I'm curious as well. Did it get added to so many reading lists that it broke through the positive barrier and came up through the negative ranks?
I did it for you this time around, but if you spot a bug next time, please use this form to report it:
I hope that whomever fixes the bug let's us know the way it was previously calculated, because I'm curious. Adding up how many people have the novel in their reading lists, then removing the duplicates with a unique quantifier is basic SQL for database pulling. I've always liked figuring out why odd results come out with database queries.
It just adds up the number of people that have the novel in their Reading List. I think the only person that has access to the means of fixing this bug is @Tony. As for what caused it, my guess is some bug involving duplicate novel entries getting merged and/or people moving the novel through their reading lists and later on deleting it... But those are just wild guesses, really.