Analyzing collegiate criterium bicycle races from the social networks and complex systems perspectives.

摘要

Ten collegiate criterium bicycle races were filmed in their entirety during the 1996 and 1997 road racing seasons using an 8mm camera. From these tapes were derived the times to the nearest 1/60th of a second that each participating racer crossed the start/finish line each lap. These data were then used as input into two network analyses techniques, one developed specifically for this research (time gap analysis), and the other widely available (principal components analysis). The results of these two procedures were then compared and contrasted, where it was found that principal components analysis was very good at teasing out the important groups of racers that formed during the races. Also, the following four hypotheses concerning the behavior of collegiate criterium racers were explored: (1) teammates tend to stay closer together than nonteammates, (2) racers from smaller teams are more likely to participate in break groups than racers from larger teams, (3) smaller break groups are more likely to succeed than larger break groups, and (4) break groups composed of only single racers from each participating team are more likely to succeed than break groups that included at least one pair of teammates.;The first, second, and last chapters of the dissertation were devoted to describing important aspects of the cycling subculture from the perspective of a participant observer, describing the four paradigms that informed this research (social networks, complex systems, evolutionary game theory, evolutionary psychology), and analyzing bicycle racing from the perspective of each of these paradigms.