I might be misremembering, but I don’t think Anki used to behave this way, at least not recently.
Here’s what I’m seeing:
I create a custom deck that pulls cards from one of my subdecks. That subdeck has its own deck-specific Desired Retention (DR) that’s different from the preset’s default DR. But when I study the custom deck, the cards are being scheduled according to the preset’s default DR, not the subdeck’s own DR.
I expected the custom deck to respect the DR that’s set on the subdeck, since that’s how it behaves when reviewing directly from that deck.
Has anyone else noticed this? I could have sworn it wasn’t doing this even a week ago.
If you have a deck which has a deck-specific desired retention, and you study a card from that deck in a custom study, deck specific desired retention is ignored and the preset desired retention is used (I’m using latest version of Ankidroid, which claims to use Anki codebase for backend operations).
Steps to reproduce the bug:
Make a deck with a preset, with desired retention 90%
Set the deck specific desired retention 80%
Make a card in the deck
Study it in the original deck; deck specific desired retention is applied
Study the same card in a filtered deck; the preset desired retention is applied
Expected behaviour:
Intervals should be computed using the deck specific desired retention. If I put low priority cards into deck A with lower desired retention, I expect the card to remain low-priority even if studied in a filtered deck.
Hi, it looks like the bug has been fixed in: https://github.com/ankitects/anki/pull/4413, but Ankidroid seems to not apply the updates, or it has applied the changes but changes have not been released yet. I moved the category to Ankidroid and I’ll close.
I’m not sure if this is a known thing, but it is an interesting quirk I found while fiddling. I have a deck at 85% retention with a subdeck retention at 90%, then created a filtered deck consisting of cards only from the subdeck. These subdeck cards are scheduled according to the main deck 85% retention, not to the subdeck 90% retention. I tested this with extremes to ensure this was the case and it wasn’t just fuzz. By cranking up retention to 99%, intervals for cards belonging to the subdeck did not change while studied in the filtered deck. They dramatically decreased when studied only in the subdeck.
I have not explored this fully, and there is crossover with an addon that could be contributing. I’m using the Ankihub smart search to create decks specific to exam blocks out of the AnKing deck. This may be the culprit, but because this lies at the intersection of so many potential causes, I’d like make the issue known but pass it off (how very self-serving of me–a saint).
Edit: I should note that the scheduling does not change after the deck returns to the subdeck/the filtered deck is emptied. Seems like due to the hierarchy system of decks, cards are only subject to main deck scheduling (?)
Version 25.09.2 (3890e12c)
Python 3.13.5 Qt 6.9.1 Chromium 122
I just saw someone who noticed the same issue a few months ago (can’t link it but tilted: “Custom Decks Ignoring Subdeck’s Desired Retention?” in Oct 2025). Looks like there’s an approved github change from Nov linked on that thread, so might an issue that is sit on until the next Anki version for now (?)