Is it possible to break Spaced Repetition?

I am new to WaniKani and to Spaced Repetition in general. So from what I understand, there is a Science behind putting things you’ve learned aside and coming back to them just before you forget them. It’s supposed to burn it into your long-term memory better, right?

Here’s my question, what if I keep hitting the Extra Study link in WK during my dead time, and I also utilize a WaniKani deck in Anki? Will that mess up the “just before you forget it” part, and sabotage my own learning?

8 Likes

Well you’re just creating more work for yourself.

SRS is just a way of getting something into memory with the shortest amount of work.

(You could review everything you know everyday and then you’d never forget but that’s just far too much work)

Time spent on extra study is, theoretically, wasted. Best spend that time studying something more fruitful like grammar or vocab, or just consuming native material.

8 Likes

Its wasted in theory, not in practice.

In theory, you would never get an SRS review wrong because you review them right before you forget them, but here we are.

If you have periods where you’re doing nothing and probably wouldn’t do other study, then its a whole lot better than nothing imo.

12 Likes

Short answer: no. The more repetition you do, the easier it is to remember something.

Long answer, it’s a bit more complicated.

SRS- is more about time-management really.

Traditionally, you do tons of repetition studying, and then a test. Then years later, you either remember some of that or you’ve completely forgotten. But, it’s a system that requires a lot of time devoted to intentional studying.

SRS let’s you do other things with your time. You don’t have to spend overly much time on the lessons stage, just some. And then you’ll simply spend below a second on answering a question about what an item mean or how it’s read. Each item doesn’t get more than maybe 10 seconds total of focus time until you’ve burnt the specific item months later. So very little active studying time. That’s the neat part of SRS.

So, while you can spend extra time studying, you might better spend that time on other parts of Japanese that requires a more complex memorization approach: such as for grammar, listening comprehension and language production.

7 Likes

Sure but the point of using an SRS is to do the bare minimum so you can free up your time to study other things

5 Likes

And the reality is there’s some time people would be willing to spend on extra study that they wouldn’t be willing to spend on other things.

It’s very easy to say “just read instead” but that misses the reality that if I’m at the mall waiting for my buddy, I’m not going to pull out my phone and start reading as a beginner (or even advanced), learn new words, or put in earphones for listening practice. What I might do though is pull out my phone and do 20 reviews or 20 extra study items while I wait. Now we are deciding between “nothing” or “something” and that something definitely isn’t a waste or a poor choice. It’s just how people work.

16 Likes

I guess I could have clarified better in my original post. The thing is, I’m retired and I spend a lot of time indoors with my wife. When I do get out, it’s for the gym, cycling, or solo-camping, but I don’t do any studying outdoors. I was running out of Netflix movies to watch, so I started adding Anki and Extra Study while waiting for my reviews to come available. I just didn’t know if not waiting till I almost forgot a Kanji, and studying it during down time, would break the Science behind it and not burn it in as well as if I had just waited. But what I’m gathering from all of these replies is that it won’t break the writing to long-term memory, but that doing the extra work is not necessary. If that’s the case, maybe I’ll just drop the Anki and keep doing the Extra Credit until WaniKani no longer feels slow to me. I’m guessing from all the stories that it won’t take much longer for me to start feeling overwhelmed and having too many reviews.

I am also reading some simple stories in the JReader and Satori Reader apps. But that’s really tiring.

4 Likes

So study some vocab :person_shrugging:

2 Likes

Short answer, yes that’s right. And I’d also agree drop the WK deck in Anki. Instead, study other vocab cards you make in Anki, or some cards you make on Satori.

Once you’re farther in and start to miss things that are coming back for review at long intervals and you miss them, the “Recent Mistakes” feature on WK would be helpful.

My understanding is the “extra study” feature is if you were starting to struggle and mix up lots of things and needed more practice generally. Or if you’re just feeling like studying something, like Vanilla said.

have fun and see you around :slight_smile:

5 Likes

No, I don’t believe there’s any danger of that. Specifically, I don’t think there’s any science to support that idea. (Anyone: If I’m wrong, I’d be interested to know, so if there is science to support it, please reply if you have a reference to something.)

What SRS is about – as far as I understand – is the efficiency of learning. In other words: How much stuff can you learn in a given amount of study time? (Which can potentially be an important factor, especially if there are a lot of items to learn, such as all these thousands of kanji to gain fluency in written Japanese.)

So, I believe the research that supports SRS says something like this:

  1. If you want to maximize the amount of stuff you can learn in a certain amount of study time, then you can learn the most if you spend as little time as possible on each thing you’re learning. Example with made-up, fictional numbers:. If you use method A, which requires on average 20 hours for each item learned, then in the long run, on average, if you only spend 1200 hours of time studying, you will learn about (1200 hours) / (20 hours per item) = 60 items. But if you use method B which only requires 10 hours per item, then in those same 1200 hours of study, you’ll learn about (1200 hours) / (10 hours per item) = 120 items. Therefore, method B is more efficient, and thus a better use of your time. This is because you only spend 10 hours per item, rather than 20.
    In other words, the less time spent on each item, the better.
  2. Compared to other methods, SRS – particularly when the timings are spaced so that they occur right about the time when you’re on the verge of forgetting – uses the least amount of time per item, in order to learn that item.
    (Note: I have not looked at the research myself, by the way, I’m just stating it based on my understanding from reading about the research from other people, such as WaniKani, who you could argue have a possible profit motive, but also people such as the people behind the free software Anki, who I doubt have any motive to make stuff up about it. Just saying to take what I’m saying with a grain of salt, though I have little reason to doubt it myself.)
  3. Therefore, if you want to learn the most stuff in the least amount of time, it’s best to use a method that uses the least amount of time per item, and one of those methods is SRS, specifically where you time it based on ‘just about to forget’ timing.

But if you spend more time studying an item than this minimum, I very much doubt that it will harm your learning of that item. It will just mean that you spend more time per item. And thus, in a given amount of time, you will end up on average learning fewer items. But you’ll still learn the items!

So, if you would minimally spend (made up numbers) a grand total (on average) of 10 hours per item using just the ‘just before forgetting’ rule, then in 1200 hours you could learn 120 items. But if you augment that with an additional 2 hours per item doing extra reviews (e.g. the Extra Study feature), then instead of each item taking 10 hours, it will take 10 + 2 = 12 hours, and as a result within those same 1200 hours, you’ll only be able to learn 1200 / 12 = 100 item. Since 100 is less than 120, then on average you are being less efficient with your time. But you’ll still eventually learn all the items. It will just take a bit longer.


In the end, IMHO, being maximally efficient with your time spent learning Kanji (and Japanese in general) is one consideration. But it’s not the only consideration. And how important efficiency is will depend from person to person.

Personally, I spend more than the minimal time learning kanji because I happen to be a bit of a perfectionist (which is not actually a good thing, IMO, it’s just part of my psychological make-up), and so I feel a certain ‘uneasiness’ with learning only to the point that I’m ‘just about to forget’. My perfectionism means that I feel like, “Well, I don’t want to be on the verge of forgetting. I ‘need’ to know it solidly,” even if that means it takes me on average longer to learn the items.

Indeed, that’s a big reason why I got the lifetime subscription from the get go. I knew at the beginning that I would end up wanting to use WaniKani for at least a few years. And these days I know that actually it will be for several years. But since I invested in lifetime, I don’t have to think about being the most ‘efficient’ in learning. I can instead take my time.

It even got to the point where I joined in on the forum thread celebrating ‘going slow and steady’ at :durtle_hello: Let’s Durtle the Scenic Route :turtle:.

So, don’t worry about ‘sabotaging’ your own learning by using something like Extra Study. At worst, you’ll spend a bit more time studying. But on the other hand, you’ll probably feel more confident in what you’ve learned as you go along. It really just depends on what your own personal priorities are for learning Japanese. Fast-fast-fast? Slow-and-steady? Or somewhere in between?

2 Likes

Every exposure is learning. Children have tons of things they learned after one exposure and things the forget even with multiple exposures. WaniKani is out of context, aka out of emotion relevance, and sometimes boring too. SRS is just attempting to activate the emotions of forgetting. And making you have to work harder, so you brain says dang this was important and I was about to delete it.

I personally wouldn’t add more SRS which Anki is. I would add more context. Try Satori reader or Graded readers or something to practice seeing vocab in context that will be much more beneficial, but to answer your question no extra study will not break SRS but it also won’t help all that much.

6 Likes

MegaZeroX researched this very topic to find out what science has to say about it. TL;DR the more often a see an item the more likely you will remember it on the long term You just have to avoid looking at it just before doing your reviews.

7 Likes

The science behind SRS does not set intervals with mathematical accuracy. You’ll find that between various SRS solutions there are very significant discrepancies in how the intervals are handled.

What matters in practice for SRS is that when you get something wrong it comes back sooner, and if you get it right it comes back later. The former is important for memorization, the latter not to get overwhelmed by reviews.

The whole idea that the system is going to present you with the item “just as you’re about to forget it” is quite a grandiose statement to make for what is a very simple system of hardcoded timers. It’s the idea behind it of course, but no you can’t really “break it” in a meaningful way, it’s not precise enough for that.

And hopefully in the near future you’ll be consuming native content on a daily basis too, at that point you won’t be able to control when stuff does or doesn’t comes back anyway. So don’t overthink it, the “science” of SRS is not that important.

4 Likes

Yeah having been using SRS daily for the better part of a decade now (for various languages) the idea that you can review an item too often is kind of silly to me. I mean after all the words and constructs we remember the best are usually those you encounter all the time in the language, I’m sure most of us don’t have a lot of difficulty remembering what 来る or する mean…

The reason you want the intervals to grow in SRS is mainly so that your review load doesn’t become unmanageable, not because getting the item presented to you again “too early” would have negative effects. It’s a way to allocate your review time more efficiently, that’s all.

5 Likes

I think all the points here are very good (efficiency of study and time spent, doesn’t hurt to study more if you want to).

What I’ll add is that besides efficiency, pushing the limit on spacing between generating a memory helps to reconsolidate that memory more than if you didn’t do that kind of spacing. More effortful recall leads to more recall strength and recall speed. Spacing helps you to increase effort in a way that’s better for learning with the added effect of efficiency over time.

But, say you’re living in Japan, and you see the same kanji/vocab over and over. You’re going to learn those too. The thing to note is that while spacing is (potentially) smaller, there’s still spacing. The reason both methods work is because you’re generating the memory (recalling it on your own) — really it’s the act of generation that allows you to reconsolidate memory and make it stronger/faster.

I don’t think you’re doing this, but the only suggestion would be to avoid rote, repetitive study. Once you generate a memory, and pull it into your working memory, every time you “recall” that memory again (before it leaves working memory), it only gets pulled from working memory. That’s why looking at something and repeating the answer doesn’t really do much. It might do something the first time you generated it, but each repetition after doesn’t do much of anything at all. You have to create space either with time, or by studying / doing something else to clear your working memory, so you can re-generate the memory again. That said, the better you know something, the less change and progress that occurs when you generate the memory. But, it’s also less likely to decay. Spaced repetition helps with all that, too. That said, I doubt any of that will hurt your studies. Like others have said, it just doesn’t do as much in terms of the time spent. If you’re enjoying it, though, why not!

As for SRS timings themselves, it does really vary. The general rule of thumb is something like:

If you want to remember something for hours, recall it in minutes.
If you want to remember something for a day, recall it in hours.
If you want to remember something for days, recall it in ~1 day.
If you want to remember something for a week, recall it in days.
If you want to remember something for weeks or a month, recall it in about a week.
If you want to remember something for months, recall it in ~1 month.
If you want to remember something for a year, recall it in months.

That said, this doesn’t take into consideration context and connections, i.e. learning related things / connected things, which does happen a lot in language study.

tl;dr you don’t necessarily have to, but study more if you want!

20 Likes

Okay, who hacked Koichi’s account

This is scaring me

10 Likes

I’m just a bot running on ChatGPT

13 Likes

THE MACHINES HAVE RISEN and their writing is very subpar

3 Likes

In those instances I think another factor comes into play, and in my opinion it’s vastly more effective than even the best SRS implementation: you can tie the vocabulary to meaningful real-world experiences.

I live in a foreign country that doesn’t speak my mother tongue and the amount of vocabulary I just passively integrated by sheer exposure in day-to-day life is astounding. Like remembering the names of fruits and plants and animals is always a source of frustration for me through SRS, I end up confusing them all, but if you actually have to do grocery shopping in a place where your target language is spoken you’ll surely end up remembering these words passively without much effort out of sheer practical necessity.

8 Likes

Please write 手 to make sure.

2 Likes