Friday, April 29

Drafting high

Sorry, Browns fans - I'm still talking about the NBA Draft here. I read an interesting story by the PD's Jodie Valade (full disclosure: I know Jodie) about the recent history of teams with the #1 overall pick in the NBA and how they fared in their subsequent campaigns after garnering the draft's top prize. As I demonstrated a couple of weeks ago, pulling for the Cavs to finish dead last in the NBA was the completely wrong thing to do as a fan, and the performance of teams drafting first overall, as opposed to second or third, furthers that view.

#1 pick
Using Valade's table of win differentials the year after drafting first, teams with the #1 pick since 2003 experience an average of +10.9 wins the next year. Not bad. The lone club to decrease in wins in this scenario: this year's Wizards with John Wall. Part of the observed increase in team wins is because of the new star - part of it is because often these clubs have nowhere to go but up. The weirdest one on the table is Toronto's rise in 2006-07 after drafting Andrea Bargnani - he really isn't all that good but the Riptors still shot up 20 wins. Does anyone know why this happened? Anyway, if you throw out the Blazers' season after drafting Greg Oden and the Clippers' first campaign after selecting Blake Griffin (neither played a single game in their first year in the league), the average improvement is +11.3. That's right: over the past nine years, teams who didn't even get to use their #1 pick for so much as a single game did slightly better than those who had their prize available in the first year post-lottery.

#2 pick
The average improvement in wins here is +10.7 - almost identical to the #1 pick. There's one less data point here because Charlotte was an expansion team and of course had zero wins the year before taking Emeka Okafor, so I obviously excluded the "increase" they experienced in their inaugural season. The other special cases for the #2 pick are Detroit (which had the #2 in 2003 because of a trade, not because they were bad; in a normal lottery Dwyane Wade goes #2 and the average here almost certainly goes up) and Seattle in 2007 (because they picked Kevin Durant but were blowing up the team to move it.) Take those two points out and you're at +16.4, considerably better than the results from the #1. Interesting.

#3 pick
Since 2003, the third pick on average garners a club +13.9 wins in the following year, the highest of the top three! The only special case here is Memphis in '08 - they got #3 pick OJ Mayo in a trade rather than drafting him. Take him and an already-modestly good Grizzly team out of the equation, and you're at +15.6 wins. Now that's impressive!

Does this mean that the #3 pick is necessarily more valuable than the #1? No, of course not, and there's certainly more to evaluating a player's impact on a team than their improvement in wins over his first season. But these numbers do strongly suggest that picks in the Top 3 at the draft (and possibly a bit further down) are pretty fungible - you just have to nab the right player. It also suggests, in case I haven't been clear about this, that rooting for the Cavaliers to finish dead last in the NBA last year was a terrible, terrible idea.

An interesting exercise would be to see at what point in the draft order the expected value of win increase in the following season turns negative. The whole NBA goes .500 every year, so the sum of all the draft slots has to be close to zero (not exactly, since teams sometimes have multiple/no first-rounders in a given season), and teams usually have to be near the top to fall, though some have done a remarkably good job at staying on top. I wonder if the +/- flattens out (whatever the inverse of a plateau is) once you reach the middle of the draft or thereabouts.

No comments: