Repeated/skipped learning steps

The FAQ on repeated/skipped learning steps says –

This issue no longer exists if you change the learning steps in Anki 2.1.55+.

That appears to be a reference to the 2.1.55 fix by @Rumo

  • You should no longer see strange behavior when changing learning steps while cards are in learning.

But this still seems to happen quite often – examples, from here or Reddit (and long enough after the Dec 2022 release that it seems unlikely they are still pre-2.1.55) –

I was in the process of writing this up as a fix needed for the FAQ, but how much effort anyone should put into updating the FAQ (and how to word it) depends a bit on what’s going on as far as development. I don’t see anything in the issue tracker about it.

Is this being investigated as a recurrence of what was happening pre-2.1.55? Or as a new issue that came up since then? Or is it unfixable?

1 Like

I have faced this recently when changing from 1m 5m 10m to 1m 10m (and doing other tweaks).

I think we should at least then list it as a known bug and not mark it off as fixed, regardless of other things.

I suspect these users haven’t been consistently syncing at the start and end of each session. A conflict could cause cards to be reverted to an earlier state, and that earlier state may have been appropriate for a different number of learning steps.

But it could also be that the fix isn’t working properly. Can anyone provide a minimum set of steps that triggers the issue, where syncing is not involved?

I looked through these examples again and many of the users were able to confirm strict syncing habits – and at the same time confirm they had definitely been changing their learning steps.

I think one of the reasons this has been so prevalent lately is that folks are removing their multi-day steps when they turn on FSRS, and

  1. If you have multi-day steps, you are almost guaranteed to have some cards in Learn at any given moment – probably a lot of them.
  2. A lot of the issues seem to arise around changing the number of steps (as opposed to just the length of the steps), so removing those multi-day steps is more prone to trigger the issue.
  3. When folks are turning on FSRS they are often keeping a closer-than-usual eye on the button-intervals.

Hopefully someone can identify a reproducible case. I’ll keep an eye out.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.