Who was a better college basketball player: Donovan Mitchell or Malik Monk? As awesome as Mitchell was at times for Louisville last season, Monk was the co-SEC Player of the Year and was rated by most as a better draft prospect. The two were in a tight race up until the draft to prove who was the better combo guard, and it seems as though Monk ultimately won; he went 11th overall to the Charlotte Hornets, while Mitchell was drafted 13th by the Denver Nuggets and was infamously dealt to the Utah Jazz. So, it seems safe to say that Monk was seen as both a better college basketball player and a better NBA prospect than Mitchell before last year’s draft.

I was not of that opinion. Mitchell finished 7th on my final big board, while I ranked Monk 12th. Even though I got plenty wrong last season (my first year covering the draft), I was sold on Mitchell’s ability to continue to improve in the NBA. He improved exponentially between his freshman and sophomore seasons at Louisville, proving that he had the work ethic required to get better. Writers and executives raved about his maturity off the court. And perhaps most importantly, he checked every one of the boxes that Layne Vashro mentioned in a piece for Canis Hoopus that I often reference about which statistics indicate “potential”. Mitchell ranked second in the ACC in steal percentage, while posting an assist-to-turnover ratio of 1.66:1 despite often playing out of position at point guard and being asked to create tons of shots for an otherwise anemic Louisville offense (steals and assist-to-turnover ratio were the two stats that Vashro identified as indicators of understanding the game that offer potential for future growth outside of athleticism).

Later on in the piece, Vashro discusses how players with “great athleticism” (he uses no-step vertical to capture this idea best) who are also in the group of “understanding the game” (steals and assist-to-turnover ratio) improve at the fastest rate of all young prospects. Guess who won the standing vertical leap at the 2017 NBA Draft Combine? Pat yourself on the back if you guessed Donovan Mitchell. In many ways, he was the ideal prospect for projecting future improvement considering his off-court maturity, the statistical indication of his feel for the game, and his most traditional sign of “potential”; outlier athleticism.

As I (and I’m sure others) expected, Mitchell improved quickly over the offseason and has continued to improve at a rate faster than his peers during the season. The 21-year-old has not only upped his per-game scoring and assist averages while transitioning from Louisville to the NBA, but he has increased his shooting efficiency while doing so. Looking at just steals and assist-to-turnover ratio, you could have also found draft steals like Jae Crowder, Draymond Green, and Delon Wright while anticipating disappointments like Jimmer Fredette and Austin Rivers, and that’s before you even take history of improvement, off-court intangibles, athleticism, offensive rebounding (according to research I have done, it’s a great indicator for non-bigs), and Defensive Box Plus-Minus (the best indicator for bigs) into account.

While I want to believe in the aforementioned system I have developed for projecting improvement, it’s important to remember that in many cases team situation is the ultimate factor in deciding a player’s outcome. Jaylen Brown and Jayson Tatum have unquestionably benefitted from the infrastructure of the Boston Celtics organization, and that has likely been the biggest factor in each of those players’ rapid ascent. Meanwhile, Victor Oladipo bounced through a tumultuous four seasons between Orlando and Oklahoma City before finding his home in Indiana and exploding into an All-Star talent. Aside from a subpar assist-to-turnover ratio, the indicators for Oladipo’s ability to improve were off the charts. However, being played out of position in Orlando and being forced into a less-than-ideal role in Oklahoma City inhibited his growth before he finally landed in a place set to maximize his skills.

Judging players independent of team context is almost impossible. But in a vacuum, it is possible to use proven methods to assess which players are likely to improve in the NBA. It is up to individual front offices to know which players fit in with their teams, and outside of constructing team-specific draft boards, it is hard to account for certain fits when ranking players. Donovan Mitchell’s success is unquestionably due in part to his seamless fit in Utah. On the flip side, Jawun Evans, a player who profiled as likely to improve, has languished during his rookie season with the Clippers. As a player who needs the ball in his hands, Evans is a difficult fit on a Clippers squad loaded with ball handlers.

How does this apply to the 2018 draft? First, noting which players are likely to improve can help us identify meaningful differences between players who are otherwise close evaluation-wise. For example, in The Stepien’s updated rankings, Robert Williams is ranked 13th and Jontay Porter is ranked 15th, both in the composite 5th tier. It’s safe to say that the two are very close in how they are perceived. Looking deeper, it’s easy to tell which is more likely to improve. Despite the fact that Williams would almost surely out-jump Porter in a no-step vertical, Porter ousts him when taking everything else into account. Per 100 possessions, Porter picks up 2.4 steals, while Williams comes up with just 1.8. Williams has a cringe-worthy 0.68:1 assist-to-turnover ratio as compared to Porter’s modest 1.05:1 assist-to-turnover ratio.

Williams has not made notable improvements in any areas of his game between his freshman and sophomore seasons, whereas Porter has gone from playing against class of 2018 AAU players as recently as last summer to ranking 4th in the SEC in BPM. It is fair to say that his performance at this level was unexpected. Vashro’s article mentions that advantages in athleticism only matter in projecting improvement when a player also has advantages in terms of steals and assist-to-turnover ratio, effectively nullifying Williams’ athleticism advantage. One other important factor to consider – because Williams and Porter both project as bigs in the NBA) – is DBPM. Research I have done has shown that it is the most important indicator for future success for big men. Williams’ 9.4 DBPM doesn’t really differentiate him from Porter, who is at 8.1 this season, and Porter’s number this year is better than Williams’ 7.2 as a freshman. Through these various lenses, it is difficult to say that Williams projects to improve at a quicker rate than Porter, who also happens to be more than 2 years younger than his Aggie counterpart.

The Porter/Williams debate is one that will surely continue as the draft approaches, but there are many other players who stand out when looking for signs of improvement. Penn State’s Josh Reaves jumps off the page most when looking at the key indicators for future growth. With a 2.95:1 assist-to-turnover ratio, an average of 4.6 steals per 100 possessions, and a robust 7.1 offensive rebounding percentage, Reaves’ numbers this season are highly convincing. His year-over-year improvement is another reason for hope; from his freshman season to now, his true shooting percentage has risen from .425 to .637, his BPM has gone from 3.6 to 12.9, and his WS/40 has exploded from .062 to .234. Oh, and he’s athletic (there are two dunks in the video):

Now, I have to temper expectations. If Reaves can’t sustain competent three-point shooting (he’s currently at 42.3%, but he’s a 29.3% career shooter), it’s hard to imagine him sticking long in the NBA. But all the indicators for him are strongly positive, and he is the type of player that I would take a bet on in the late first round, especially because at that point in the draft, he’s likely to land on a team where his do-it-all role player skill set fits well.

Kenrich Williams, Gary Clark, and Troy Brown also pop when reviewing the indicators that signal “non-traditional upside”. None of those guys are particularly athletic, but they all create for others while limiting turnovers, they all play defense (and generate steals!), and they all rebound particularly well for their positions. Clark and Williams are on the older side, but it’s foolish to think that even a 23-year-old player is done progressing.

After reviewing the standouts, who is on the flip side? Kevin Knox, for all his awesome floaters, looks awful when measuring up to these standards. His assist-to-turnover ratio is just 0.62:1, he generates just 1.2 steals per 40 minutes, and his offensive rebounding percentage is a minuscule 3.5. Brandon McCoy has over five times as many turnovers as assists, he produces just 1.1 steals per 100 possessions, and his 0.8 DBPM is a cherry on top (I do not think McCoy is an NBA player, at least at this stage in his development). Most importantly, let’s look at the top bigs in this class. Mohamed Bamba’s A:TO ratio is at 0.21:1, and he picks up a modest 1.7 steals per 100. Luckily for him, his DBPM of 10.3 seriously bolsters his profile. Deandre Ayton is at 0.8:1 and 0.9 per 100, respectively, paired with a leaves-you-wanting-more 3.8 DBPM. Marvin Bagley III disappoints at 0.67:1 and 1.7 per 100, with a particularly concerning 2.9 DBPM. Jaren Jackson Jr. sits at 0.58:1 and 1.9 per 100, but his 11.2 DBPM extinguishes most questions about his projection. For context on these stats, 0.4:1 seems to be the critical point for A:TO ratio for bigs. Since 2011, only Andre Drummond stands out as a successful player from the group of players who have posted an A:TO ratio that low or lower during his last year in college. I think that discussions about Bamba, Ayton, and Bagley III need to include conversations about these indicators, because each player is seriously flawed in at least one key area.

It’s important to note that all of the players mentioned above are freshmen and that the players I mentioned as standouts (with the exception of Troy Brown – he’s good) have all spent at least three years in college. Still, these are important numbers to consider. In regards to judging the top bigs, I trust DBPM more as an indicator for future success for bigs than A:TO ratio and steals, but I think that it is important to consider those numbers when picking apart these prospects; particularly in the case of a guy like Bamba, whose A:TO ratio is a very real issue. The importance of DBPM for bigs is why I have qualms with labeling Ayton and Bagley III as top prospects on the same level as Luka Doncic, Jaren Jackson Jr., and Trae Young.

Predicting player improvement is just about as much of a crapshoot as the draft itself. I myself have done plenty of analysis trying to figure out what I (and many others) missed with Kyle Kuzma, but I have come to realize that most of what he has done can simply be chalked up to unexpected improvement. Without access to Kuzma and his training, there was no way to tell that this would happen. As a draft writer who doesn’t get to have the same intel as NBA teams, there is only so much I can rely on to find players who are likely to get better. For now, I will trust in assist-to-turnover ratio combined with steals, past improvement, the little info I have about off-court maturity, “athleticism”, offensive rebounding for guards and wings, and DBPM for bigs.