Draft Philosophy is a new series my colleagues and I are starting to examine different philosophical questions that plague draft analysis. Some of these posts will be very short, some might be excessively long. Enjoy responsibly.
In the NBA, talent is distributed something like this. (Excuse my terrible artistic skills and don’t take the labels too seriously.)
There are few truly great or terrible players, and the majority of players are slightly below-average. The thing is, that graph is only representative of the NBA talent pool. The graph of all players who have a chance of playing at the NBA level looks something more like this.
This might not seem revolutionary, but it is an important distinction. For every Monte Morris sitting at the end of the Nuggets bench, there are at least 30 point guards playing professionally in the G-League or overseas of comparable quality. That number might even be low because the point guard position is the most abundant by virtue of height distribution in the population.
Even for more established NBA players like Tyler Zeller, there are probably 5-10 guys in other leagues of similar quality. The reason for this discrepancy is the constraints of a 12-15 man roster. But why does it matter?
Well, when constructing draft boards it has implications on the way to set up tiers and view the talent pool. Let me throw one more chart at you and then get to the meat of the analysis. This chart was produced by Seth Partnow way back before he got hired by the Bucks. The article it was posted with no longer exists so I don’t know the exact distinctions of Bust/Rotation/Good Rotation/Star, but I trust Seth’s analysis. For the purposes of this article only focus on the top row of the graphic.
On average, out of 60 players, 41 end up as busts, 11 end up as rotation guys, 5 as good rotation players, and 3 as stars. Now, that is just an average. In a deep draft maybe 25 guys end up as rotation players or better. A top heavy draft might have 10 good rotation players.
For example, this year’s draft looks particularly top-heavy. There are six guys I’d give bonafide star potential to, and another couple who profile as high-end good rotation players. Where this connects back to my initial graphs is the way we go about constructing tiers in a draft board.
Depending on a draft’s quality, the top ~6-12 players should be split up into two or three tiers. Next, one big tier of ~10-15 guys should make up the middle of the first round. After that, #20 or so on a big board is probably closer to #60 or even #100 then they are to #5. It’s not that the gap between the 20th and 5th best player in a draft is huge, but that the gap between the 20th and 100th best player in a draft class is tiny.
This framework shouldn’t be taken as gospel. As stated earlier, every draft has different strengths and weaknesses. One should feel comfortable deviating from this tiered structure. But one should really consider why they think a draft will upend the historical norm in whatever way they’re projecting.
This is also not to say that every player ranked below a certain point on a big board is going to be a bust. The draft is a game of uncertainty. You are ranking guys based on range of possible outcomes. Not knowledge of a specific outcome. Instead, a ranking of below 20 or so on a big board should indicate you think it’s unlikely that player will turn into an NBA rotation guy.
Generally, people seem to intuit these notions of talent drop-off in the first 20 or so picks, but try and delineate too much after that. Right now, the top-12 of The Stepien board consists of 4 tiers, and the 5th tier stretches from 13-20. This is good. It reflects the way talent usually ends up dispersing across a draft. After 20, things get a little more dicey.
Not to call out my colleagues (totally to call out my colleagues), but Cole Zwicker, Mike Gribanov, and Jackson Hoy’s boards all divvy up the post-20 range of their boards into three or more tiers. Based on historical data and the general distribution of NBA talent, this is an attempt at finding granularity that does not really exist.
I’ve divided up similarly in the past, but I’m going to think longer about doing so again in the future. I might have preferences or think I can differentiate between players in this range, but I am probably fooling myself in doing so. Being cognizant of the talent distribution is a useful framework to have in mind when doing draft analysis.