clock menu more-arrow no yes

Filed under:

MIT Sloan Sports Analytics Conference

New, comments

Better known as Dorkapalooza.

The conference was held on March 6th, headlined by luminaries like Rockets GM (and '00 Sloan School grad) Daryl Morey, author Michael Lewis, Colts GM Bill Polian, Mavs owner Mark Cuban, and ESPN Bah-stahhnophile Bill Simmons.

It attracted over 1,000 fans, job seekers, knowledge enthusiasts, sexless dorks, and rationals. Dan Shanoff offers a capable recap.

I thought this piece of advice from Shanoff was particularly meaningful:

Here's how I see it: If you really want to be a team front-office executive, you are better off spending your summer internship not fetching coffee for a team or leading stats-oriented Web site. Start your own (blog) and own the hell out of some segment of analysis. Post daily, post brilliantly, gently pass your stuff around to folks in media who would appreciate it. If it's good -- and if you're not good, you might as well not even try -- the teams will notice. There is no barrier (at least no publishing barrier) to becoming your own expert and putting your talent on display. THAT is the fast track.

There is some precedent here. 1970s Sabremetrics pioneer Bill James was hired as a special consultant to at least one major league team on the strength of a hand-stapled newsletter (after years of being ignored and aggressively mocked). In that sense, being a protester to the orthodoxy wasn't much different from Thomas Paine two hundred years previous or Martin Luther another two hundred before that - you were essentially a pamphleteer of grievance. The internet has changed everything.

The Blogger-to-Front-Office hire is coming.

Hardwood Paroxysm tells us that Moneyball analytical methods are already cracking the NBA establishment:

Trade and free agency decisions are increasingly being made with more statistical information, and Oliver broke down how widespread this is all becoming and stated that he knows of eight teams that have actually integrated advanced analytics into their decision-making (Boston, Cleveland, Dallas, Denver, Houston, Oklahoma City, Orlando and Portland).

Like MLB, the NBA and NFL will see an increased use of unconventional analytics. Anecdotal and experiential data - traditional scouting, the accumulated wisdom of former players and coaches - will be complemented, if not supplanted, by methods more friendly to falsification, verification, and the extraction of value. As methodological sophistication increases, we can all anticipate Goldman Sachs bundling Kevin Durant futures expressed in plus-minus, tying them to the Greek economy, sprinkled with a Laotian kip counterweight hedge.

A sure thing!

The challenge that the NBA and NFL face is that they are games featuring mutually dependent interplay and teammate-contingent results - unlike baseball, which is essentially a team game played in individual parallel, and thus lends itself to easy analysis. Breaking down the importance of the RB vs. the OL vs. the scheme isn't easy to do.

Unless you've followed Mike Shanahan's career.

My own guess is that the New New Thing in basketball will be the ability to marry meaningful plus-minus metrics with advanced performance statistics (say shooting percentage sorted by distance from basket, time on shot clock, good shot/bad shot, defended/undefended) and actual game film such that it becomes a teaching tool and is capable of behavioral impact, not just as an ex post facto evaluation tool.

It's one thing to tell a player that a 2-pointer shot from just inside the 3 point line is the worst shot in basketball or that the 3 pointer from the corner is one of the best, it's quite another to be able to integrate it as a real-time teaching tool.

Imagine game film where the player watches himself run through the offense, a small read-out above his on-screen head calculates the adjusted value of each potential shot depending on his spot on the floor. Or defensively, seeing the same metrics as the offensive player gets an clear-out iso at the top of the 3 point line with 12 seconds on the shot clock. As a GM, you're now informing action rather than just reacting to a season's body of work.

Jeff Ma recounts a part of the conference that I found most interesting as it speaks to what is most difficult to do in decision-making - the ability to pull the trigger on what your convictions and analysis tell you to do when all others are losing their heads:

Particularly interesting and relevant personally was a discussion between Pollian, Kraft and Simmons regarding Belichick's infamous 4th and 2 decision. Pollian and Kraft both supported Belichick's decision while Simmons' remains staunchly against the decision. Simmons' main argument centers on the Patriots' poor use of time outs and play calling preceding the 4th and 2 and less on the decision itself, but he remains adamant that it was the wrong decision.

Simmons' failure here is the inability to marry analysis with rationality, though he fancies himself a rational. Which is pretty much his calling card as it relates to Boston sports teams. This is primarily a failure of will, the refusal to see what is right in front of you, the graveyard of Wall Street analysts, Enron execs, utopian politicians, and the captains of wrecked ships. It's the most insidious sort of peer pressure. If you want to understand why Warren Buffett is a billionaire, you need look no further than this aspect of his character.

During this discussion, Polian made the point that using statistics to make this type of decision is irrelevant since there are so many factors that can't be taken into account by the statistics. Polian is actually citing something called "reduction bias" where in order to solve a complex problem, one reduces the number of variables to make the situation easier to solve. But what Polian fails to acknowledge is the importance of using numbers to help you make these types of decisions. Of course numbers are not going to definitively tell you what is right to do here, but in this case the numbers supported that there was an option besides punting the ball - an observation that most in football would not accept as feasible.

Right. It's not that the 4th and 2 call was clearly good or bad. It was a near thing, statistically, and contextually within the game. I'll even call it a coin flip. Yet the post-game reactionaries viewed a close proposition as something more like 95-5 against. Confirming that most of our beliefs are based on superstition, anecdote, and legend rather than hard numbers.

I'm guilty of it too.

I watched that Patriots-Colts game live and can recall my own instant bias: how can the Patriots not punt this ball? I began to mouth the familiar orthodoxy: It's 4th and 2, you must kick it and play defense! Then my rational mind kicked in, I started running the numbers, and realized that this was far from madness - in fact it was defensible. Probably advisable. However, the 4th down wasn't converted, so that it was the wrong call became ironclad fact. Now it was about media and fan perception and the predictable dumb recitation of football tradition and time-honored platitudes.

Also known as ESPN.

I listed to Bill Simmons post-game podcast and Simmons, one of the proponents of the New Analysis, spent an hour excoriating Belichick's play call, offering that it proved Belichick was being contrarian to reinforce his own personal brand as a risk taker to the detriment of the team, that he was now likely washed up, that it was an act without any rational explanation. Hmm. That Simmons is then a lead panelist at a conference examining a rational approach to sports is proof that when it comes to applied analysis, human psychology is always the final barrier.