the big dance starts with 64 teams. over the course of several rounds, teams are eliminated: from 32 to the sweet 16 to the elite eight to the final four and then the last two teams remaining play for the national championship. year before last, we were the national champs. i thought we were going to do it again last year, but alas...uconn had something else in mind. a moment of silence, please. anyway...each round can be considered a small world. or maybe a bracket is a small world.

so, in a few weeks, the selection process for the big dance takes place. generally, the champions from the 26 different conferences are invited plus the remaining at-large bids, the other "best" teams in the nation. there's a random element to this selection process. an inevitably, one team becomes the "cinderella" team--the one team that surpassed all expectations and predictions. in 1996, austin peay was the "cinderella" team.

clustering coefficient -- the degree to which things in a network are inter-related. let's look at the big east. in our conference, there's a large clustering coefficient because all conference teams play one another, and i'm using the conference schedule to define the teams' relationships. if a big-east conference team plays a team from outside the conference, like we did this season, that can be considered a kind of "solaria" relationship--one that is "random" and "independent." if we expanded our gaze and look at all the teams in the ncaa, the clustering coefficient is much smaller since not all teams in the ncaa, or in the tournament for that matter, will play one another. so i guess the big-east example is analogous to the cave scenario, and we can contrast that with the ncaa/solaria example: "Back on Earth, life is lived in the security of interlocking and mutually reinforcing ties, and intiating a relationship with a random stranger would be inconceivable. But on Solaria, all interactions are equally accessble, and prior relationships are relatively unimportant to the establishment of new ones" (74-5).

at this point, i'm using watts' network theory as a way to understand how networks form, and that especially makes sense coming from my/our disciplinary affliations. we can chart our various memberships in various networks. however, it seems to me that in other places, like at the nih, network theory can be used in a more preventative fashion, like trying to predict a pattern of infection. how can we use network theory in our field? would our uses be prescriptive or descriptive?