The Secret Link Between Human Stats and Stories

Right now I’m reading Bill Simmons’ epic tome, “The Book of Basketball:  The NBA According to the Sports Guy”, which is one of the funniest books I’ve ever read and which has also renewed my love and interest in basketball.  Simmons has written what is essentially a history and ethnography of professional basketball, which is a considerably easier task than it would be with other sports because pro ball has not been around that long.

At any rate, Simmons has now joined several other writers who have recently written about the volatile relationship between Moneyballization and a statistical revolution in basketball.  There was first Michael Lewis’s Moneyball: The Art of Winning an Unfair Game, which was about Bill James, sabermetrics, and the Oakland Athletics bringing on a statistician who sought to buy under-rated, cheap baseball players who had uncanny abilities to avoid outs on offense and get outs on defense.  What this took was an alternative measurement scheme, which valued less the older statistics:  homers, hits, RBIs.  What it valued more was an ability to get an extended pitch count while at bat, to draw walks, to foul off pitches, to have high on-base and slugging percentages, and to throw a lot of strikes consistently.

http://thephoenix.com/BLOGS/soxblog/archive/2009/05/06/bill-james-gets-animated.aspx

The site 82games became the source for alternative basketball stats after the statistical revolution of Moneyballization hit.  On it, you can view “clutch” rankings, the best player pairings, etc.  But these do not capture the whole picture.

Michael Lewis recently went on to basketball, to write a formative article about Shane Battier, Houston Rockets forward, which made Battier out to be a thinking man’s forward, studying tape and stats to figure out the best way to limit his opponent’s scoring game.

The FreeDarko blog guys (think Bethlehem Shoals) published a book called The Macrophenomenal Pro Basketball Almanac:  Styles, Stats, and Stars in Today’s Game, which pushed further, featuring stats, yes, but stating that team ball no matter interested them.  What was fascinating was looking into the heads of the agonizing, troubled superstars of the NBA.  Think Kobe and his Colorado hotel room, or T-Mac and his back (and skeletons in his closet), or Gilbert and his crazy half-court heaves.  This book focused on the personalities and the styles of players.

from http://www.freedarkobook.com/

Bill Simmons’ book seems like a blend of both.  Simmons points out, accurately, that basketball is not a perfectly measurable sport like baseball is.  Baseball is very much individualistic, with controlled standoffs between batter and pitcher.  Basketball, for Simmons, is about The Secret, which is finding players who are willing to give up personal stats in order to play more team ball.  And this chemistry requires that you watch a game in person to see how players react to each other, to see the killer instincts or lack of them.

And yet Simmons says that you can still look at stat lines in basketball and reconstruct a game fairly well:  points scored, free throws attempted, rebounds, assists, blocks, steals.  These can tell you a lot.  Experimental stats like the +/- used now just do not work very well.

So Simmons’ book is genius because it combines statistics with stories.  Yes, Wilt Chamberlain scored over 100 points in a game, but all his teammates hated him and he didn’t set up his teammates to be better players like Bill Russell did:

And yes, certain players put up amazing years, but when it came to the playoffs, they couldn’t pass the ball away faster to a teammate because they didn’t want to be clutch.  They were afraid of failure.

http://www.youtube.com/watch?v=mtKi9SCEJK4

This nexus of stats with stories seems to be unique to basketball as a team game.  You don’t see these sorts of studies in baseball, or football, because those seem to be more individual-oriented.  What I find interesting is how unofficial anthropologists seem so engaged with basketball.  As if the personalities suck in social science hobbyists more in basketball than with other sports.  For instance, Malcolm Gladwell wrote a New Yorker article on a girl’s team led by an outsider Indian dad who always wondered why teams let the other team walk up the whole court without being challenged.  He decided that the only way his less talented girl’s team could win was if they executed a full-court press THE WHOLE GAME.  And it worked fabulously.  While violating the norms of the game.  So other coaches and parents became furious!  These are the sorts of articles you can find about basketball.

From an interview with Simmons on The Onion’s AV Club:

AVC: When you talk about everybody seeming happy on their team… You have these observations that are unquantifiable. Like, for example, the fact that Los Angeles Clippers players hate their coach, Mike Dunleavy, Sr.

BS: That’s why I don’t even really go to other games anymore, other sports. I can watch any football game in HD and probably have a better time than I would at the stadium. The tailgates are obviously more fun, but the actual game experience is more fun on TV at this point. Baseball depends on the seats, and these games are so freaking long now that, I don’t know, I’d rather watch them at home, for the most part. If I lived near Fenway, I’d want to go to games. Basketball’s the one sport that you just pick up so much more when you’re at the games, especially if you have decent seats, or if you’re close to the court. You can watch the guys interact, you can watch them watch the JumboTron during time-outs, and see how they react to the coach. All human-nature stuff. It’s the most human of all the sports. And it’s my favorite part. That’s why I love going. It’s why I freaking pay the Clippers $15,000 every year to watch their latest shitty team.

Bill Simmons

My own experience with basketball has been bittersweet.  I didn’t start playing until after college, since I played tennis (poorly) in high school and baseball as a really young kid (and did okay…).  I’ve got no handles, no moves, and no killer instinct within the paint to finish a layup.

I really could have used some coaching.  Coaching would have taught me more aggressiveness and footwork, which I think are crucial towards making an average player a good player.  What I always felt was my advantage was my defense.  But in pick-up games, defense is not rewarded.  I was willing to run down fast breaks or cut off a slashing player driving to the basket.  And I had the length to disrupt shots or poke a ball away.

I felt like I worked a lot harder than other people on the court.  That doesn’t count much in a pick-up era which is based on ballhogs dribbling at the top of the key and then driving to the basket, ignoring teammates, and drawing a foul.  It’s enormously more frustrating because this me-first, my-stats-mean-everything mentality leads to pick-up games where people argue CONSTANTLY about every call and play and stand around peacocking.

So my love for playing this game has always been inhibited by pick-up stupidity, and I wonder what would happen if I had ever played on a team with a bunch of team players.

Getting to the Point:  Galapag.us

One of the issues for Galapag.us that comes up when I talk about it with people is that they refuse to be reduced down to numbers.  The repulsion is so extreme that the discussion gets shut off just at that initial point.

Certainly it is true that people are not just amorphous bundles of stats that can predict their every behavior.  But at the same time, peoples’ habits, rituals, and ticks are remarkably observable and predictable.  We tend to create myths about ourselves which may not necessarily be borne out of stats. And that variability, that defaulting to personality or character or defect, that is what really makes us the most interesting.

So a system that is built to support one’s reputation or identity, as Galapag.us will, should be built to wed the statistical side (quantitative) with the mythical/qualitative side.

This could be the holy grail of identity and reputation memory, human story-telling and myth, and statistical analysis improving social sciences.

Can someone figure out how to make a web site that allows someone to study his own stats to see where he could improve quantitatively, yet also somehow measure the intangibles, such as his character or quirks or funny stories about him or his nickname or heroic feats he’s done that didn’t square up with his numbers?

I mean, how do you design a system that accurately keeps someone like Dick Cheney in historical perspective (as a universally loathed and feared death dealer) despite on paper having superior credentials?

How do you capture Dwyane Wade’s night-in, night-out hustle on the court, scrambling constantly and relentlessly, while evoking support and inspiration from his teammates who love playing with him?  How do you capture a will for teamwork (or lack of it), determination (my only talent), and charisma (as in those guys who can get a girl to grab the phone out of hand and put her number in it, or an MLK Jr.-like leader who can get people to transcend themselves?).

How would Jesus Christ come out in a purely quantitative system?  In a world of monetary worship and diploma envy, would a guy who never accumulated wealth and who ended up being crucified rank well against a Barack Obama?

I envision a system underpinned with statistical analysis and automatically-fed stats coming in from peoples’ self-quantification devices.  But on top of that is a layer of human subjective input:  things like “if ‘the most interesting man in the world’ existed in the real world, this guy would be him” or “I would never want to work with this guy ever again” or “her children looked at her always with such reverence that others became envious”.  Did people like being around this person?  Was this person a stat whore who didn’t care about being a team player?  These are questions Simmons uses for his basketball analysis, and I love it.

Maybe that will be a component…along with your stats, there are questions such as “Was he a total dick?” or “Could you trust this person to take care of your cat?”  “Did he have killer instinct?  A swagger?”  “Was she so funny that people would laugh at anything she would say before she finished telling a joke?”

I want to figure out how this system will work — I really do think it’s the convergence of anthropology, economics, statistics, technology, sociology, politics, banking, everything.  And I think I’m uniquely placed to pull this off.