It’s May and the 2019 NBA Draft is a little over a month away. Soon we will watch the future stars of tomorrow excitedly walk onto the stage and shake Commissioner Adam Silver’s hand. Of course, mixed in with the future stars will be many draftees who will simply never pan out.
The draft night is so intoxicating precisely because, on this night, we can’t distinguish the stars from the non-stars. Yes, by the time we reach 2029, the events that took place on draft night 2019 will be oh-so clear. Team X saw the potential in player A and got an amazing player at pick #8. Team Y thought they were getting a franchise changer at pick #2, but player B just never panned out. But on the the actual draft night of 2019, none of this will be known.
What we can do is establish some baseline expectations for how likely each pick is to actually work out, based on historical data. To this end, I collected data on all players drafted from 1995 to 2012. For each player, I recorded the number of All-Star game appearances and All-NBA teams made. Using this data, we can estimate the likelihood that a player will eventually make an All-Star game based on their draft position:
The black dots in the plot above represent the actual proportion of players from 1995-2012, at each draft slot, who made an All-Star game in their careers. The red line is the smoothed probability given from a logistic regression.
One of the striking things to me is that, even at the top of the draft, there is really no such thing as a sure thing. At the #1 overall pick, 12 of the 18 players in this time range made at least 1 All-Star game. But at #2, only 28% of draftees (5 of 18) actually reached this level. That means that only 47% of top 2 picks became All-Stars. There are roughly the same number of Greg Odens as there are Kevin Durants.
Another thing we can see from the graph is that star potential goes down very quickly after the first few selections. Based on the logistic regression, the #1 overall pick has a 64% chance of becoming an All-Star. By the time pick 5 comes around, that probability is more than halved to 30%. Pick #10 has roughly half the star potential of pick #5, at 16%.
The situation is even dicier from here. Historically, picks in the back half of the first round become All-Stars only 8% of the time. Hoping to snag the next Nikola Jokic in the 2nd round? Well, only about 3% of these 2nd rounders reached All-Stardom.
Of course, the All-Star threshold is not the only marker of player value. There are plenty of extremely useful players in the NBA who have never made an All-Star team. Ask the Rockets how important P.J. Tucker has been for them. Getting a player who will someday be a league-average starter is extremely valuable at, say, pick #15.
But the NBA is a league built around stars. The quickest path from watching the playoffs on the couch to advancing far enough to be beat by the Warriors is to acquire at least one franchise-changing contributor. Every team is hoping on draft night that they will snag an All-Star of tomorrow. But for selections made after the top 2 or 3, that is unlikely to happen.
For reference, here is a table of the estimated probability of making an All-Star team by draft pick: