I have used 10m 1d as my learning steps in the past, with a 10m relearning step, but from recommendations by Anking and other sources, I recently adopted solely a 10m learning step. Apparently this is also what is recommended with FSRS?
An issue I have been having is that todays “new” cards, are already “dues” tomorrow, which feels fraudulent. They are never really in the learning stage, apart from my first day, and I have read about how multiple same-day reviews (pretty frequent for me) are not good for long term memory formation.
Anyone have any thoughts on this, and is my thought process flawed? Thank you!
When you only have one learning step, a card only needs one Good grade to graduate from Learn to Review. There’s nothing “fraudulent” about it – the card graduates when you get it right. If you want to have multiple short learning steps, you can. There’s no prohibition against it – it’s just usually not necessary.
Adding a “1d” learning step is something different though. That prolongs the Learn stage unnecessarily and postpones the time when FSRS can take over the scheduling. Some folks had reasons for doing that under the default SM-2 algorithm, but with FSRS it just makes you fight with the algorithm.
You don’t mention having any issues with the length of the intervals your cards are getting when they graduate to Review. So you should just let FSRS take over the scheduling after the first day.
Thank you for the reply and insight. I assume that adding multiple learning steps then, maybe a 2d and 4d, would essentially do the same thing you mentioned with postponing the time when FSRS can take over the schedule?
With my previous settings, I would usually have a decent amount of Learns that I would tackle in an early day study session as they were typically harder (or more time consuming) that Due cards, and would then work on the Dues throughout the day. If I wanted to adopt a similar study methodology with the new single learning step, would a setting like “Ascending retrievability” for the Dues be effective? Another way to phrase my question: on average, would more recently graduated cards be earlier in the retrievability rankings compared to older cards that have been in circulation (and have longer intervals)?
That’s correct. I was referring to any steps 1d or longer, and that also applies to any steps/series of steps that are too long to complete in one day (e.g. 23h, 12h, etc.).
Hmmm… I wouldn’t use Retrievability for that. Retrievability can change quickly a variety of reasons, so the set of cards that would bring to the front would not necessarily be the cards you’re looking for. It sounds like you’d want “Ascending Intervals.”