The battle for the No. 1 position in ESPN Analytics’ NBA draft projections has never been closer, with Gonzaga’s Chet Holmgren and Duke’s Paolo Banchero virtually tied for the top spot.
Holmgren and Banchero both project to have a 51% chance at becoming an All-Star-caliber player — higher than Anthony Edwards‘ odds prior to his selection in the 2020 NBA draft but just short of Cade Cunningham‘s in 2021. Each forward is projected to have an average real plus-minus (RPM) of 0.96 in his fourth season, the first time in the model’s eight years that the top two prospects have had the same projection. Auburn’s Jabari Smith follows with a 48% chance to play at an All-Star level and an average projected RPM of 0.68.
ESPN’s NBA draft model forecasts the range of outcomes for players entering the league. A full breakdown of the methodology can be found below. The abridged version is that the final model is based on the combination of five individual models built from:
• Opponent-adjusted NCAA production
• League-adjusted international production
• ESPN draft expert rankings
• AAU/grassroots box score statistics
• Combine measurements
Bottom line: It uses a wide range of inputs to probabilistically predict future performance of prospects about to enter the NBA.
The top three prospects
Holmgren’s combination of height, length and impressive on-court production have had him at the top of our scouts’ rankings since early 2020. Standing at 7 feet with a reported 7-foot-5 wingspan, Holmgren might have ranked No. 1 overall by the combine-based model, had he attended. Among our top 100 players he ranks:
• No. 2 by the NCAA stats-based model
• No. 13 using AAU stats
• No. 2 in adjusted block percentage in college
• No. 8 in defensive win shares
• No. 2 in adjusted effective shooting percentage (eFG%)
That last ranking is especially impressive. Holmgren became the only 7-footer since we began our model to be in the top 15 in adjusted eFG% among our 100 prospects while also being in the top half of 3-point attempt rate. All of this puts Holmgren at the top of our projections with a 51.3% chance to play at an All-Star level, edging out Banchero at 51.2%.
Banchero’s on-court production and scoring have carried him to a virtual tie with Holmgren in the model’s projections, despite him going No. 3 in Jonathan Givony’s latest mock draft. Among our top 100, he ranks:
• No. 1 by the NCAA production-based model
• No. 1 by the model from AAU stats
• No. 3 in points produced, a measure of offensive contributions created by Dean Oliver
Banchero’s strong performances at events such as the Nike EYBL (AAU) in 2019 (he averaged 22 points on 51.5% shooting along with 2.5 blocks per game) and the NCAA tournament, during which he averaged 18.8 points, 7.6 rebounds and 3.4 assists while shooting 53% from 3 amid Duke’s Final Four run, have been impressive and consistent enough that the model views him as the best prospect in the draft before the expert ranking is factored in.
Smith ranks third overall in our model’s projections and has a 48% chance to play at an All-Star level, with a small gap separating him and our top two prospects. The gap is mainly explained by a slight drop in his adjusted college stats and an ineffective AAU showing. Among our top 100, he is in the top 12 in:
• Adjusted defensive rebounding rate
• Adjusted block percentage
• Defensive win shares
• Turnover percentage (TOV%)
Despite this, he only ranks No. 5 by our NCAA data-based model, trailing Banchero and Holmgren, plus Tennessee’s Kennedy Chandler (more on him below) and Duke’s Mark Williams. Smith’s shooting numbers from the National Basketball Players Association Top 100 camp (AAU) in 2019 were not impressive. He had an eFG% of 42%, which was below average as compared to his peers, and he ultimately ranked No. 69 of the 75 players for which we had AAU data. Smith’s No. 3 ranking is a combination of solid NCAA performance and a high ranking from Givony, who has him going No. 1 in his most recent mock draft.
Top 30 players
We view this as an above-average draft in terms of top to bottom strength.
Underrated
The 2021-22 All-NBA Team featured 14 first-round picks, 11 of whom were selected in the top 10 of their respective drafts. Two were selected each from picks Nos. 11 to 20 and Nos. 21 to 30. The Denver Nuggets‘ Nikola Jokic is the lone second-round pick. Odds are good that there will be a premier player selected outside of the top 10 in this draft. Among candidates to outperform their draft slot are Chandler and Duke’s Trevor Keels, both of whom have higher analytics grades than their current mock draft position.
Kennedy Chandler | PG | Tennessee
Chandler is No. 25 in Givony’s mock draft but is No. 11 in our projections. Chandler grades well because he is No. 3 in our NCAA production-based model and No. 2 in our AAU stats-based projections. He ranked No. 4, No. 2 and No. 7 respectively in adjusted assist rate, steal percentage and defensive rating among our top 100.
Trevor Keels | SF | Duke
Keels holds the No. 25 spot in our model’s rankings but is currently going No. 32 in the mock draft. Keels performs quite well in each of our categories, ranking No. 21, No. 9 and No. 6 in the combine measurements, NCAA stats and AAU stats, respectively. Some of his strengths include TOV%, steal percentage and foul rate, where he ranks in the top 15 in this class.
Honorable mentions: Mark Williams (No. 15 in mock draft vs. No. 10 in our model), Marquette’s Justin Lewis (No. 43 vs. No. 35), Louisiana Tech’s Kenneth Lofton Jr. (undrafted vs. No. 49)
Overrated
Ochai Agbaji | SF | Kansas
Agbaji is No. 14 in the latest mock draft. Our model projections have him ranked at No. 41. There are a number of factors at play here. In addition to the draft expert rankings, we have NCAA, AAU and combine information for Agbaji. He ranks No. 56, No. 62 and No. 29 by those respective models. His age (22) is taken into account in the NCAA stats-based model, which is part of the reason that he falls all the way to No. 56. His shooting, one of his main appeals to scouts, isn’t terrible, but he ranks No. 24 among our top 100 in true shooting percentage (TS%). In adjusted rebounding, assists, steals and blocks, however, he never cracks the top 40. Our model sees him as a clear second-round prospect, not a lottery pick.
Jalen Williams | SG | Santa Clara
In ESPN’s latest mock draft, Williams nearly makes it into the lottery, going No. 16 to the Atlanta Hawks. Our model sees him as the No. 26 best player in this draft, having a similar profile to that of Agbaji. Williams’ shooting is attractive to scouts, and he ranks No. 20 among our prospects in TS% and No. 9 in adjusted assist rate. He doesn’t, however, crack the top 50 in adjusted rebounding, steals, blocks or TOV%. Our NCAA stats-based model ranks him No. 35, with a No. 18 ranking from AAU data and No. 33 by combine measurements.
Honorable mentions: Gonzaga’s Andrew Nembhard (No. 31 vs. No. 67) and Wake Forest’s Jake LaRavia (No. 28 vs. No. 58)
International prospects
We don’t currently project any international players in the lottery. Ousmane Dieng, a small forward from France, is the top international prospect at No. 14 in our projections.
This is the lowest ranking we have for the top international player in our current model since 2017, when we had Frank Ntilikina (also from France) as the No. 14 prospect. Although Dieng’s average ranking is 14, there can be upside in selecting a premier international player. Since there have been fewer international players in the NBA than former college players, there is a lot more variance in projecting those players. Increased variance means that Dieng is 12th most likely to play like an All-Star, so perhaps taking a flier on him could pay off if you’re a late lottery team hoping for higher upside (Thunder, Hornets).
Rest of the rankings
Methodology
At ESPN Analytics, we combine our draft experts’ scouting grades with four other statistical models to get an average projection for each player over his first contract. We seek to project a predictive version of our RPM metric for each player at the end of his fourth season and obtain a probability that the player will reach the level of play of an All-Star, starter, bench player or non-NBA player in that time period.
The five individual models that are aggregated to achieve the final projections are based on NCAA production, AAU stats, international play, combine measurements and scout rankings. Some notes on the information used:
• Scout rankings are based on ESPN draft experts’ rankings.
• The NCAA and AAU models considered opponent-adjusted per possession box score statistics and composite statistics, such as individual rating and win shares.
• International statistics include similar box score metrics and adjust for the strength of the league.
• The combine model is based on body measurements, such as height, weight, wingspan and body fat percentage relative to position.
For each player, all relevant categories are considered if available.