There is a bit of a complicated algorithm to it. It doesn’t go just by the average score of a book, but the quantity of reviews are factored in as well. There are also some other factors which contribute from user data.
One extreme example:
– 19 people rate “Killing Floor” by Lee Child 10/10. 1 person rates it 9/10. This gives it an average rating of 9.95/10.
– 1 person rates “Die Trying” by Lee Child 10/10. This gives it an average rating of 10/10.
Technically, “Die Trying” is higher based on average rating, however our system would list “Killing Floor” as the highest rated book, due to more 10/10s and the overall quantity of reviews compared to “Die Trying”.
The system we have in place is also set up to prevent “review bombing” scenarios as it factors in specific users and their review history.