After all the arguments about how bad/not bad Grant's drafting has been I thought I would sketch out a simple system for rating draft picks within their own draft. This is what I have come up with to start, I will lay it out piece by piece and explain why for each piece. I am hoping to get a discussion going, not to say "look at this system, It works perfectly (or even great)". If you have an objection to something- say you hate win shares- then please suggest an alternative, or a way to improve the use of that stat. There are plenty of opportunities to post "win shares suck", please let us try to have a conversation here.
The system so far-
1. Take a players WS as a proxy for the quality of career that they have had so far. This should be self explanatory.
2. Compare that win share to the average of the 5 players taken immediately after that player. The reasoning- a pick can only be as good as what is available, using the average lets us differentiate between a bust pick where there was 1 very good pick that could have been made and a bust pick where there were 3-4 other good picks that could have been made. I view one error as worse than the other. I am starting with 5 picks after only, but I am not married to the actual number. I do not think it is productive, however, to say that Charlie Villanueva was a bad pick at #7 because David Lee had a much better career after being taken #30. Basically we are using the 5 picks after a proxy for who it would be reasonable to take there. I am way, way open to arguments that it is a terrible number to use and we should use #x.
UHHHH- thats it. I said it was simple.
For those interested I ran it for drafts 2003-2009 and the top 10 picks were
The bottom 10 were
Obviously a lot of these bottom guys basically didn't play in the NBA, so their starting point was 0 win shares.