100+ Ao3 Most Written Ships 2023
An Archive of Our Own, a project of the Organization for Transformative Works
Number of new works added between 4th of August 2022 to 7th of August 2023 within 369 days. Using this query: created_at:["2022-08-04" TO "2023-08-07"]
#5 Drarry
I’m sure many have seen another list:
On which Drarry is ranked at #34 instead.
Why such huge difference in ranking (#34 -> #5)? This is because the definition of “new” works are different between the 2 lists.
We have to look at the Q&A of the widely circulated list.
An Archive of Our Own, a project of the Organization for Transformative Works
From the Q&A:
New Works - The difference in the numbers of fandoms between 4th August 2022 and 7th August 2023, which is equal to the number of new works created in that time period minus the number of existing works deleted or locked in the same period.
It is not possible to directly access the "This Year's List" data from an AO3 search, as it requires having a snapshot of last year's data to compare to. When creating the "This Year" and "Year In Review" lists, I subtract the number of unlocked works which existed at the start of the time period from the number which existed at the end. This gives a result equal to the number of new works added in that time, minus any older works which have been deleted or locked. As a result, if a large number of older works are deleted or locked, the pairing may appear much lower on the list than expected, not appear at all, or even be listed with a negative number of "new works".
What they essentially acknowledge here is, instead of tabulating the actual number of new works added over a time period (gross growth), they calculate the difference of publicly available works over a time period (net growth).
For example,
At the end of 2022 period, Drarry has 56,506 publicly available works.
At the end of 2023 period, Drarry has 59,048 publicly available works.
Diff = 59,048 - 56,506 = 2,542 “new” works
But using the actual query: created_at:["2022-08-04" TO "2023-08-07"]
There are actually about 6.9k new works posted publicly for Drarry between August 04, 2022 to August 07, 2023. But over the same period, there were about 4.4k works deleted/locked. Which is why there is only about 2.5k “new” works.
TLDR:
Between 4th August 2022 and 7th August 2023, for Drarry:
- 6.9k new public works added (rank #5) -> gross growth
- 4.4k works locked/deleted (rank #4)
- 2.5k “new” public works available (rank #34) -> net growth
Follow up: So the 2023 ship stats lists are released:
On the traditional list, Drarry ranks #14
An Archive of Our Own, a project of the Organization for Transformative Works
On the alternative list, Drarry ranks #7
An Archive of Our Own, a project of the Organization for Transformative Works
Why the huge difference?
The traditional list tracks net growth (# of new works - # of deleted or privatized existing works) instead of gross growth (# of new works). It also doesn’t take into consideration newly added private works.
The reason why drarry ranked so low on the traditional list is because during 2023, many drarry fics were deleted or locked (i.e, taken out of public circulation) due to AI data scrapping concerns.
To put it simple,
Assuming all newly added works are public, if drarry has 100 new works added (gross growth) but 50 existing works are deleted or locked during the same period, the traditional list reports 100 - 50 = 50 new works (net growth).
Now if we take into consideration newly added private works. Say, there are a total of 100 new works added (gross growth) and 20 of them are private works. During the same period, 50 existing works are deleted or locked. The traditional list reports 100 - 20 - 50 = 30 new works (net growth).
In both scenarios, the alternative list reports 100 (gross growth).
Most people look at the traditional list and interpret the “New Works” column as gross growth, while it’s in fact net growth. Net growth can even be negative if there are more fics deleted or locked than there are fics added during the same period. I think the appropriate column label should have been “Net New Public Works”.
