[Bladesss]

This idea may not be possible or just beyond the scope of this plugin.

If you have the rating working and have a way of marking books as read/not-read. Can you use the ratings to recommend a book.

Concept:
using a temp variable. Sum the ratings of the books not read.
get a random number between 0 and the sum.
Add up the sums to find which book that random number represents. Jump to that book.
(this way you are 5 times as likely to get a 5 star recommendation then a one star recommendation).

Have a method to indicate you don't want that book. Repeat until it recommends a book you want to read.

Varient 1: Use the cube of the ratings. I would much rather read a 5 star book then a one star book. By cubing the ratings, I am 125 times more likely to get a 5 star book recomendation.

Varient 2: For books that dont have ratings, is it possible to use the average of the other rated books by the same author? Or if no rating by the author, have some method of noting and favoring award winning authors.

Varient 3: Have 2 ratings fields. The first is to store ratings from Goodreads or amazon. The 2nd, duplicates the first is available. The average of other books by the same author if available. Add one if it is an award winning author if known. Or use a default of 2 for an unrated book (even an unknown book is better then a book with a rating of one). Use the 2nd rating to select random books. The user should be able to enter ratings into the 2nd field, so that if he reads one good book by an author and rates it, then it will recommend more books of that author.

I do something like this in a spreadsheet. It would be nice is Calibre could do something like it.