In ESPN's latest mock draft, one huge question is how the teams at the top of the lottery see prospects such as LaMelo Ball, Anthony Edwards and James Wiseman.
Meanwhile, in the ESPN ranking of the top 100 draft prospects, Ball is first and Edwards second, based on the consensus from around the NBA.
Who is No. 1 according to the ESPN Sports Analytics draft projections? And how do we rank the top 100 prospects?
Let's break down the top picks, the top international prospects, the most overrated and the most underrated.
How does the model work?
The ESPN Sports Analytics draft projections model considers information from five categories before forecasting the range of possibilities for each prospect's NBA career:
Opponent-adjusted NCAA production
League-adjusted international production
ESPN draft expert rankings
AAU/grassroots box score statistics
Combine measurements, such as height, weight, wingspan and body fat (excluded this season because of the lack of a traditional combine)
For every player, the relevant categories are considered. (Only some players have an international track record, for example.)
We updated the model this year to estimate a predictive version of our Real Plus-Minus (RPM) metric.
The model projects both the average RPM by the end of a prospect's rookie contract (fourth season) and his chance to play at the level of an All-Star, starter, bench player or non-NBA player in that time.
The No. 1 prospect
The ESPN Sports Analytics model prefers Edwards to Ball but also views Edwards as the weakest top prospect in the model's data set (since 2001), a far cry from last year's No. 1 pick, Zion Williamson, who was the top projected prospect over the past 10 years.
Edwards has a 41% chance to play at an All-Star level by his fourth NBA season, which bests Ball's 28%.
Why does the model prefer Edwards to Ball?
Edwards put up great numbers in the Under Armour amateur events and in the NBAPA Top 100 Camp in 2018, making him the top prospect based on AAU-level events. Ball, on the other hand, had limited exposure in the AAU circuit before he went overseas to play professionally.
Edwards' numbers in college were not earth-shattering. In fact, Iowa State's Tyrese Haliburton -- not Edwards -- projects highest based on opponent-adjusted college production. (Haliburton is also second in Kevin Pelton's projections.)
Based on those metrics, Edwards is most similar to Ronnie Brewer, Donovan Mitchell and Randy Foye. That's a pretty wide range of outcomes and a sign of why predicting a prospect's NBA success is difficult: Edwards could be the next Mitchell, but he also could turn out to be another Foye.
In the model's eyes, Ball has to overcome having limited data to compare to that of other prospects. Despite a stellar assessment from scouts, Ball is a riskier prospect, as are most draftees from overseas leagues. Despite the success of players such as Luka Doncic, there simply have not been enough prospects from each overseas league to be confident in their projection.
By comparison, top NCAA players such as Edwards and LaMelo's older brother, Lonzo Ball, the top projected prospect in 2017, were able to play 30-plus games against collegiate competition, giving us a bigger sample of past prospects to compare them to.
Top 30 players
This is not a strong draft by the numbers.
In fact, we project this draft to have only 2.7 All-Star-caliber players on average, the worst mark in our database. The second-worst such projections belong to 2016 (when Ben Simmons was the top pick), which had only 3.2 projected All-Star-caliber players.
Top international prospects
Deni Avdija, SF (Israel) & Killian Hayes, PG (France)
The model and the scouts agree on Avdija and Hayes as the top international prospects in this draft. While the model hedges on Ball's ranking due to his limited minutes in the NBL, Hayes and Avdija are the same age and have more professional experience, both against top European competition.
Avdija does not have a great outside shot but still had an 79% effective field goal percentage in the most recent Euroleague campaign, per Basketball Reference. At 6-foot-9, he rates as one of the top shot-blockers and rebounders, which are among the most predictive stats for any prospect.
Hayes would have the top projection based on international stats. After our adjustments for the varying levels of international competition, Hayes is above average for an NBA prospect in every advanced metric considered by the model, but he excels most at creating assist opportunities for his teammates. He averaged 6.3 assists per 36 minutes in the French League and Eurocup.
Is the top college player overrated?
Obi Toppin
Toppin won the Wooden Award as the most outstanding collegiate player last season, yet his projected Real Plus-Minus is 22nd. Toppin has a pretty high ceiling by our metrics, as he is seventh in All-Star percentage (how likely he is to play at an All-Star level), which might justify a team using a lottery pick.
Ultimately, the main issue for Toppin isn't his game -- it's his age. He's 22, and most of the other top prospects are under 20. If you didn't enter the NBA draft by age 20, it is likely because you weren't good enough. Toppin was a late bloomer who made a big leap his last year at Dayton.
Why is being a late bloomer a problem? Well, on average, NBA players improve their predicted RPM by 1.8 in their age 19 to 21 seasons, but Toppin has already played those seasons in college. From age 22 to 25, a player's predicted RPM improves by 1.1 on average. Toppin might be a great player, but he is closer to his peak than the other draftees.
Who are the underrated prospects?
Our system ranks several prospects higher than the scouting consensus, including these four:
Vernon Carey Jr.
You would not expect the NBA to overlook the best player at Duke, yet here we are. He's no Zion, but Carey was the most efficient player in college last season, per our opponent-adjusted win shares per 40 minutes.
Devin Vassell
The most similar player to Vassell, according to our metrics? Tyler Herro, who made a surprisingly successful NBA debut. Don't be surprised if Vassell follows.
Jalen Smith
Smith's 13 rebounds and three blocks per 40 minutes are eye-catching. Basketball-Reference estimates that he contributed 9.9 win shares in a shortened season for Maryland.
Kira Lewis Jr.
When he was on the court last season, 28% of his teammates' field goals came from his assists, which is in the top 5% in opponent-adjusted assist rate. His projection is hindered by his outside shooting, but if he isn't taken until the back half of the first round, he could be a major contributor and distributor for a playoff team.
ESPN NBA Draft Model Accuracy
