Some Game Theory

I am reminded of a counterintuitive result that more people nearby will result in less action. The typical example in a game theory course is a crime witnessed by n people. As n grows, the likelihood that the police will be called actually falls. While this phenomenon has been well documented, I have recently observed a case where this is not true. An email was sent out to a mistakenly listserv. The recipients hit reply to all to inform the sender. As a few emails were sent, this took on a life of its own. Between the initial emails, replies to those emails, and emails telling people not to reply to all there were about 50 emails in all.

A couple possibilities

1. The game theory result is still correct. Where n increases, the probability of any reply falls. However, if the first reply to all is sent everybody will jump on the bandwagon.
2. This situation calls for a different set of assumptions than the typical game theory situations of shared responsibility.