future star of the league, then at the very least a young rotational player that may be moved for more valuable assets. However, this isn’t always the case. There are instances when a top 3 pick doesn’t turn out as planned and completely misses the mark, with the player putting up extremely underwhelming career numbers in the NBA. These players are commonly referred to as ‘Busts’ by the fans.
Using the following analysis, we identify the standout negative outliers, or busts, in the collection of all the top 3 NBA draft picks from seasons 1989-2021. Scatter plot below shows us the two players who are significantly below the rest of the group when it comes to career points per game. Hasheem Thabeet (2nd Overall Pick) and Anthony Bennet (1st Overall Pick) are the only two top 3 picks in the history of the league to put up under 5.00 points per game across their NBA career and are considered the two single biggest busts in the history of the NBA to date.
3. Biggest Steals in the Draft (Picks 45–60)
On the other hand, the dynamic of draft picks changes as we delve into the late picks of the draft. Last quarter of players picked in the draft tend to have minimal expectations and are considered to be subpar and at the very most, teams hope for them to develop into a useful rotation option for the future of the team, or an asset used as a part of a deal to bring in a more prominent player into the team.
The graph below looks at all the players who were selected in the last 15 picks of the draft (45-60). The bigger range is used here as teams have historically sometimes not selected a player at all within the final few picks of the draft, so the range was enhanced to accustom for this tendency.
The analysis singles out some of the standouts in the visual such as the San Antonio Spurs legend Manu Ginobili, who has managed to win a 6th man of the year trophy in the 2007/08 NBA season, and Isaiah Thomas, who was the last overall pick in the 2011 and was the franchise player for the Boston Celtics in the mid 2010s which included an Eastern Conference Final run that saw the Celtics miss the finals at the hands of LeBron James’ Cleveland Cavaliers.
4. Best Producing Colleges
The next analysis will take a look at the best producing colleges over the last 32 years. Looking at all the college data available we take a look at which college has had the most players drafted overall and furthermore split the results into 1st and 2nd round to evaluate which colleges actually have the most 1st round picks as the first rounders tend to have more influential roles in the league throughout their career.
For this analysis we use a clustered bar chart which can tell us about the highest producing colleges over the 32-year period as well as splitting them into 1st and 2nd round picks to get a better understanding of the quality of the players they send to the NBA
The 2 juggernauts in this analysis came out to be Duke and Kentucky as the 2 highest drafting colleges in the 32-year period, only separated by 1 total pick where Kentucky has the bragging rights.
5. Statistical Differences Between Draft Rounds
Next up we want to distinguish a historical difference between a first and second round pick. Every team owns a single 1st and a single 2nd round pick for each draft from up to 5 years into the future. These rights they hold for these picks are able to be given away for an active player in return or may be used to sweeten up a trade deal offer for a team if the other party feels the value of the offer isn’t sufficient.
Therefore, to differentiate the value each pick holds as teams might want to be able to hold on to a certain pick if they believe that the value doesn’t meet the sufficient standard in a trade offer.
In the following figure, we compare 1st round and 2nd round picks among 3 most popular basketball statistics a player outputs in a single game: Points per game, Rebounds per game and Assists per game using box plots. The data shows a consistent difference between 1st and 2nd round difference that justifies the value the 1st round pick holds over the 2nd round pick.
6. Career Longevity vs Average Minutes Played
Something the data provided in this dataset allows us to do is assess how different variables are tied to career longevity in the NBA. The 3 different variables assessed for this analysis will be average minutes played, points per game and value over replacement.
One aspect that must be mentioned in this section is the missing data of active vs retired players, as some players who have only recently joined the league and have performed exceptionally might be assessed under the same bracket as some retired players who haven’t had many seasons in the NBA due to subpar output. This might lead the results to be misleading and sets grounds for improvement in the future of data collecting.
Average minutes played is used mainly as a control variable here just to maintain the idea that staying in the NBA is directly correlated to how many minutes a game each player gets.
7. Career Longevity vs Points Per Game
Points per game graph shows a positive correlation between a player’s points output and the number of years they spend in the league.
8. Career Longevity vs Box +/-
Penultimately, the correlation between box plus minus and career longevity showed a positive correlation and shows that box plus minus is a great indicator to convey player value as its consistent with player longevity within the league.
9. Career Longevity vs Value Over Replacement (VORP)
Finally, the value over replacement’s (VORP) correlation with career longevity is consistent as well. Using a bar chart, we can see that players with a higher value over replacement tend to hang around the NBA longer, proving that VORP is a great indicator for predicting a player’s longevity in the league.
10. Career Longevity by Draft Round
Additionally, to put into perspective, teams tend to be aware of player quality very well come time of the draft. This can be seen clearly through the visual where we compare the two draft rounds and how long players’ career last depending on which round they were drafted in.
Being drafted in the first round you can be fairly confident to get a good 5 years out of the player’s active status in the league. Whereas, the same can’t be confidently said for second round picks according to this box plot.
11. Volume Scoring vs Shooting Efficiency
The data set provides a plethora of shooting data for each player in the last 32 drafts. A player that scores a high number of points per game is undoubtedly a great scorer however, are they at the same time an efficient scorer?
To assess this, we create a scatter plot that includes all 3 statistics color coded to fit the graph. Due to all 3 statistics not having a consistent range of measurement, each variable is adjusted for to a 0-1 scale for better understanding.
Correlation is shown mostly through 3-point percentage and free throw percentage. Field goal percentage doesn’t show any significant correlation for the relationship between scoring output and efficiency, however, 3-point shooting and free throw shooting percentages do. Meaning those two variables are directly tied to the scoring ability a player possesses.
Additionally, one fascinating outlier, as marked on the plot, is Eddy Curry. Who is shown to have a seemingly average career point output however holds 100% efficiency from 3-point shots. This does not unfortunately mean Eddy Curry was the most elite shooter the world of basketball has ever seen, this however is the result of him only attempting two 3-pointers throughout his entire career and of course making them both. Still a pretty impressive feat for a Center, which is a position who traditionally don’t shoot threes.
12. Greatest 2nd Round Picks of All Time
The second round of the NBA draft is not where players of stardom level tend to be selected at, however there have been instances where exceptional players have come from being drafted in the second round. Overlooked by every team, these players proved the doubters wrong by elevating to elite level when no organizational body or coach believed they could do so.
The table below takes all the players who have been selected in the 2nd round of their respective draft, and simultaneously have over 100 career games played, sorted and highlighted by the career points per game statistic, which is the most important key performance indicator showing the quality and impact a player has had over their career.
2 things this table doesn’t account for is the player accolades each player might have achieved over the course of their career, as well as doesn’t signify which players are still active and which are retired. Looking at the graph, one would assume that Gilbert Arenas is the greatest second rounder of all time, however, with additional context that Nikola Jokić, who is second on the list, is a one time NBA Champion and a three time MVP of the league, we’d look at the ranking differently as Gilbert Arenas hasn’t managed to achieve either of those accolades.
The reason Nikola Jokić’s statistics are lower are due to him still being active in the NBA meaning his averages are still held back by his early seasons when he was still a developing role player.
| Best NBA 2nd Round Picks | |||||||||
| Sorted by Career Points Per Game (min. 100 games) | |||||||||
| Player | Draft Team | Draft Year | Pick | PPG | RPG | APG | Games | FG% | Win Shares |
|---|---|---|---|---|---|---|---|---|---|
| Gilbert Arenas | GSW | 2001 | 31 | 20.7 | 3.9 | 5.3 | 552 | 42.1% | 51.3 |
| Nikola Jokić | DEN | 2014 | 41 | 19.7 | 10.4 | 6.2 | 527 | 54.2% | 79.6 |
| Michael Redd | MIL | 2000 | 43 | 19.0 | 3.8 | 2.1 | 629 | 44.7% | 55.9 |
| Monta Ellis | GSW | 2005 | 40 | 17.8 | 3.5 | 4.6 | 833 | 45.1% | 41.9 |
| Isaiah Thomas | SAC | 2011 | 60 | 17.7 | 2.4 | 4.8 | 550 | 43.4% | 45.3 |
| Khris Middleton | DET | 2012 | 39 | 17.1 | 4.9 | 3.8 | 651 | 45.8% | 50.6 |
| Dino Radja | BOS | 1989 | 40 | 16.7 | 8.4 | 1.6 | 224 | 49.7% | 14.3 |
| Carlos Boozer | CLE | 2002 | 35 | 16.2 | 9.5 | 2.2 | 861 | 52.1% | 80.3 |
| Cuttino Mobley | HOU | 1998 | 41 | 16.0 | 3.9 | 2.7 | 747 | 43.3% | 52.2 |
| Malcolm Brogdon | MIL | 2016 | 36 | 15.5 | 4.2 | 4.8 | 333 | 46.4% | 24.2 |
| Jordan Clarkson | WAS | 2014 | 46 | 15.4 | 3.3 | 2.5 | 600 | 43.9% | 21.7 |
| Stephen Jackson | PHO | 1997 | 42 | 15.1 | 3.9 | 3.1 | 858 | 41.4% | 36.1 |
| Bojan Bogdanović | MIA | 2011 | 31 | 15.0 | 3.6 | 1.6 | 603 | 45.7% | 33.0 |
| Rashard Lewis | SEA | 1998 | 32 | 14.9 | 5.2 | 1.7 | 1049 | 45.2% | 90.9 |
| Dillon Brooks | HOU | 2017 | 45 | 14.5 | 3.0 | 2.0 | 272 | 42.1% | 5.3 |
13. Evolution of NBA Players Over the Years
The last analysis that will be conducted is the evolution of NBA’s 3 most significant statistics, how they’ve evolved over generations and what this tells us about the players of the game of basketball.
We used points per game, rebounds per game and assists per game as the variables to see how the league has evolved over the years. This takes into account the year a player was drafted not the season in the NBA, which is also why we use statistics only up to 2015 as data for players who were drafted in the 2020s would portray only early career statistics compared to previous generations where the whole career is taken into context for our understanding.
Players are also split into different bins spread by 5 years, as significant changes may only be visible over a multiyear period as opposed to just one.
The graph shows a few inconsistencies; however, the most certain thing is that all statistics peak or just about peak in the most recent generations of players. This is due to the fact that over the course of decades, an average NBA game contains a lot more possessions than it previously would have, as the game is speeding up by the day, so are possessions.
Naturally, higher scoring games will cause an increase to points per game, assists per game but also rebounds per game. As there is more possessions in a single NBA game that means more missed shots as well which results in the increase in rebounds per game as well.
