Algorithmic observation—we eagerly buy books about other people doing their jobs

An article in Popular Science this week details the success of a computer algorithm designed to predict which novels are—and are not—bestsellers. Given a pile of 5,000 hits and misses to sort, the algorithm was able to tell bestsellers from flops with 80 per cent accuracy, on average.There’s a lot here for readers to find interesting. Despite the huge success of the 50 Shades series, for example, close analysis of the results indicate that, on the whole, sex doesn’t sell particularly well. Violence does a lot better.Bestselling novels use more active verbs than novels that don’t sell. As the article

Read More - The Fraser Institute