Advanced Analytics in Ultimate

It was only a matter of time.  Apparently, folks are trying to bring the sabermetric revolution to ultimate frisbee (H/T to Dan Drezner).  Ultimate has gotten increasingly serious over the course of my time playing the sport.  The people who start playing in college now have been playing for years and are far more athletic than when I started.  There are now a couple of professional leagues in the US, regular world championships with a fair amount of representation from around the globe, an 11th edition of the rules, and so on. 

Indeed, one of the newest rules is somewhat counter to the old spirit of the game.  It used to be that if you committed a foul, you would call it on yourself.  However, at the most competitive levels, this became a trick that a player could do to stop play if his/her team was out of position.  I even got into a mild argument last week because I told the player I was guarding that he might want to call a foul on me (I had kicked his hand trying to block the disk he was throwing), and he said I should call it.  I eventually explained that the rules had changed....

Anyhow, the idea that we can learn about how to play better via more data makes complete sense.  I had one team that did try to collect stats--they would audio-tape play by play so that we could count how many passes someone threw well or errantly, how many dropped passes we would have (I compensate for a lack of speed/height by very rarely dropping the disk #humblebrag), how many defensive plays were made and so on. 

The WSJ story focuses on the usual stuff--how to measure which players contribute the most--but also strategy.  

Mr. Childers and Mr. Weiss crunched the numbers and were able to prove, for example, that if a player has the Frisbee in a corner close to his own end zone, then the optimal pass is actually a backward pass to the middle of the field, even if doing that seems counterproductive.

 This is actually far from counter-intuitive, but glad that we have data that backs up a conventional wisdom: tis better to move the disk to the middle of the field, even by going backwards, than stay stuck along one sideline.  I have long observed that teams that move the disk laterally do far better than those that do not.  The terms vary--in the US and Quebec it is called dumping and in Ontario and beyond it seems to be called bailing--but the idea is the same.  You can only hold onto the disk for 8-10 seconds [the opponent marking the person with the disk calls out 1 2 3 and so on--some say mississippi, some say steamboat (Canada/English) or bateau (Canada French)], so you have to keep the disk moving and throwing it backwards or sideways re-sets the stall count again and again.

The funny thing is that teams in rec leagues and even many teams that compete at tourneys kind of suck at dumping the disk.  It is not a hard pass if the two players are committed to it, but one is often distracted by the fleeting temptation to throw the disk up the field.  Again, the good teams do this quite well.

The article's story about the NY team that does not score enough when close hits home with me, as I have played with many teams that are good for most of the field but then turn it over when close to scoring.  It is not only frustrating but can be game-breaking. 

There is far more that can be figured out--like which kinds of zone defenses work the best, which kinds of offensive plays are more or less successful, does breaking the mark (throwing to the side of the field the opponent is trying to take away) as rewarding as it seems to be?

Just as baseball has this finite thing that has become the obsession--outs which limit the opportunities, ultimate has something pretty basic too--turnovers.  The team that turns over the disk less wins.  So, I think the ulti-metrics should focus on exactly that question--the causes and prevention of turnovers.

We live in interesting times...