Nylon Calculus: Projecting the top 23 and under NBA players

The NBA is filled with great, young talent under 23 years old. Recently, Dunc’d On completed their annual top-10 NBA prospect podcast which led to many variations being posted on Twitter. Using those lists as my basis, I used my player projection model to find the top 23 and under NBA prospects.

The projections are built around my player projection model, the same one I use to project season win totals during the offseason. The played model I developed compares a slightly modified and regressed version of what a player has done over their previous three seasons to a database of historical season. For rookies, the model regresses current season performance to how the player projected out of college or overseas. The model finds similar players statistically/stylistically and uses how those players’ games changed as they aged as a crutch to augment a generic aging curve to better tailor it to the specific player’s style of play and skill set.

Read More: The Thunder are who we thought they were

A disclaimer before getting any further: the model is attempting to project events that are four to nine years from now depending on the player. The expected error on these is, rightly, pretty large. The younger a player is, the more error in the model. A rookie being ranked better or worse than you personally think is not an indictment on the player.

The way I decided to rank under-23 prospects was by projected Wins Added from age 26 to 28, generally a player’s three most prolific seasons. Wins Added is my cumulative impact statistic, comparable to RPM wins or VORP. For reference, the average starter provides about 5.0 Wins Added in value each season.

Terminology

There are seven columns on the charts below and here is what they all mean:

• Rank – Where the player ranked by peak Wins Added
• Player – The player’s name
• Age – Player’s age as of today, I used the Basketball-Reference method for what age season this is for a player
• Peak %ile – Where the player’s projected 3-years Wins Added peak ranks among all players in my database (going back to 1973-74)
• Wins Added – The numbers of wins a player is projected to add during their age 26 to 28 seasons
• Minutes – A player’s projected minutes during their age 26 to 28 seasons
• WA/48 – Wins Added per 48 minutes played, a conversion of Wins Added to a per minute rate to compare players with different minutes projections

To help you read WA/48, keep in mind that an average player provides 0.500 WA/48 and a replacement level player provides 0.426 WA/48.

A link to the entire list along with notes on each player’s projection and what the model might be seeing or missing is available at the bottom of the page.

Without further ado, let’s get to the list!

23-and-under prospects: 41 to 50

The most surprising part of this group was seeing Juan Hernangomez, Domantas Sabonis, and Julius Randle come in this low. Each projects to provide above average production on a per minute rate, but their low minutes projections pull them down to the bottom of the list. Randle getting out off the Lakers would drastically improve his ranking. Hernangomez continues to be played next to no minutes on the Nuggets. Sabonis is certainly playing more for the Pacers so I expect his projection to rise as the year goes on.

Outside of those three, the most surprising name is Kyle Kuzma. Kuzma is certainly held back by a relatively lower college projection, but the largest issue right now is defense. By my metric, he is one of the worst defensive players to make this list. If he is able to improve on that end while maintaining his offensive skill, look for him to move way up on this list.

23-and-under prospects: 31 to 40

The most interesting part of this group is max player Andrew Wiggins coming in 34th overall. While Wiggins has never been loved by advanced stats, he has usually rated slightly less bad by my metric. This season, that has not been the case. Despite being asked to do less, Wiggins is still settling for the same poor long mid-range shots and his 3-point accuracy has dropped off. In a film study I did on Wiggins, I found that he performed much better when the offensive system created a shot for him than when he was asked to do it on his own.

Outside of Justise Winslow, all of the players in the group project to above average peak production. Frank Ntilikina is the most intriguing. While he does not project to be an elite player individually, his defensive prowess and 3-point shooting make him a great compliment for Knicks star Kristaps Pozingis.

23-and-under prospects: 21 to 30

This group is the first to include projected starting level players based off Wins Added. Jaylen Brown currently holds the largest single year increase in my value metric from last season. Brown is showing the 3-and-D prowess that made him a high draft pick last season. Also interesting is how well Mailk Monk rated out in the projections. While the model never sees him becoming anywhere close to a defensive player, it really believes in his offense as the second coming of J.R. Smith (who was Monk’s closest statistical comp).

Brandon Ingram ranking this low may be disappointing to some Laker’s fans, but even still he has made quite a jump from where he would have ranked last year. The Lakers do Ingram no favors by asking him to create so much offense for himself. He is solid at it, but the best way to use Ingram right now is as a complimentary glue type player similar to Nicolas Batum.

23-and-under prospects: 11 to 20

This group is the start of the future stars of the league. John Collins is likely the only player in this section who I think is being overrated by the model, with his pick-and-role defensive issues being hidden by the data I use. Devin Booker is a player likely to improve in his projection as the season goes on. Booker is having the best year of his career, and that becomes a larger portion of his projection sample will move him up this list.

Another guy likely to move up is Donovan Mitchell. Mitchell struggled early in the season, but has been absolutely crushing it for the last month and a half. I fully expect Mitchell to move into the top-10 by the end of the season.

It’s important to note that Markelle Fultz rates this well likely because he has played so little, if that makes sense. For rookies, I blended their projected production from college or overseas with their actual production to help stabilize hot or cold starts for projecting them forwards. Because Fultz has played so little this year, he has a much higher weighting on his projected production than his actual production when compared to other prospects. Once Fultz is back and healthy, I expect his projection to drop to the lower end of this group.

23-and-under prospects: 1 to 10

Giannis is very good at basketball and his projected Wins Added peak of 51.5 would be the 12th highest all-time total in my database. Jayson Tatum is likely overrated here despite blending college production with rookie production. Do not get me wrong, Tatum is an A+ prospect, but he probably should be seventh on the list with everyone else moving up a spot. I expect the data to show that as the season goes on.

Clint Capela is quietly one of the best young players in the league. He is not a flashy upside prospect, but he is an elite rim-protector and roll man who is playing with two of the best pick-and-roll guards in the league.

Joel Embiid is hurt by a low minutes projection here, but I think the 76ers would trade Embiid playing relatively lower regular season minutes for his health during the many playoff runs that are sure to come.

Next: Nylon Calculus -- Kyle Lowry is an underrated star leading an underrated team

Myles Turner is the forgotten unicorn. Gary Harris is going to be perennially underrated by fans. Towns and Jokic and Simmons are all young monsters. The NBA is in very good hands.

A place for improvement in this exercise would be to blend this ranking list with a similar list made by a scout. With draft models, the best success is found by blending in a scout’s perspective to account for things that the data might miss. I believe that this process could also create an improved statistical ranking of the best under-23 prospects in the league.

Here is a complete list of players along with notes on each player’s projection