What Is a 书单? Chinese Booklists, and Why They Beat Algorithms

A 书单 (shudan, literally "book list") is a collection of novels assembled by a reader around a theme, published on the platform for everyone else to browse. Think of a public playlist, for books.

They are the least glamorous feature these platforms have and the most useful one, and they are almost invisible from English. There are around 97,000 Qidian booklists and 14,000 Fanqie ones indexed here, averaging well over a hundred books each.

Why a human list beats a filter

A filter can only ask questions the database has columns for. Genre, tag, word count, status, reader counts. Those are real questions and this site is built to answer them, but they have a hard ceiling: you cannot filter on a judgement.

A booklist can be titled "cultivation novels where the protagonist is actually clever", or "books that stay good after chapter 500", or "female leads who are not there to be rescued". Those are the criteria people genuinely select on, and not one of them is a column in any database.

A tag says what is in a book. A booklist says why somebody thought it was worth your time.

A recommendation algorithm cannot produce this either. It optimises for engagement against behavioural similarity, which is why it will happily feed you five more books exactly like the one you just finished, including the ones you did not like about it. A curator has a thesis.

What they look like in practice

The lists cluster into a few recognisable kinds.

Theme lists. The largest group, and the most useful. A specific premise, setting or character type, gathered across genres and platforms. This is where you find the thing you cannot describe to a search box.

Canon lists. "Essential xianxia", "the ones everyone has read". Very useful as a newcomer, because they encode the consensus that regular readers absorbed over years.

Completed lists. Curated around books that actually finished, which solves the single biggest failure mode of the format.

Author lists. One author's whole output, in order, usually with notes about where to start.

Deep-cut lists. Books the curator thinks are underrated. Highest variance, highest ceiling. If you have already read the obvious things, these are where the value is.

How to actually mine them

Three techniques, in increasing order of payoff.

Sort by recently updated, not by popular

The default instinct is to look at the most-viewed lists. Do not. Popularity accumulates, so the most-viewed lists are usually the oldest, and a curated list that has not been touched in three years is a snapshot of a catalogue that has moved on.

Sorting by when the list itself was last updated surfaces the curators who are still curating. That is the Qidian booklists and Fanqie booklists default here for exactly that reason.

Work backwards from one book you liked

This is the single best use of the feature and almost nobody does it.

Every book page here shows the booklists that book appears in. So when you find something you love, look at who else put it on a list. Those curators have demonstrably overlapping taste with yours, verified against a real data point rather than inferred from click behaviour. Everything else on their lists is a recommendation from a person you have just checked.

Do it twice and you have a small set of curators worth following, which is a better recommendation system than any of these platforms ship.

Read the list titles as a genre map

Even without opening them, scanning a few hundred booklist titles teaches you the vocabulary of what readers actually want, which is not the same as the platform's category tree. Recurring phrases in list titles are the real genres. This is how you discover that a trope you have been circling has a name, at which point the tag filter can take over.

The two platforms handle them differently

Qidian's booklists are a first-class feature with a browsable directory, curator profiles and a long history. Around 97,000 of them here, and the culture around them is mature: there are curators with reputations and lists that function as canon.

Fanqie's are harder to reach. They exist as "topics" rather than as a dedicated booklist surface, mixed into a general community feed, and the app gives you no systematic way to browse them. About 14,000 are indexed here, which took a fair amount of work to reach at all. The gap between the two numbers is mostly about how discoverable each platform makes them, not about how many exist.

What they are not

Not ranked. Order within a list is usually arbitrary. The list is the recommendation, not the position.

Not maintained. Books get delisted, authors stop writing, and nobody goes back to prune. Expect a proportion of dead links in any list more than a year old.

Not neutral. Some lists are promotion. A list of twenty books by one author, or one publisher's whole roster, is marketing wearing a curator's hat. The tell is usually uniformity: real taste is lumpy.

Not translated by their creators. Titles and descriptions here are machine-translated like everything else, with the Chinese kept beside them. List titles are short and idiomatic, which is the worst case for machine translation, so read them generously.

Where to start

If you have a genre in mind, start from the Qidian booklists, sorted by recently updated, and open anything whose title makes a claim rather than naming a category.

If you already have a book you love, find it in the catalogue and look at which lists carry it. That is the shortcut.

And if booklists turn out to be too fuzzy for what you want, the precise tool is on the other side of the site: filter by trope tag and sort by sustained readership. The two approaches answer different questions, and using both is how you stop running out of things to read.