Posted by Neil Paine on January 25, 2010
Today I wanted to tackle the question of "team talent" -- which team has the most, and which team wins the most with the least? Unfortunately, there's no real way to quantify "raw talent" in a sense of "what is Player X's basketball potential right now?" The best you could do is to look at past years and regress to the mean for a projection, which we did a lot over the past offseason when attempting to predict player performance in 2009-10. But that's a different way to define "talent" than I wanted to look at here -- that method assumes a player has at least shown some flashes of potential in the past, upon which you can base future expectations. But about players who are still considered prospects despite having done nothing at the NBA level? In cases like that, I suppose the only way to really capture "talent" is to use draft position as a proxy, since draft decisions are made on the basis of raw talent the majority of the time. So who are the teams right now with the best collection of talent, based on where their players were drafted?
(Note: Average Draft Pick # was weighted by 2009-10 minutes played. Undrafted players were assigned a value of 1 + the overall pick# of the final player selected in the draft prior to their rookie season.)
So by this measure, at least, the Thunder have the best collection of raw talent in the NBA, a notion that isn't all that hard to believe. However, this method treats a 35-year-old #4 overall pick as the equal of a 25-year-old, which doesn't seem very equitable, since I'm pretty sure nobody would say the 2010 version of Rasheed Wallace had more potential than the 2010 version of Chris Paul (unless we were talking about the potential to draw fines for criticizing officials). How do we fix this problem? Well, the issue lies in determining how to weight the expected production of a draft pick at each age, and since we have a nice general metric of production in Win Shares, I simply took every player taken in the modern (1966-present) draft, calculated their WS/min. at each age, and then averaged those rates together by pick# (undrafted players were expected to contribute 0 WS/mp). Using these "expected" Win Share rates and each team's actual minutes played, I established the # of win shares each team would have if their players performed like the average player taken at their draft position did at the same age:
Don't worry, those numbers look incredibly low to me, too. Why are they that way? Because, simply put, most draft picks don't work out, especially the further down the board you go. When players drop out of the league, they stop producing WS at all, and are replaced by someone who can get the job done -- and when those players no longer cut it, they're replaced by someone else, ad infinitum. This is why I took a non-weighted average of all of the WS rates, because a weighted average (by minutes, or what have you) would have been biased toward the "survivors", the players who were good enough to stick around. I wanted to capture the average performance rate of every pick at a given age, even if he wasn't in the NBA anymore, and use that as a baseline, rather than skew the baseline towards the few players who actually stayed in basketball for a number of seasons.
So in addition to the fact that now Denver supposedly has the perfect blend of drafted talent and age/experience, what's interesting here is that if teams stubbornly held onto all of their draft picks, even if they produced no wins at all, they'd win approximately 1/3 as many games as they do by Darwinistically weeding out ineffective players. None of which explains the 3-40 Nets, who by both measures have the talent to far exceed their meager win output in 2010 so far.