||03-20-2013 11:00 PM
The Raps, Analytics and Giving away Secrets
Since there were no posts on this, here is an article in Grantland about the proprietary Analytics programs the Raps use to help them on a day to day basis:
It's pretty detailed, with specific examples (pretty cool, actually) and made me wonder: why in the world would a team give away a proprietary form of analytics? I wasn't the only one to think this (I've bolded a couple items for consumption):
Early this season, Philadelphia 76ers coach Doug Collins was asked if he believed in analytics — the attempt to superimpose the order of numbers onto the chaos of basketball.
“No. If I did that, I’d blow my brains out,” Collins told the Philadelphia Inquirer. “There’s 20-page printouts after every game. I would kill myself.
“My analytics are here,” Collins continued, pointing to his head, “and here,” pointing to his gut.
This is an increasingly untenable position in the NBA, at least publicly. For saying what many coaches still believe, Collins was painted as the league’s longest surviving Neanderthal.
The angry reaction to Collins’ stodginess (and the season-long pratfall by his club) was a tipping point. Analytics are to this year’s NBA season what beards were at the Academy Awards — something not so new as to be shocking, but still cool enough to attract would-be hipsters.
This may be why the Raptors decided recently to open up the guts of their own analytics skunkworks and show them to Grantland writer Zach Lowe. Lowe’s piece is fascinating inside baseball.
It explains how the Raptors use a camera technology called SportsVU to break down every play of every game into a series of computer coordinates. Fifteen NBA teams use SportsVU. The software that determines how the data is turned into a graphic representation of the way the game is played is proprietary, unique in each case.
Fourteen teams decided that sort of information is best kept to themselves. One team decided to pull back the curtain and let ESPN see Oz at work.
Included are a series of animated diagrams, provided by the Raptors, showing Toronto at work on defence. Players appear as a dot moving across the court. Alongside each is his “shadow.” The shadow Raptors move in what the program has decided was their optimal defensive posture. There isn’t much similarity between what the real players decided to do and what the computer thinks they ought to have done.
Three animations accompany the text. When told that on Tuesday afternoon, Raptors coach Dwane Casey reared back in surprise.
“What?” Casey said. “Not the actual thing?”
The actual thing.
Casey was left briefly speechless. It is difficult to credit that the people guarding the team vaults are also doing guided tours.
The SportsVU perspective is a merciless judge of instinctive decisions made hundreds of times every game. The program thinks a guy should be standing in the paint; instead he was out on the perimeter. The program thinks he should be sliding off to guard a higher-value shot-taker, when he’s stuck on his man.
Were you only watching events from SportsVU’s vantage, it appears the Raptors play basketball wrong.
That appears to be the view of the backroom boys interviewed by Lowe.
“A lot of coaches will say how great it is that analytics confirm what they already see,” Keith Boyarsky, an analytics consultant with the team, tells Lowe. “The fact of the matter is, that’s not really true.”
This is a guy who lives in L.A., who visits with the team from time to time, apparently burying the coaching staff. You know, the men who played the game, spend every day with the players and see them as something other than white dots on a monitor.
This is not an indictment of analytics, though that branch of study appears increasingly cultish and removed from the reality of the human decision making process under pressure.
It’s also important to remember that basketball analytics isn’t Moneyball. It’s far more revolutionary than that. Moneyball was an evaluative tool. Moneyball didn’t change the way baseball was played (at best, it put a larger emphasis on not playing, e.g. taking walks).
Basketball analytics would entirely change the way fully formed players approach the game. From a new onus on three-point shots to sneaky zone defence, its deepest converts advocate massive upheaval.
As Lowe’s article makes clear, the neutral territory is thinning out. The Raptors braintrust unwisely moved the battleground right into the middle of their organization.
According to Casey, the players never see the SportsVU animations. They watch only game tape. Showing them the utopian vision of “ghost” basketball would be too confusing (and, based on the small sample obtained by Lowe, too embarrassing).
Instead, all those gigs of data are boiled down to “four or five stats” that players can plug into their games — such directives as shoot more from this or that position.
That sounds like a sensible team-wide approach. What isn’t sound is publicly trumpeting a needless sense of friction between the geeks and the jocks. This all smacks of pointless grandstanding and destructive over-sharing. Casey cannot be cast as the dinosaur in this fight, if that’s what it really is. He comes out of Dallas, where Mark Cuban had the Mavs fitted up like NASA. Now he’s forced to deny any rift.
“The challenge is using (analytics) in the right way, where it’s not the only tool you’re using,” Casey said. “Everybody wants to get their point across. Analytical people want to say, ‘This is the tool to use.’ The scouts want to say, ‘My tool is important.’ The bottom line . . . is winning.”
It should be.
Unfortunately, Casey’s bosses appear too interested in being seen to be on the cutting edge, while undermining the reason why you go out there in the first place.
|All times are GMT -4. The time now is 09:54 PM.
Powered by vBulletin Version 3.5.4