This is HN, hopefully someone smarter than I could come up with an algorithm that could model content bandwidth activity.
Content creates a spike of activity when it is "born" on the internet.
Content maintains activity over the life of its availability.
Activity may be a function of the contents [age, quality, rating, medium, popularity, relation to other content, etc]
\It is a sequel, part of a set, etc.
It would be interesting to be able to utilize the rating from netflix/imdb to predict how much total activity a piece of content might create.
The fresh content will dominate top-N lists, sure. It will also dominate headlines, because it will more often be the subject of newsworthy early-leaks and enforcement actions.
But there is so much more old content, and the collection/completist/sampling motivations so important for unauthorized sharing, that I could easily see the old content swamping the new in terms of total copies or total bandwidth.