Learning Steps Not Working as Expected

[milkywayapple]

Hi everyone,

I’m having an issue with my Anki deck where the scheduling for new cards doesn’t seem to work as expected. I’ve set up my Learning Steps as 3d 7d 14d because I want new cards, when I press “Good”, to be scheduled for 3 days. My understanding is that the first step should be 3 days, but this isn’t happening.

When I click “Good” on a new card, instead of being scheduled in 3 days, the card is scheduled for just 1 day later. This is not what I want. I expected that pressing “Good” would schedule the card for 3 days.

Here’s a summary of my current setup:

• Learning Steps: 3d 7d 14d
• Graduating Interval: 14 days
• Easy Interval: 14 days
• Starting Ease: 2.5

What I want:

• For new cards, when I press “Good”, I’d like the next review to happen 3 days later, not 1 day.

I’ve looked through my settings and can’t seem to figure out why this is happening. I’ve attached a screenshot of my deck settings.

Has anyone else run into this issue? Do you have any suggestions for how to fix it?

Any help would be appreciated! Thanks in advance.

[Danika_Dakika]

First, some blunt advice you didn’t ask for –

1. Keeping cards in Learn for weeks is excessive. I don’t know if these Leitner-box-type intervals are your idea or someone else’s, but one of the main benefits of Anki is that you can just let the scheduling algorithm do its job. It’s also best to stick close to the defaults until you know what you’re doing with these settings. That will help you avoid errors like pointlessly setting your final step, Graduating interval, and Easy interval be all the same length.
2. Since it sounds like you’re new to Anki – you should use FSRS instead. It is an even better scheduling algorithm than the SM-2 algorithm you are currently trying to prevent from doing its job.

---

Just in case I haven’t convinced you to choose a better path, I’ll answer your question –
I think you are misunderstanding how the learning steps work, and also possibly looking at the wrong numbers.

When a New card is introduced, it is already on the 1st step. After you grade your answer Good it moves to the next step. So even with those learning steps, you wouldn’t get what you’re trying to get.

Obviously 1d isn’t any of the steps you have set, so I suspect either (a) you are looking at the learning steps for the wrong deck – perhaps the deck you are clicking to study, instead of the subdeck the cards are actually in, or (b) you changed your learning steps while you had cards in Learn, and they haven’t adjusted to the new steps yet.

If you want to dive into what is happening with any specific card, post the Card Info and we can figure it out. But I’m also happy to tell you more about how to switch to a different system of doing things!

[milkywayapple]

Thank you very much for your helpful response, Danika_Dakika!

It turns out the issue was, indeed, that the deck I was studying was assigned to the wrong options group. I’ve now corrected this and set the steps to 1d 3d 7d 14d, which works the way I originally intended (though I understand this might not be optimal). I had mistakenly assumed that the subdeck would inherit the parent deck’s options.

The reason I wanted to use a 3-day interval as the second learning step (when I press “Good”) is because I’m using Anki for a somewhat specialised use case. I’m preparing for coding interviews and studying complex coding problems from the LeetCode platform, specifically following the “Neetcode 150” list of problems. My deck consists of subdecks, each corresponding to a category of problems (e.g., “Binary Search”). I work through these categories, adding solutions and notes to the cards as I solve each problem for the first time.

The workflow I’ve been following is to revise each problem the day after I solve it (my first review). Then, for the phase where I review and try to memorise the entire category, I prefer to go through each problem in order exactly once before seeing any problem again. The list has a logical order of problems. My goal is to avoid seeing the same problem too many times at this stage, as I want to solve each problem three times in total: once when I solve it initially, the second time during the next-day review, and the third time when I go over the whole category. That’s why I wanted to use a 3-day interval – to ensure I won’t see the same problem again for a few days, by which time I will have covered the rest of the category.

I realise this approach might not align with traditional spaced repetition practices, but it’s currently helping me strike a balance between learning new material and avoiding over-revision early on. The complexity of the problems (which can take 5-20 minutes to solve) also factors into this, as each review involves writing out solutions and running the code on LeetCode, rather than simply checking the back of the card.

Once I’ve worked through the entire list of 150 problems (which may take several months), I plan to memorise the problems more thoroughly by going over the cards in a randomised order, adhering more closely to spaced repetition principles.

That said, I would welcome suggestions on how I might optimise my workflow and make better use of Anki. I’ll definitely look into FSRS, as you’ve recommended.

Thank you again for your time and advice!