People who think advanced stats are stupid (people like this guy, who should be looking for a new line of work if he believes what he wrote) should read this post Hoopdata’s Tom Haberstroh wrote at Hardwood Paroxysm last week. Go read it.
No matter how complicated the math behind them might be, the best advanced stats are rooted in common sense. They represent attempts to test out hypotheses smart people already think about before the stat exists or before they are aware of a stat’s existence. In the pre-Internet mid-1990s, I had a sense (as did many of you, I suspect) that walks were a very good thing for a baseball player to get, and I wondered why more announcers didn’t seem to care about them. Then I went to college, logged onto the Internet and found a dude named Rob Neyer writing at ESPNnet.sportzone.com and realized a whole discussion had been going for years about walks, slugging percentage and other stats I’d never heard about.
The math can be complex, but the conclusions they try to get at are easy to grasp.
Case in point: Haberstroh’s piece on assists. There’s all sorts of nasty-looking math in there (though it’s actually not that hard), with parentheses and capital letters next to lowercase letters in symbols like wAPG.
But it’s a fairly simple concept: Assists aren’t all equally valuable, and if you want to know who the world’s best passer is, you should find the guy who racks up the most valuable assists, right?
And that’s what Haberstroh is trying to do. Steve Nash dribbling into the paint, drawing the defense and dishing to Amare Stoudemire for a dunk is a different thing than Paul Pierce handing the ball off to KG for an open 20-foot jumper. The first one is a tougher pass that leads to a shot with a much greater chance of going in.
So Haberstroh created a new version of assists that gives players extra statistical credit for passes that lead to a) lay-ups/dunks; and b) three-pointers. The reason is simple: Lay-ups and dunks usually go in, and three-pointers are worth three points. If your passes are leading to those sorts of baskets, you’re doing great work.
Under Haberstroh’s system, the following players get the biggest boost in their assist totals, with wAPG representing how they do with the extra credit for high-efficiency assists:
Notice who’s missing? Rajon Rondo. I wondered how Rondo would fare here. Based on passing stats already available at Hoopdata, I figured he would fare pretty well. Good news for us: In his post at HP, Tom links to a spreadsheet showing the weighted assist numbers for (basically) every player in the league.
And where does Rajon rank? That would be 4th, the same place he ranks in regular assists. Here are the top four:
Steve Nash: 11.2 assists/g, 11.7 weighted assists/g = difference of +4.4 percent
Chris Paul: 11.2 assists/g, 11.4 weighted assists/g = difference of +2.3 percent
Deron Williams: 10.3 assists/g, 10.7 weighted assists/g = difference of + 4.2 percent
Rajon Rondo: 9.8 assists/g, 10.2 weighted assists/g = difference of +4.2 percent
Rajon does really well here. It’s obviously harder for a guy who already averages a ton of dimes to get a huge boost—percentage-wise—from Haberstroh’s method. And still: There’s Rondo, right with Nash and D-Will.
Some other tidbits: Of Rajon’s 599 assists at the time this piece appeared last week, 65.6 percent led to baskets at the rim or three-pointers. That’s not quite as high as LeBron’s percentage here (an off the charts 75 percent), but it edges out the marks of Nash, Williams and Paul.
Only two players (Nash and Williams) have assisted on more baskets at the rim. And only two (LeBron James and Nash again) have assisted on more three-pointers. (Note: Paul would almost certainly have more assists leading to threes than Rajon were it not for his knee injury, and that’s only sort of because the Hornets scorekeepers may be inflating assist stats).
Sure, this stuff is dependent (to a degree) on personnel. Any good point guard playing heavy minutes with Paul Pierce, Ray Allen, Eddie House and (yes, even him) Rasheed Wallace should assist on a decent number of threes. The at-the-rim numbers are more impressive, considering the C’s don’t have an interior finisher as talented as Stoudemire or Carlos Boozer.
The bottom line: Rondo has developed into the best kind of passer for this team. The next step will be adjusting to new personnel, but we don’t have to worry about that yet….right?