Hello! I’m not one to really make post honestly but I need some help/clarification. So I have been on WaniKani for a few months now, and when I hit level 11 I had a lot going on and I thought that I wasn’t performing as well as I would like to with current review cycles. Therefore, I stopped picking up new lessons to improve my performance overall, at least that was the idea.
However, once I stopped doing new lessons my performance didn’t really improve, while doing the reviews before stopping lessons I would average about 88% (referring to the percentage that appears in the top right corner). I also would appreciate if someone could explain to me how this percentage is calculated, I’m not really too sure how it works so I’d really appreciate it. But anyhow at the end of the reviews when you get the review summary my answered correctly percentage is usually about an average of 75% to 80%.
I thought that if I stopped doing lessons I would naturally improve at both of these, nothing wild but just a bit better and then I could pick up lessons again and try to have a more balance relationship with lessons and reviews. But I essentially ended up getting about the same results, or even worse on some occasions. I was wondering if there’s some reason that I’m just missing completely of why this happens to be the case. It’s alright if there isn’t any particular reason but I’d also appreciate some tips or advice on how to improve reviewing, or any techniques that I could employ while reviewing. I’m slowly starting to pick up lessons again but I would like just be sure I’m not missing anything important.
When do you do your reviews? For example if you just learned something it should be back in 4 hours, but if you’re like me I will sometimes do my lessons at night and do my reviews like 14 hours later and it should be no surprise that I got everything I just learned wrong, as well as the usual things people will get wrong. That brings your percentage down.
Basically if you get the reading or definition wrong, it doesn’t matter if it was only one, the whole thing is wrong, I believe.
If you do your reviews at the correct time this should help, if you already are doing so, then I can’t say you’re missing something, it’s just that you have quite a few things that aren’t sticking and you need to carefully sit down and read the mnemonics or make your own.
you might just be struggling with specific items. Then it’s best to do some specific cramming to improve your accuracy, like using the Self-Study Quiz, the Item Inspector-script or the Leech training script. You don’t really have to pause lessons for this, that’s more a matter of personal preference.
I’d say 88% is quite good enough accuracy to not be concerned. There’s a cut-off minimum percentage (I forget how much) that when you go under it, you’re moving more reviews down the SRS ladder, rather than moving them up… That means not progressing. Someone else might be able to say where it is, but I don’t think you’re in any danger with those accuracies, even if 75% is on the low side.
So, maybe just do a bit of extra studying of the items you keep failing. There are bound to be those as you move through the levels, while others are easier. I wouldn’t think too hard about it.
I might suggest spending some extra time with everything you get wrong if you aren’t doing so already, like ekg said. SRS is great, but I think sometimes people can get caught up thinking that things will eventually stick because they’re using SRS. As a result, they end up forgetting, or just choose not to review the words they get wrong. I’m guilty of once doing this, but putting in the extra time into relearning the thing you couldn’t remember is really beneficial. Also, the way to go about spending extra time with what you’re getting wrong can be done many different ways, but personally I think that studying it right when you get it wrong is preferable. The reason is because you have fresh in your head what your thought process was when you were trying to recall it. Therefore, when you’re studying the word you got wrong you can incorporate that thought process to try and link that with the meaning/reading.
One other thing - if it’s taking you a long time to remember the meaning/reading of a certain review, I might suggest getting it wrong on purpose. The reason behind this is because the longer it takes for you to remember, the higher the chances are that you’re eventually going to start losing focus. Once you start losing your focus, it’s going to be harder to remember the rest of your reviews. Secondly, say it takes you 30 seconds to remember the reading of a kanji and it gets passed to a 2 week interval. The likelihood that you remember the meaning of that kanji in 2 weeks are pretty low if you don’t end up seeing it in the wild. By purposely sending these reviews into the earlier stages of SRS, you’re seeing them more and getting more reps which helps solidify the memory. In return, this will improve accuracy in the long run and make reviews go by faster.
I would recommend you this userscript. If it really bothers you and interferes with your studies, it’s better to look at numbers after the review session ended.
As to why two percent values are different. It’s because they are different values, wild huh. The top right corner is a ratio of failed questions to all questions asked, so the meaning and reading count as two different questions. The number on the post-review screen is a ratio of downgraded items to all asked items. And 80% is totally all-right.
For new items I usually tried to get them done as they come up. There are times when they slip my mind and I don’t do them until 3 or 4 hours after they are due. But with older reviews, which is mostly what I have now, most of the time I do them in one go in the mornings. My SRS mostly at the point where I usually get all the reviews them around a three hour frame in the mornings.
@ekg@_josh I think I will certainly spent more time and just picking out the things I’m struggling with. I usually always review the mnemonics as well while doing the reviews and I thought that would keep me on track but it seems that’s not doing it. I have yet to dabble with scripts so I’ll have to do some reading on setting that up.
I’ll keep in mind the accuracy, I was a bit concerned because I think I spent enough time not doing lessons to where my apprentice number should be very small or none at all and I keep having periods where I do really poorly and it comes back up. I’ll also certainly keep in mind of how long it takes me to recall a kanji, is not very often but sometimes I do stay for a while on some of the readings that I feel I should know for sure, though I was under the impression that doing that wasn’t necessarily bad as I thought the harder you try to recall something the more importance the brain put’s on remembering the next time. Any thoughts on that? However I do think that you are correct that there are times where getting it wrong on purpose is useful, to reinforce the reading sooner before going to longer intervals.
Thank you so much for you comments, really appreciate it.
Yeah so I think this is true; however, I’ve passed many things where it took me a long time to recall the meaning/reading and when it came up the next time it was pretty much the same process 50% of the time. It was either I couldn’t remember it at all, or 25/25 whether I got it fairly quickly or it took me a long time to recall it. It’s up to you what you want to do, but I don’t think there’s much harm in purposely failing reviews that take a long time in the early stages since seeing them a lot in those early stages will help improve speed in recalling them later on. What I’d do personally is fail a review after a certain amount of time (whatever I feel like is too long), but still try and recall it, and if I feel like there’s no way I can remember it, then I’d look at the meaning/reading.
Wanikani is an SRS (spaced repetition system), and the way many SRS programs are designed, you’re not meant to get 100% success rate every time. It might be counter-intuitive if you’re used to traditional classroom learning, but getting a few things wrong every session actually turns out to be more efficient in the long run so a lot of SRS programs are actually designed with that in mind. I don’t know what the designed targeted rate is in Wanikani, but I know that in Anki at least it’s around like 80-90%, so I’d guess you’re probably doing about as well as the average user here. Personally I don’t really pay attention to my percentage since I know that “failure” is actually a very important part of the learning process. So if you ask me, I probably wouldn’t worry about it too much, you’re doing fine.
It doesn’t really bother me, but it could be helpful to use that script to not distract from focusing on the reviews. I’ll consider adding it when I learned how to add userscripts. Also thank you for explaining the differences I was certain that it used different values but was curious as to how it was making the calculation. Really appreciate it.
Wow I think you have a very healthy view in terms of your learning process, I certainly agree and think I have probably to some extent I unnecessarily slowed down my progress. I can certainly implement the mindset that failing is just part of the long process here.
It’s true that if you try to cram too many ‘new’ items into your brain, then your accuracy tends to go down. You’ve already taken the right action: stop doing lessons for a while to get the apprentice count down a little and then resume. You should see an improvement in accuracy… however that would be visible after you resume the lessons. Why? As your apprentice count drops, your daily reviews will (percentage wise) start consisting of more and more reviews from higher stages. Accuracy tends to be lower on those, and as a result you see your accuracy going down rather than up. This is not strange. Your accuracy will rebound after returning to lessons, giving your review sessions back the original mix of new and older items.
Note also that accuracy tends to drop over time / levels. Like people said, I wouldn’t worry if you regularly score 88’s.
That’s interesting, it didn’t occur to me. Thank you for commenting, I appreciate it. I will definitely keep it in mind as I start doing lessons again.
In addition to the effect that @rwesterhof mentioned, the SRS timings conspire against high accuracy numbers in atleast one other way. During a lesson break, easy items that would raise the accuracy will become rare relative to harder ones, since they rise through to Master and beyond, while tougher items remain at Guru and Apprentice. Without new lessons, the SRS effectively refines your review pile into a progressively more difficult set of items.
Thanks for the link,didnt know about this script. Those percentage arent helping me at all, and at best, they crush my sanity and fill me with despair. I appreciate it.
Your 75 to 80% average accuracy sounds great (to me). Mine’s in the 60s, and I feel I’ve done well if I hit the 70s : )
But I also have a mindset that I can get into (sometimes), which is that reviews are lessons in disguise. When I get a review correct I feel good of course. But when I get it wrong, that’s the gold! It’s my chance to relearn it.
When I think like this, % accuracy seems less important, as the lower it is, the more of these bonus lessons (that I obviously need) I get.
I also like that instead of thinking I should get the darn reviews out of the way before I get to the important stuff (the lessons), I think (as I start the reviews), wooh hoo, it’s lesson time!
WaniKani. All lesson. All the time. (Except when getting the right answers to reviews?)
