It's all kind of subjective. I have very little expectations of our rookies and doubt any become starters. I'd have no qualms trading the whole lot of em for a shot at healthy Embiid. Given that we tried to give up twice as much for the #9 pick, I don't feel too awkward suggesting it.
But if there's a better than 50% chance that Embiid's career is over, I guess it would be a bit crazy to trade for him. Nobody knows what his actual situation entails though.
It's a fair call and is (of course) subjective.
Personally though, I disagree on
This is a pre-draft prediction, based on all the variables stated above.
My understanding is that:
a) The probabilities are generated based on semi-objective data gathered from past draft's days (e.g. in the past 30 years, 15 players taken at #16 turned out busts - hence there is a 50% probability that a player taken at #16 will be a bust).
I say semi-objective because obviously there needs to be at least some subjectivity in what separates a 'rotation player' from or a 'good rotation player', a 'good rotation player' from a 'star' etc.
Is Rudy Gay a 'good rotation player' or is he a 'star'? How about Tyreke Evans, or Demar Derozan? Is Amir Johnson a 'rotation player', or a 'good rotation player'? How about Avery Bradley?
You get what I'm saying.
b) The predicted draft positions are based on subjective analysis - somebody sits there, watches tape of a guy, and says "I think this guy fits team x, and they are picking at #8, so I'm thinking he'll probably go at #8". But then somebody like Cleveland goes and selects that player at #1. So is the probably of success for this player based on a #8 slot (which was subjectively determined as a spot that makes sense for their talent level) or a #1 slot (where the real world team was actually willing to draft them)?
For example, Danny apparently really like Rozier - hence why he got him in for a second workout. Really felt the kid had sky high potential. So, assuming Danny landed higher in the draft than he actually did - how high would he have had to go in order for him to take another player over Rozier?
If he drafted at #14 instead of #16, he may have still taken him. If he drafted at 11 (and Winslow was gone) he may have still taken him. We really don't know much - all we know is that if Danny in #10 or higher, he wouldn't have taken Rozier (since we already know he was willing to trade up for Winslow).
So now we're at a point where the team selecting has a strong influence on where a player gets drafted. If Danny was drafting at #11 and still selected Rozier, then is the probability of Rozier becoming a star still '0.7%'? Naturally no, and yet in reality this probability hasn't changed because he is exactly the same player regardless of where he was taken.
This is why applying the logic post-draft to the players taken at position X doesn't really work. If you're analysing the chance of certain players becoming stars (etc) then you'd probably get much more accurate results by basing it on their talents compared to similar players from past drafts.
For example, in Embiid's case:
How many long, athletic, 7'0" centers who are good at shot-blocking, rebounding and post scoring have gone on to be stars / rotation players / busts.
I wonder if anybody has a statistical system that works based off attributes like that? I've no doubts it'd be easy enough to come up with one that based on individual aspects that fall within the following broad criteria:
a) Physical attributes (height, weight, size, length, vertical, etc)
b) Position
c) Medical risk (foot history, back history, shoulder history, ankle history, no history, etc)
d) Tangible talents/skills (rebounding, passing, shooting, etc)
e) Intangible talents/skills (IQ, work ethic, selflessness, etc)
People have always been quick to point out that certain skills and/or attributes are more likely to translate to the NBA than others. If you captured information on all of the above, since the first draft in 1984 (or whatever it was) with a conclusion on how good that player ended up becoming...I'm sure that would give you a pretty powerful tool for predicting the likely success of future draftees.