Who cares about my opinion? 
Everyone can have their own method, and people should be given the opportunity to choose what they want.
I don’t use the FSRS, although it’s good in many ways. Well, if you study regularly… but what if you don’t?
Sometimes I’m a proponent of the idea that you either know it or you don’t. That is, there should be two grades: you either cut the red wire or the blue one, and maybe there won’t be an explosion… but like in medicine, you can figure it out yourself
Sometimes you can’t make mistakes, so even if you wanted an A, we’d sometimes get a grade lowered immediately for just one mistake—it’s very stressful.
But there are many fields of study that don’t require such strictness, but management personnel should still have excellent knowledge, without any “oh, I almost got it right… so what if the rocket is now flying toward the Sun” attitude. So, the A students are chosen to become managers, those who simply do well will become engineers, and those who get satisfactory grades will simply obey and listen to their superiors.
How should you grade cards if you have to pass an exam?
But how do you approach knowledge? Some teachers aren’t strict, saying if someone forgets something, just answer a couple of other questions. They seem to know, so they give them good grades. If a student agrees to this, why bother studying with 90-95% accuracy if it’s enough to review all the cards, discard the most difficult ones, and focus on the easy ones first?
As you can see, there are different strategies. You’ve also noticed that you need something different, and then give them another button… someone else will come along and give them another button, and so on. Well, if you want flexibility, you should let the user set a time limit for each button to determine when they should review a card. You say this exists… well, yes, you can set a deadline, and postpone a card until tomorrow. This has already been done, but it doesn’t address the convenience of the buttons.
Look at the available add-ons. Perhaps you could write something yourself, then try it all out, test it for a while, and then you can confidently say how convenient and necessary it is.
And further:
So, how do I proceed or advise proceeding before the exam? So, you have a deck or a subject, it doesn’t matter. You’re given a list of questions, and let all the questions be cards.
You review them and divide them into three categories:
- I don’t know or this is very important (about 60%)
- I remember something and they might ask about it (about 30%)
- I know, I remember, or this is not important at all and I can at least somehow answer it on the exam (about 10%)
During the study process, you might discover cards that simply don’t want to be remembered, so they’re put into a separate category (subdeck, deck) “for processing.”
And it makes sense to start studying with the first category, the ones you don’t know or the ones that are very important.
But this isn’t always the case. Ideally, you create decks based on what you already know, but many people don’t create their own decks, but study something directly from the anka. So the algorithm will be different: in the first step, the filtered deck is quickly reviewed in the order laid out by the author. In the second step, each card is studied in the same order (if you know a card, then you obviously don’t study it). In the third step, the cards you remember are marked, say, with a green flag. In the fourth step, you mark the cards you were able to remember the second time in a different color. Finally, in another step, you can select the cards that you’re having trouble with (don’t study from the cards, go read lectures and books). This way, you identify the varying difficulty of the cards and review these cards repeatedly, memorizing them.
Caution!
A typical mistake: you’re given 300,000 cards to know for an exam, and you simply start going through them all in order, thinking they’re all connected. Competent authors divide their decks into subdecks based on small topics and master one topic 100% before moving on to another.
The solution: take an anki and have it toss you a card after 10 minutes. It seems to work, but you have to consider how much new information (cards) you’ll receive during that time. If you look at 300 new cards and think the algorithm will work, you’re mistaken, as you always have to consider the amount of new information (not cards, but information blocks, so for one person a card might be nothing new, they know everything, while for another it’s all new and they’re just fooling themselves, thinking they didn’t remember what they remembered; they’ve learned to recognize it). The algorithm will only be able to figure this out if you repeatedly fail to remember a given card, although you could identify such cards earlier and set specific parameters for that subdeck: for simple cards, you can do it after 20 minutes, for complex cards, maybe even after 3 minutes. I can’t tell you how to do this with FSRS, but people seem to be satisfied over long periods of time, and those who are dissatisfied don’t post on the forum, but instead go on to find their ideal program or method.
So, time certainly plays a role in spaced repetition, but so does information volume… I haven’t heard anything about that, but teachers have known and talked about it for a long time: learn 3-5 words in phrases and in a circle, and be sure to repeat them out loud. After you’ve learned these phrases, move on to the next ones, then combine those 5 with the previous 5 and repeat again. Learn the entire text this way.