I think the boffins may have pranged a kite here.

In an attempt to define a unit of fame (the “Warhol” perhaps?) they seem to have neglected its overriding property – that fame is comparative. The fame of each participant affects that of the rest of them. In an n-body world of mutually orbiting celebrity, the behaviour of the fame indices would seem a perfect candidate for not a linear or exponential relationship, but a chaotic one.