For the first time in nearly a decade and a half, we’ve reached the All-Star break without Timberwolves fans obsessing over the draft. With the team in great position for a spot in the playoffs, and even favored to finish with home court advantage, the question of who to add to the team in June is much less urgent. So long as the team does not pick another center, most fans should be happy.
Habits, however, are hard to break. And Minnesota does have a pick this year, from Oklahoma City via Utah, that is projected to fall in the early 20s. And I have revamped SMILODON, the intuitive but innovative draft projection system I developed last year, so I’m not going to let a little thing like “the team being good” stop me from writing about the draft. Let me begin with a re-introduction to the method of my madness.
SMILODON splits players into three categories: guards, forwards, and bigs. I’ll explain how each category (now) works, using a series of examples.
Unlike last year, when point guards and shooting guards were evaluated differently, this year, all guards are evaluated on the same scale, from “Poor” (Red) to “Elite” (Blue), under the same rubrics in five separate categories: shooting, driving, passing, defense, and age.
Shooting is determined by a formula that takes into account free throw percentage, number of three pointers made (per 40), and jump shot percentage, with the first two factors weighted more heavily.
Driving is determined by estimating the player’s unassisted makes at the rim in a halfcourt setting, excluding offensive rebounds. Last year, the system did not exclude transition baskets, but I’ve been persuaded that was a mistake.
Passing is determined by an adjusted assist to turnover rate. First, turnovers are adjusted for usage, so high usage players will see their turnovers reduced, while low usage players will see their turnovers increase. Next, the adjusted assist to turnover rate is adjusted for height, which can make a big difference in a player’s ability to see passing lanes in the NBA. This also has the effect of slightly dampening the projections for small point guards, a player archetype that tends to under-perform its SMILODON in the NBA.
Defense is determined by a player’s steal, block, and defensive rebounding rate, with steal rate receiving the highest weighting. Compared to last year, blocks have been de-emphasized. I think of shooting and passing as measuring a player’s skill, while driving and defense are good athleticism measurements. Gaudy defensive numbers may also indicate untapped potential.
Age is perhaps the most important part of the system, as young players have much more room for improvement than older players.
Here is how this looks, using some of last year’s top prospects.
Donovan Mitchell profiled as a high level 3&D guard - the perfect #4 option on a championship contender. This is actually the best projection of any of the last six #13 selections, and a profile that would usually be indicative of a top ten pick. The positive aspects of this projection were mostly driven by Mitchell’s free throw percentage, three point volume, and high number of steals. SMILODON was optimistic about his shooting and his good defensive numbers indicated a very good athleticism & awareness combination; however, the system was less sanguine about his handle and vision being that of a lead guard. Of course, it seems that Mitchell will make this projection look overly pessimistic.
Now, here is Lonzo Ball.
As anyone could tell you, Ball profiled as a great passer and posted solid defensive numbers, falling just short of a green flag there after accounting for UCLA’s fast pace. The system did, however, identify that Ball may have problems scoring, with subpar shooting and driving indicators. That orange flag for shooting is more heavily influenced by Ball’s poor free throw percentage than his stellar three point percentage. Indeed, SMILODON’s general disdain for field goal percentage gave Ball a lower projection than in most statistical systems. Lonzo was a unique prospect. His ability to shoot a little separated him from players like Michael Carter-Williams, while his height separated him from smaller point guards with statistical similarities, like Tyler Ennis.
Who was a better prospect, Ball or Mitchell? SMILODON is not an “all in one” projection system, so there is not a definitive answer to that question. I think of a blue and a yellow flag as being roughly equivalent to two green flags, but it is certainly arguable that shooting is a more important category than passing. In short, it’s subjective when projections are close, and certainly open to mental adjustment from individual scouting impressions - which I view as a plus.
What does a typical second round pick look like? Here are Davon Reed and Frank Mason, both picked in the 30s.
Each player has a red flag, which indicates a serious hole in his game. Reed’s line is more typical of 3&D prospects than Mitchell’s, while Mason is portrayed as highly skilled, but less athletic. As older players with only one clear strength, each profiles as a backup.
Now, here are some of this year’s top guard prospects (as of February 11th).
Trae Young has the best statistical profile of any guard of the past seven drafts. For purpose of comparison, here is how that compares to a few other smaller guards taken in the top ten.
When you compare favorably to Damian Lillard, that’s very impressive. Meanwhile, Collin Sexton shows the ability to get to the rim, which means that he’ll probably be able to score in the NBA, but has not yet shown the skill or defensive impact that would suggest eventual stardom.
Lonnie Walker IV is a player who SMILODON is cautiously optimistic about; he is young, can shoot a little, is putting up decent defensive numbers, and, subjectively, is very athletic. He looks like a fine choice in the teens.
Hamidou Diallo, on the other hand, has probably the most pessimistic projection of this year’s first round guards. The shot isn’t there, the lead guard skills aren’t there, and the athleticism isn’t even translating to the defensive end. Kentucky has a history of suppressing statistical output, but not by that much.
The system to evaluate forwards is very similar to the one described above. What follows are the (small) differences between the two.
In the “shooting” category, forwards have a higher emphasis placed on free throw percentage. There is also a different, far more forgiving, formula used to calculate the “passing” category. Finally, forwards have a small team component built into their “defense” formula.
Here is Jayson Tatum.
Don’t let Trae Young’s profile throw you off - this is what it usually looks like when SMILODON loves a player. Tatum had the best projection of any 2017 prospect. The combination of green flags for shooting and age is usually indicative of a special offensive player, and Tatum does not show any weaknesses by this method - which is almost unheard of for a young freshman.
Why does Tatum excel by this method while posting seemingly unimpressive stats? For one, his shooting projection is largely driven by his fantastic 85% from the line, and good enough volume from deep, while only marginally effected by his mediocre percentage from three. For another, Tatum got to the rim by himself more than would be predicted from his reputation as a two point jump shooter - very narrowly missing a green flag in the “driving” category. Finally, it is also possible to have an assist to turnover rate of less than 1 and be rated a decent passer. As I said above, the passing category is very forgiving to forwards, as many of them seem to immediately improve their assist to turnover rate upon reaching the NBA because of the change in roles.
Here is OG Anunoby
It’s worth noting that this projection is based on only 16 games, making it less reliable than the others I’ve shared so far. Despite that, it is a good summary of Anunoby’s strengths and weaknesses as a prospect. His surprising green flag for “driving” hints at some nascent offensive skill that could be unlocked if he ever learns to shoot. The vast majority of players with a red flag for shooting never learn, though young, athletic forwards seem to have the best chance of learning to shoot well enough to be valuable (see: Jaylen Brown, Moe Harkless, Rondae Hollis-Jefferson). Anunoby is shooting 38% from deep in his rookie season, so he could prove one of those success stories.
Here is a pair of second rounders.
Ojeleye does not look good by SMILODON. He could shoot, but was very old by prospect standards and had poor athleticism indicators (defense & driving). As mentioned above, when evaluating a perimeter prospect, I’ll use defense & driving as an indication of functional athleticism, while shooting & passing can be combined to estimate a player’s offensive skill level. For example, in the 2014 draft, Nik Stauskas scored well on the skill indicators, but poorly on the athleticism indicators, while T.J. Warren did the reverse.
Bacon’s ability to get to the rim was important, but he was also older and didn’t put up numbers on defense. He will probably top out as a bench scorer. Ojeleye and Bacon have posted PERs of 4.3 and 4.5 respectively this year, but have received more than their share of minutes due to the current shortage of competent NBA wings.
Some of this year’s forwards have very good projections.
Miles Bridges has elite shooting thanks to his nearly 90% mark from the line, while Mikal Bridges is percentage points away from a green flag for defense thanks to a recent slump. Even beyond these players, this year’s forward class is much stronger on skilled players who might have enough athleticism than the reverse.
The big man position has had the most comprehensive set of changes since last year, starting with two entirely new categories. There are now five big man categories: two offense, two defense, and age.
Shooting is calculated in the same way as for forwards, but with lower baselines for players to hit to be considered “decent”, “good”, etc.
Creation is a combination of assists, adjusted turnovers, and unassisted makes at the rim. This includes transition makes, unlike for perimeter players, because knowing that a big man can consistently lead fast breaks is actually useful information and an impressive trait. This category attempts to assess how much potential a player has to become an offensive hub. Big men who can finish are a dime a dozen. Big men who can create opportunities are very rare, and highly valuable.
This is probably the category for which a mental adjustment for athleticism is the most important. Athletic players who score well here, like Aaron Gordon or Kyle Kuzma, tend to succeed despite a myriad of weaknesses. Less athletic players, like Frank Kaminsky or Mike Muscala (or John Shurna, Seth Tuttle, or other players that are not white guys from the Midwest, I promise), tend to be overrated in SMILODON thanks to this category.
Defensive Awareness is a combination of steals, team defense, defensive rebounding, and personal fouls. Steals make up the biggest component, and team defense is weighted more strongly for bigs than for forwards.
Rim Protection is block rate adjusted for standing reach.
Age is calculated the same way as for forwards and guards - a player’s age on February 1 of his draft year.
This system does not measure rim running, finishing, or offensive rebounding, so it’s very possible a player can put up perfectly good numbers in the NBA through those skills despite a pessimistic SMILODON projection. However, functional big men with those skills are freely available on the trade and free agent market, and bigs that significantly move the needle tend to have the skills measured by SMILODON.
Here is Lauri Markkanen.
The combination of elite shooting and a young age is almost as enticing as that defensive projection is frightening. In his defense, Arizona has a system that seems to suppress defensive stats. On the other hand, we still don’t know if Markkanen will turn into an acceptable defender.
Here is Justin Patton.
Patton is young and projects as a decent rim protector, which is nice. He doesn’t project as either a floor spacer or an offensive hub. All in all, this is the profile of a generic center - not a bad player, merely an easily replaceable one.
Rabb is another generic big man. He probably won’t protect the rim as well as Patton, but might have a slightly better chance to stretch the floor. I think Patton would be a better pick, due to his size and the importance of rim protection, but it’s pretty close. Jordan Bell’s profile is marred by age, but he has an intriguing defensive profile with enough offensive skill to get by. Subjectively, I definitely underrated Bell’s passing and ability to translate his shot blocking to the NBA when initially assessing his chances.
Here are a few big men that will be in June’s draft.
Ayton has a good offensive projection, as an athletic big who can shoot a little and will probably be able to create his own offense. His defensive profile is dicier, and if I was drafting him, I would hope it was another case of Arizona suppressing a prospect’s defensive numbers. Otherwise, he could end up as a “good stats, bad team” guy in the NBA.
Bagley should be able to score in the NBA. He is athletic, very young, and can already create offense on his own. But how many championship caliber teams start a big man that contributes little beyond volume scoring? Unless Bagley massively improves his shooting, passing, or defensive awareness - definitely possible, as he’s very young - he might not be the best use of a top five pick.
Bamba looks like an awesome defender, but a long shot to contribute much on offense. His shooting numbers have improved over the past month, so it’s worth paying attention to see if that continues.
Here is last year’s lottery by SMILODON. Eagle eyed readers will notice a projection for Frank Ntilikina. I am testing projections for European players this year. I am less happy than with the results for NCAA players - mostly because of worse data and the massive differences between leagues - but it’s a start. Ntilikina has a uniquely bad projection that he’s certain to outperform.
SMILODON 2.0, from 2012-2017, can be found here. International players are now included for the first time, but those results should be taken with a massive heaping of salt, as I have not yet adjusted them for league quality. Thus, Bruno Caboclo’s stint in the Brazilian Beer League is assumed to have the same validity as a season in the Euroleague, which is transparently nonsense. Still, when comparing players from the same quality of league, I believe they can be useful.
Coming next time...some early 2018 projections. Who else looks good this year?