In previous mocks, say, 5.0 and before, we used predictive markers (steals, rebounds, two-point percentage, mutual exclusion properties, among many others) to paint the canvas and equate a player with a specific number. These predictive markers, drawn from established approaches, often can tell whether a player has the goods to translate their work to the NBA level. And, for the most part, it works. But there is something invisible--like the dark matter of the universe--that can't be seen through general stats. Namely, this affects athletic point guards and athletic centers, usually in a the proper range (say, 6'10" to 7'1"). So I've incorporated conditions where incorporating numbers where PGs or Cs have to meet a certain barometer in order to get the raise that's needed for the NBA. As you'll see, there are names in the third chart below where there is very little chance, statistically, you can construe, say, a Taj Gibson, a Russell Westbrook, a Lance Stephenson, a Taj Gibson, a DeAndre Jordan, an Andre Drummond or a Chandler Parsons as first round picks. It's just incredibly hard to find that medium. I've created enough conditions in my algorithm where all of them land as first round picks, although many of them become centralized in the mid-first round (or for Stephenson's and Westbrook's case, a very early 2nd rounder).
This is obviously not a foolproof method--for starters, Shawn James, shotblocking extraordinaire and Cedric Simmons, two notable busts, ranked well in this algorithm--because it's hard to find severe weaknesses to really dock them for. But Hassan Whiteside and Alan Williams, two guys notable for their combination of zero passing and very low 2-point percentages, are docked. Also, guards or forwards who really can't score after a certain age are docked, and there are only a few of these cases. The age is key--Ricky Rubio and Holiday came into the league extremely young, and are exempt for this condition. You'll see this hurt Briante Weber a bit--there's a reason scouts aren't that high on him, and it's because he's on the older side and doesn't score.
Without further ado, here's the results of this mock. I'd like to consider this my more accurate one, because this builds off what I thought was quite accurate in 5.0 and adds a historical context to allow others who might fare well:
The 2015 mock and the rankings of some historical players (who influenced this algorithm) are listed below the first table:
No comments:
Post a Comment