I switched from SM-2 to FSRS. I miss the fact that I got to review all learned cards the next day. In FSRS, a learned card that I answered correctly both times, would be scheduled for review in a few days. I understand that this may be optimal scheduling. But I really like to cement yesterday’s (re-)learned material today, not in a few days; however inefficient that may be. After this point, or if I press easy, the algorithm can do its thing. Is there a way to make FSRS to always have 1d for good for cards (re-)learned today?
Solutions that I thought of:
Raising retention. I have to raise it to 0.93 to get the 1d interval, that’s too high. It would affect the learned cards too much.
“multi day” learning phase. ie setting the last learning step 1d. Not really compatible with FSRS according to the manual:
(Re)learning steps of 1+ days are not recommended when using FSRS.
Currently, I just start decks with an SM-2 preset and change the preset to an FSRS one once a deck of say 100cards is done. But is there better way?
I’m using: Version 24.04 (429bc9e1)
Edit: Apparently the problem is not clear:
Problem: With FSRS the first interval after having learned a card is at 3-4 days.
Desired solution: set that interval to 1 day without affecting retention rate for old learned cards.
I didn’t really understand your problem. Do you need the first interval for new cards in 1 day?
Yes. That is the problem. By default (retention at 0.9), the first interval for new cars is at 3-4 days. I want it to 1d, as it was with SM2.
It is 3-4 if you answer the “good” card. For “again” and “hard”, you will most likely get about 1 day.
But if you want to achieve this without changing your habits, then you can try to do it through a filtered deck or through naming FSRS parameters.
It is better to use a filtered deck. Other options are not recommended.
it should look like this
(introduced:2-introduced:1) deck:name
But “-introduced:1” not supported
Therefore, use such a search query. And do it before you study the new cards of today.
introduced:2 deck:name
https://docs.ankiweb.net/searching.html#first-answered
Maybe someone has a better solution?