Royal Road, Goodreads and Audible

The page joins two lists. The first is the top 500 of the "Best Rated" list on Royal Road on 29 October 2026. That list shows followers, but not how many people rated a story. It shows no status for about a quarter of the stories, mostly the stubbed ones; those count as "unknown".

The second is litrpg.lo5.me, a hand-kept list of about 1,500 LitRPG series by u/neoporcupine. It gives the Goodreads and Audible rating of each series' first book, and whether the series is complete. Series not marked complete count as "unknown". The copy here is from 28 October 2026. About 80 stories are on both lists.

Each Royal Road story was searched on Goodreads by its title. A hit counts only when the base title matches, without subtitle or volume number. The book also has to be from 2012 or later and have at least 5 ratings. Generic titles drew only namesakes and were checked by hand. Matches that look plausible but are not certain are drawn as hollow dots. Fan fiction is not searched.

When the search found nothing, a story can still get its Goodreads rating from the LitRPG list. That happens when the list has exactly one series under the same title. A list entry with the same Goodreads book or the same author also confirms an uncertain search match.

A searched story uses the volume with the most ratings, almost always the first. Later volumes are rated only by readers who liked the first, so they sit higher. The LitRPG list always gives the first book.

An average from few ratings is uncertain. The page offers two ways to account for that, for Goodreads and Audible alike.

"Shrunk" pulls a rating with few votes towards the mean of the same platform in the chosen list: (200 × mean + n × average) / (200 + n). A book with 20 ratings ends up close to the mean, whatever its average.

"20th percentile" treats the true rating as a distribution around the average. With fewer ratings the distribution is wider, and the page takes the value that the true rating exceeds with 80% probability. The prior here is weak (10 ratings' worth at the mean), so the book's own average counts for more. The width assumes a spread of 1 star per single rating, because the sources do not show the spread. For a book with 20 ratings the 20th percentile lies about 0.15 stars below the central estimate, for one with 2,000 ratings 0.02 stars.

The combined score works on percentiles, so the compressed Royal Road scale and the wider Goodreads scale become comparable. The percentiles are taken among the stories that have a value on every chosen platform. With one platform chosen, the score is simply its percentile. The lower of two percentiles punishes a story that only one platform loves. It also leans low, because the lower of two noisy numbers is low by chance more often.

Tags and fan fiction exist only on Royal Road. Those filters leave the series that are only on the LitRPG list alone.

Goodreads and Audible ratings have a long low tail. By default the plot leaves out the lowest 5% on such an axis, so the rest spreads out; the table keeps them. Dot size follows the logarithm of the Goodreads rating count, so one bestseller does not dwarf the rest.

Royal Road readers rate a free serial while it is still running. Readers who drop a story rarely rate it. Authors often ask for ratings in their chapter notes.

Goodreads readers rate a finished, edited, paid book. Only stories that sold well enough get a book at all, so the Goodreads side is already a selection.

Audible listeners rate an audiobook. The narrator shapes that rating as much as the story does. Audible ratings also bunch up near the top: in the LitRPG list the mean is about 4.37 stars on Audible and 4.14 on Goodreads. The percentiles take that offset out.

Within the Royal Road list the Royal Road and Goodreads ratings correlate only weakly. Part of that is the selection: every story there is already near the top of Royal Road. The narrower the selection, the weaker the link. Across all 500 stories with a Goodreads match the correlation is about 0.26, among the top 200 alone about 0.08.