We had two bags of grass, seventy-five pellets of mescaline, five sheets of high-powered blotter acid, a saltshaker half-full of cocaine, and a whole galaxy of multi-colored uppers, downers, screamers, laughers… Also, a quart of tequila, a quart of rum, a case of beer, a pint of raw ether, and two dozen amyls. Not that we needed all that for the trip, but once you get locked into a serious drug collection, the tendency is to push it as far as you can. The only thing that really worried me was the ether. There is nothing in the world more helpless and irresponsible and depraved than a man in the depths of an ether binge, and I knew we’d get into that rotten stuff pretty soon.
Hunter S. Thompson, Fear and Loathing in Las Vegas
This weekend I’m back in Boston for the MIT Sloan Sports Analytics Conference. If you’ve never been, the conference is nowhere near as debased an enterprise as the above quote implies. But it is an experience containing a certain amount of visceral flair. This is Rumspringa for the sports analytics community and the Sports Media Industrial Complex is right there to document every moment, creating a few of their own.
I know many of you will spend the next two days following along from afar, live-streaming the panels and staying glued to your Twitter feeds. But those tiny digital portals don’t capture the true essence of the event. Have no fear, I’m here to help. Over the next two days I’m going to be keeping a live (mostly true) journal of my experiences here on Hardwood Paroxysm, continually updated throughout the conference. Some anecdotes will be explicitly factual, while others will be accentuated with a seasoning of exaggeration (or outright falsehood). But they will all be true to the experience, that’s my promise to you.
Friday – 8:45 AM
I left my pint of ether at home, but it’s still been a slightly groggy morning. Last night members of the APBRMetics forum gathered at a local drinking establishment to research the effects of mixing sports, Jameson, and Guinness. As of now, it appears the results are inconclusive (although there are no concerns about the sample size).
It was a slow crawl at registration, shuffling feet, eyes roaming across pastel ties and laptop bags. It’s always struck me as a little weird that each conference attendee must anoint their foreheads from a granite basin filled with Wade Boggs tears, but no one seemed to be complaining. We are all converts already. And if it makes Bill James happy, then we’re happy.
First panel for me this morning is Heads and Arms: How Big Data is Changing Professional Arm Wrestling. It’s a whole new world.
Friday – 10:20 AM
Turns out arm wrestling analytics are not nearly as interesting as actual arm wrestling. About halfway through I snuck out and moved over to the panel, In-Game Innovations: Genius or Gimmick?
This panel featured, among others, Shel Silverstein and the bad guy from Lethal Weapon 2. Silverstein’s rhyming couplets were confusing but they added some much needed humor. Ultimately, the group was clearly in agreement that gimmicks are genius and that Axel Foley was a loose cannon. It’s not shocking to hear these maverick minds recognizing and lauding loopholes in the rules the create inefficiencies to be exploited. It’s also not shocking to hear them advocate rule tweaks that would potentially create more of those inefficiencies. Loopholes set-up the interplay between risk and reward; that interplay creates stylistic and productive separation between teams and players. It also gives operations departments for sports teams something to analyze and the media something to hyperbolize. This group seemed especially interested in lengthening and sharpening that knife’s edge balance between risk and reward.
But I was left with a question. Is there a point where expanding the risk/reward trade-off wipes away the stylistic spectrum and creates two polar opposites — teams who take the risks and teams who don’t. It seems like there may be diminishing returns on the entertainment and productive value of risk/reward, the maximum benefits of that interplay may come not at its maximum disparity.
One other gripe. I know it’s the nature of this conference, but most of the discussion at this panel focused on the big three sports — baseball, football and basketball. They completely ignored the ever-expanding popularity of POGS and the expansive and sweeping rule changes that were recently implemented by the Federated International POG Association (FIPA). With the new use of fiberglass compounds for the construction of slammers, the old paradigm of POG strategy has completely changed. And it was completely ignored.
I’m off to the first research presentation of the day. Rudy Gay is presenting his paper, “Take These Stats and Shove Them Where The Sun Don’t Shine.” Should be groundbreaking stuff.
Friday – 11:21 PM
It turns out Rudy Gay’s presentation was mostly just him shouting and pounding on the podium. I wanted to be a good listener but when he started flicking the lights on and off it was time for me to get out of there.
From there I went to the “Automatically Recognizing On Ball Screens” presentation. As near as I can tell, these well-meaning researchers have created a way of having a computer analyze SportVU Player Tracking data to automatically recognize when an on-ball screen has occurred. But I’m cheating, because I snatched that shallow piece of knowledge right from the title of the paper. Other than that, this one went mostly right over my head in the exact way you’ve probably envisioned Sloan research presentations unfolding. Here are a collection of words and terms I didn’t really understand from this presentation — linear kernel (sp?), feature vector, feature extraction, linear SVM classifier, supervised classifier, pairwise interactions, Receiver Operator Characteristic, Area Under Receiver Operator Characteristic.
For someone like myself (and I assume, many of you), with no formal training in statistics this is often the defining experience of Sloan. The research questions make sense and the answers seem reasonable. But the techniques, even when explained slowly and carefully, are so far beyond our previous experiences that on some level we are left in the position of having to accept the findings on faith.
Speaking of faith, a non-denominational ceremony of spiritual celebration is scheduled for just before lunch. A giant pyre has been built outside the convention center upon which we will all be encouraged to place an antique abacus, symbolizing the exorcism of primitive techniques and tools. I’m a little worried though because I saw someone heading outside with what looked like an effigy of Doug Collins.
Friday – 12:24 PM
Well, that got a little out of hand.
I’m back inside the convention center, short an eyebrow and one shoelace. Burning archaic statistical tools quickly spiraled out of control. After the last abacus had been burned someone brought out a crate of sextants. Next thing I knew an Apple IIGS was crashing into the center of the bonfire scattering sparks everywhere. One of those sparks found it’s way to the tasseled fringe of Sam Hinkie’s gilded litter and the elephant carrying the litter reared up, startled by the fire, sending Hinkie sliding to the ground. His attendants were there quickly, pulling from the path of the stampeding elephant. If ever there was a reminder about the dangers of decadently constructed fire-ceremonies, this was it.
Back inside I was able to catch the presentation “The Hot Hand: A New Approach to an Old Fallacy.” I admittedly missed some of the methodology, but I did catch that J.R. Smith is 1,245% more likely to take the next shot…..regardless of whether or not his last shot went in. The researchers also demonstrated, unequivocally, that no matter how “hot” a player’s hand is there would never be enough heat transfer to the basketball, or heat conservation through the flight of the basketball, to actually set the net on fire when the shot is made. Kudos to the authors on proving that the “Hot Hand” is not a fallacy, but that the graphics in NBA Jam were.
Friday – 3:44 PM
The Basketball Analytics panel has just begun, with Zach Lowe leading us right into the question of communication and we connect the “jocks” and the “nerds.” No word on how to involve the drama kids or the stoners.
Also, no one seems to have noticed the elephant in the room.
Friday – 4:30
As with last year, Stan Van Gundy is the star of the Basketball Analytics panel with his passionate arguments against analytics. He bounced back and forth between critiquing the reliability of data collected and analyzed, calling into question the fundamentalist adherence to things demonstrated by data and just having himself a curmudgeon party. Van Gundy had good points — there are places where the analytics tell an incomplete story, there are places where the data itself is of questionable quality. But he did himself no favors with the generally confrontational tone of his comments:
Stan Van Gundy giving wedgies at the door on the way out of this panel, his buddies manning a wet willie station
— Danny Nowell (@dmnowell) February 28, 2014
Things got especially uncomfortable when he and Brad Stevens refused to speak to each other (dialogue is partially accurate).
Van Gundy: Zach, will you please tell Brad that minute limits are bull s##t?
Lowe: Stan, I’m pretty sure he can hear you. You have a microphone and you’re sitting ten feet away from each other.
Stevens: That’s okay Zach. Will you please tell Stan that we hardly ever practice in an effort to keep our players fresh and that his mustache looks silly? At that point Van Gundy continued to raise his voice and made a scene as he left the room.
I walked by Stan Van Gundy and he slapped my trapper keeper out of my hands and skateboarded away, smoking.
— Danny Nowell (@dmnowell) February 28, 2014
On a positive note, I believe Stan Van Gundy has earned himself a job with his performance tonight. The buzz around the conference was that the 1994 Miami Heat had made him an offer.