This is something I’ve wondered about…I know that the simplest answer is to set new cards to zero and then just persist doing daily reviews, letting the number of reviews go down naturally over time. However, if someone is using anki very heavily and then for whatever has to dramatically change their daily anki usage, sometimes there simply isn’t time. So I was wondering if there was a sort of generally accepted best practice for this sort of situation…of course short of the above plan (set to zero, continue reviews, let it go down naturally) I’m pretty sure any solution will sort of be a suversion of SRS/anki, so I’m curious which solution is “least bad.” Esp given our lovely developer posts here
Here are the options I’ve thought of thus far…I’d be very open to any other options, or which of the following would be “the most principled” (of course saying something is “least bad” means you have to justify what is bad, which involves tradeoffs etc)
set max review limits. calculate about how many daily reviews you can handle and set accordingly. that said, I have no idea how anki chooses what cards to show in this case…is it arbitrary? is there some principle at work? it seems like a pretty severe subversion of SRS, but I guess all the options are!
time box anki usage, and simple let the numbers be what they are gonna be. the downside here I guess is mainly a psychological one compared to 1…behind the scenes it’s likely the same? (though I have no idea how anki chooses which reviews to prioritize when you don’t simply do all of them) but you would see all of your total reviews potentially skyrocketing over time. sort of tough psychologically
“anki bankruptcy.” basically, set a large swatch of cards to new. this is the most in line with anki’s philosophy…kind of? except you’re not leveraging your anki history at all, which seems undesirable? that is to say, a “principled” approach feels like it would be able to make use of all of your review and scheduling data somehow, instead of just throwing it all away
I know there are plugins that move around reviews…I’m not sure what they do, but there’s definitely a chance that someone has already make a plugin that does something like this. the plugins I’ve seen are more oriented around temporary vacations versus a permanent step change, but I’d be curious if anyone has tackled this directly
again, I’m totally open to thoughts on this. thus far I’m in an ok position and can sort of work on organically getting my reviews down, but there’s a non-zero chance that if life circumstances change I’d need to dramatically reduce anki usage…which has made me wonder what approach I should take! welcome your thoughts and suggestions!
This work just great to me, when using scheduler 3:
The new count is now capped by the review count. If you have limits of 200 reviews and 20 new cards, and 190 reviews are due, only 10 new cards will be introduced. The Anki v3 scheduler - Frequently Asked Questions
It really depends on your priorities. If you build up a growing backlog of pending reviews, you’re going to start forgetting more of that material, and you need to decide whether you want to try keep maintaining all of that material at a lower recall rate, or whether you want to prioritize some of it, eg certain decks being more important than others, certain cards being more important than others, certain card states being more important than others (wanting to focus on recently introduced cards over old cards or vice versa), etc.
you nailed my concern…if you have X pending reviews doing it every day and all of a sudden set a cap of X/4…the remaning 3X/4 will just build up, and sort of destroy your recall.
what I guess I’m most interested is whether or not there is some way to leverage card history to achieve at least a more principled balance. not sure what that would look like but as a start, anything that is very mature seems “safe.” you can also project out what future reviews will likely look like, then sort of figure out a buffer…but beyond that I don’t know.
I guess the only really principled thing to do is to throw out cards. hrm…
Anki shows the most-delayed cards first by default, and recent versions have an option in the v3 scheduler to sort by relative overdueness, based on each card’s interval. You may want to consider suspending or deleting less important material in addition to that.
Set longer learning steps and make new interval after lapse something else than zero. I have the former at 1d 6d and latter about 40%. Changing the new interval to something sensible is one of the most important things to do, imo. Also consider automatically suspending cards with a low leech count, helps getting those cards that just wont stick from wasting your time.
Been adding on average about 15 new cards a day for more than a year and don’t usually go much over 120 reviews a day. Even with a pretty average retention.
