EU by the numbers: ranking the Star Wars Legends adult novels
One of the many debates that can be had within the context of the expanded universe is the debate of the best and worst. It is a debate that still gets made to this day, and there seems to be no genuine consensus within the community of what is the best. But I had an idea, not on how to solve this debate once and for all, but a way to add context to it with data.
My idea was to somehow take the opinion of every person that has read these books, correlate that to a number, combine that number with the number of everyone else who read that book, and then rank everything based on those numbers. This is of course next to impossible as there is no way to find every single person that has read a Legends novel and get their thoughts on every novel they've read. But there is a way we can at least get close, and that is in user star ratings. Many websites allow users to review books and give star ratings, these websites make these public, so if you could combined all of these star ratings into one score you could get a number that more or less represents the general consensus of everyone who read and reviewed the book. Then all you would need to do is repeat this for every Expanded Universe book and rank them.
I did just that and collected 1,522,923 user star ratings from various book rating websites to get a consensus user star rating for every Star Wars Legends Novel. This article will serve as the first in a series of articles where I use my findings from taking the user star ratings of every single Legends book to look at various trends and rank this data in interesting ways. In this article I will simply use the findings to rank the adult novels from worst to best based on these numbers. This isn't my opinion, as I haven't read all of these books and don't even agree with how the list came out. This is simply the hard numbers from my research and calculations, so don't get mad at me if your favorite book doesn't rate as highly as you think it should.
But first, my methodology. The websites I chose to use were Amazon, Barnes and Noble, Goodreads, LibraryThing, and anobii. I am well aware of the fact that these are not the only websites on the internet that have user star ratings for books, in fact I looked at over two dozen sites that I either found or were recommended to me by people I consulted with in the process of making this series. The reason I did not include these other sites is generally because they did not have a sufficient number of ratings to effect the overall totals. I'll explain more when I get to how I calculated this but generally any rating site that didn't have double digit ratings for most of the books weren't even adding or subtracting a thousandth of a percent to the book's overall star rating. Maybe had I added them all it might move the needle to almost a full percent but it would be too small to have an appreciable impact so I left out any website that didn't have at least double digit ratings for at least a quarter of the novels. So that just leaves those five, but if you know of a major one that I missed that has dozens of ratings for all of these books let me know.
I also know that in the time since I collected these ratings that new ratings have likely been added but again any additional effect on this data the handful of new ratings had would be miniscule and thus I won't be adding them over the course of the next few months as the later articles come out. This is also to prevent any bias as I imagine after these articles come out there may be some people who go to these websites to support or criticize the books they like or dislike and I'd like to keep this data free of that bias.
Next the math. Every website I used has a five star rating system so all of their overall star ratings are out of 5 stars. Though an issue that comes up when comparing multiple websites or sources is how do you value the number of ratings on each. Say one has 200 ratings of a 3 and 300 ratings of a 5? if you average the ratings you get 4, but that doesn't seem fair as 100 more people gave the book a rating of 5 than did a 3. My solution was to use weighted averages.
To the non mathematically inclined this is in this case simply taking the number of user ratings and multiplying them by what that website considered the overall star rating to be. This was done for each website and were added together. Once added all websites then had their total star ratings added together, I divided by the total number of ratings.
Using the earlier numbers in my hypothetical as an example, site A had 200 ratings and an average of 3 stars , and site B had 300 ratings that averaged 5 stars. So 200*3 equals 600 total stars and 300* 5 equals 1500 total stars, add these together and you get 2100 total stars the book received across the two websites. Then divide by the total number of ratings to get the average stars per ratings. So 200 ratings + 300 ratings is 500 total ratings and 2100 total stars divided by 500 total ratings is 4.2 stars. So in this case the book has an overall ranking of 4.2 stars.
One final thing that I did was in regards to series of books that were later collected into single books. For the most part I only included reviews for each book on its own not the collection though there were two exceptions to this rule and that was for the Brian Daley Han Solo Trilogy and the L. Neil Smith Lando Trilogy. These were both originally released over 40 years ago as single books but after Bantam took over in the 90s have only been available in the 3 in one collections. As a result reviews for the individual books are far more scarce than reviews for the three in ones. So I decided to add the ratings of the three in ones to the ratings of the individual books so that they would be on par in terms of number of ratings as the other books.
So without further stalling the results. I found that Goodreads and LibraryThing generally had more ratings than the others but also rarely had star ratings higher than four stars. I think this is because these two are social media sites not book sellers. Barnes and Noble and Amazon being sellers means that people may be giving books higher ratings as they are factoring in the book purchasing experience which is generally pretty good. But that's just a theory of mine. If you're curious the average star rating and number of ratings per book for each site was.
Amazon - 4.568 Stars, 275.4 Ratings per Book
Barnes and Noble - 4.115 Stars, 33.1 Ratings per Book
Goodreads - 3.747 Stars, 4966.8 Ratings per Book
Library Thing - 3.4785 Stars, 96.3 Ratings per Book
anobii - 3.797 Stars, 9.8 Ratings per Book
Yes I know anobii doesn't make the at least a dozen ratings threshold and Barnes and Noble also in on the low end of ratings per book, this is because these two also didn't have any user star ratings on over a dozen of the YA books. I decided to keep them as they were adding a lot of reviews for the adult novels but they were essentially non factors when it came to the YA novels.
The average star rating across all Star Wars books is about a 3.83 stars with a standard deviation of 0.2526. This gives a standard deviation graph that looks something like this.
It is a pretty nice bell curve with the exception of a longer tail at the lower end. As for the actual list going by these metrics we have the following. Rated from worst to best for maximum suspense. Star rating is out of five, and total ratings is how many people left a rating for that book.
I didn't include it in this list as it wasn't an official release but Heart of the Jedi did have 30 Goodreads reviews and 24 Amazon ratings at the time of the last snapshot I have on the Wayback Machine before it was taken down. The latter reviews are probably inflated because of the hype many had of that book being on Amazon but if your curious it had a overall rating of 3.91111 stars which would put it 68th on the list.
Honestly there's a lot to unpack here which I will be doing in several follow up articles. I know I was quite surprised and even upset at some of the ranking positions but that's just where the numbers put them. But for the moment what do you think of this list? Let us know and I'll be back tomorrow with some graphs and tables further analyzing these numbers.
This article is part of a series of articles, click here for the next article in the series.
To see more EU news, updates, and original content about the Expanded Universe, click here!