Review method - deep recall or quick reaction

Generally, I think this depends on the level you’re at with the review item.

If the item I’m learning is relatively new, I’ll give myself about 30 seconds to figure it out. If I still can’t I’ll just fail it and study the item closely for a bit. If it’s an item I’ve known for a while, I’ll give myself about 10 seconds. When doing something like speaking or listening, if you can’t recall the meaning of something after 10 seconds, I’d say it’s functionally equivalent to not knowing it at all, so I consider it a fail.

In the end, I know that everyone is different, and you may be able to deeply memorize things by spending that time, while others benefit more from frequency. I’d say try both and see what works best for you.


If it feels like I know it but I can’t remember it, I’ll spend some time trying to recall it, but usually, I’ll either know it almost immediately or not recognize it and just guess (or I think I know it but am confused and get it wrong). I typically aim for 3+ reviews per minute and often average 4 or more.

I disagree with the comments that not knowing something immediately is just as bad as not knowing it at all. If you can dredge something up from your memory, that reinforces the memory, even if it takes time. The real issue is to use your judgement to see if you feel like you can actually remember it or not. There’s no point wasting time if you can’t.


This nails it on the head for me.

I rarely spend more than a few seconds before providing an answer on an item, but sometimes I think I “know” the item and spend more time trying to dredge it up (lots of research shows this is helpful).

Of course, sometimes I then dredge up an entirely incorrect answer, which doesn’t help at all!

So it’s a judgment call, but personally, I think it’s almost never worthwhile to spend more than an agonizing minute or two trying to recall an answer. Even if you get it right eventually, an “extra” few reviews from answering incorrectly won’t hurt your long-term retention.

Remember that only way the SRS knows you find something difficult is if you answer incorrectly! You’re paying for a service to show you items you find difficult more frequently — you should never feel bad about providing a wrong answer.

My go-to “I give up” answer, by the way, is “ke” or “け”: quick to type, never matches meanings, almost never matches readings, and is almost conversational (kind of like “que?” in Spanish, kinda like “huh?” in Japanese).

Lastly, if you’ll forgive me plugging my own script, you may be interested in my GanbarOmeter user script. The difficulty gauge uses a slightly more sophisticated algorithm than just “keep 100 items or less in the Apprentice stage”. It also keeps track of the number Kanji in the first two stages as well as the number of misses in recent reviews, suggesting you slow down on lessons if either gets too high.


Interesting. My “I give up” answer is just my best guess for what it is. Every once in a while, I’ll actually randomly guess it!

1 Like

That more than the other, though I admit if something seems “on the tip of my tongue” I’ll sometimes give it a minute.

But I wouldn’t say that. When I fail, I open the “item info” right there, read the correct answer, mnemonics, etc. and do a little 2-minute study BEFORE moving on to the next. It makes the review session take longer, but it IS the studying so I don’t have to go back later. I suppose you could instead make a failure list or have the Self Study Quiz script quiz you on recently failed items, that would have the same effect. But I don’t think just failing and moving on is enough exposure to make next time’s outcome any different. (For me, everyone’s mileage varies)


:frowning: That’s the worst thing that could happen, I accidentally promote something I don’t know to a higher level where I will certainly fail next time. :stuck_out_tongue:

I use ‘nn’ - might as well destroy that n key even more than wanikani already does.

1 Like

When things get shaky, I do find it beneficial to try to ponder on it for around a minute. Personally, attempting to dig it up from the depths of my brain is amazing reinforcement. I’m telling my brain that it’s important enough to recover that far back.
If I truly don’t remember after that I just type idk.


Supposedly it actually helps our memory if we struggle to remember something but ultimately manage to call it to mind (rather than just automatically giving up and looking at the answer). Because of this, I generally spend a little time/effort on tricky reviews before giving up. Often I am able to remember them, and I generally have better luck on my next review of that item! But if I look at something and immediately know that I don’t know it at all, I’ll just fail the review and spend a little time reviewing the correct answer before moving on.


This is the primary reason why I like to fail quickly.

For me this is the key. I get much more information out of a failure:

  • did I mistake it for something else? I need to review both side by side
  • did I forget the mnemonic? I need to review the mnemonic and figure out why it wasn’t memorable or whether I was remembering the wrong parts of the mnemonic.
  • did I completely blank on it? Maybe I didn’t spend enough time on the lesson. I need to spend a bit more time with it during the session review.

The key to failing fast that hasn’t been mentioned yet is going over all your failed items during the session review. I fail fast so I can spend more time with each item on this step.


Thanks everyone - this has been very helpful to read the different approaches. I say different but I sense the majority are ‘fast failures’ but some are then more inclined to delve into the failure there and then. I am going to give that a try for a few weeks and see how I get on. There is still immense satisfaction when I do dredge something up but some of the review sessions can seem endless and I need to manage the time more effectively.


This interests me. There definitely seem to be two camps: review misses after the session (but keep on chugging while doing your reviews), and review misses as you go along.

I’m in the latter camp myself, but I can see pros and cons to both. To my mind, the main wins with reviewing right after missing include:

  • The info I need is RIGHT THERE, just a click or two away.

  • My mind is already in the context of that specific item. When I see the correct answer, I usually know immediately why I missed it. I also tend to have a good feel for whether I’ve been missing it a lot lately and if it’s worth taking the time to figure out why (often by adding the character I’m confusing it with to the Niai script if it isn’t already there and really focusing on the differences, sometimes creating new mnemonics).

  • When I finish my daily review session my only decision is how many lessons to do (sometimes zero). I’ve honestly never looked at anything other than my miss rate at the end of a review session!

On the other hand I can understand wanting to think somewhat modally: Doing reviews (testing yourself) then reviewing all the mistakes.

Personally, I don’t think I could stand knowing I had a small pile of misses that need further review at the end of a session. I’m not sure I’d have the discipline to work through them all.

In houses with a “clean your plate” mandate, I’ve seen kids pick out the vegetables they don’t like and eat them all-at-once either before or after eating everything else so they can enjoy “the good bits” without irritation. I’ve also seen kids that take an occasional bite of the despised vegetable as they go along, expressly so they won’t have a big pile to work through (the occasional small bite isn’t as much of an ordeal). I was that latter kid!


Yeah, I think it’s going to be personal preference which one you go with. I can completely understand your explanation of why that works for you.

I always finish my reviews before I do lessons, so reviewing missed items and then directly going into new lessons puts me in the right mindset.

The missed items are like the warmup. They’re easy because I only have to tweak a bit to learn them. And then I go into learning something completely new.

But again, it’s personal preference.

“Listen to everyone. Follow no one.” :wink:

1 Like

Haha, that’s exactly me.

1 Like

In this we are exactly alike. Doing lessons first would be heresy!

(Seriously: You gotta do your reviews almost every day, but it won’t kill you to miss lessons occasionally — that’s why I do 'em in that order. I might run out of time or energy!)


Well, and just logically, if you have too many reviews to do, stop adding more

1 Like

I had a similar question last year. I took a poll and found the results interesting, see below if you are curious. For me rapid recall results in many mistakes and leeches. I like to spend a minimum of 5 seconds, since it greatly reduces my mistakes. If i think i know the kanji i’ll spend up to 30 seconds or so trying to recall it. If I know that i don’t know the answer then I will fail intentionally in less than 30 seconds.

I have been trying to slow down, to eliminate rushed-fingers typos* and the thing where I’m not thinking, just literally sight-reading a reading for two individual kanji instead of the word they make together. Those kind of typos.

But that’s not the same as sitting there trying to come up with an answer when I don’t have one. If I find myself without any answer, “nn” it is. After all, real conversation moves too fast for “uh, what was the word again?” all the time to be practical. It’s just a matter of what your personal standard is.

* I refuse to use the double-check script to retroactively “ok” typos. I consider teaching my fingers to type it correctly part of the deal. I know I’m the crazy one on that, but :man_shrugging:

If I can’t get the right answer within ten seconds or so I deliberately get it wrong. The more times you see something, the better it will get ingrained in your memory. Getting answers wrong is really good for learning.

Great topic! I have been wondering about this myself.

Like a lot of other posters have said, the best method seems to come down to a combination
of approaches.

I burned through the first levels and felt great about how things were going. If
I got a kanji or vocab word wrong, I just let the SRS bring it up for me again.
I never spent more than a few seconds on any one word or kanji - I either knew
it instantly or I didn’t.

However, I eventually built up a huge backlog of chronic leeches that have now
slowed me down.

So, I am now at a point where it seems to make more sense to stop and think
for words that give me trouble, and to make notes of any I get wrong. After a
certain point, it seems to require some conscious work/study to get past leeches
that I’ve been getting wrong over and over again.

1 Like

The best thing to do for your memory is to give your brain some time to think so that it has a chance to pull the information it’s looking for from your memory, otherwise you aren’t really practicing anything. If you can not get the answer after a few seconds try to find something in there to help you jog your memory, like try to find the radical or try to remember when you did the original lesson (what other words were in the same lesson that you do remember?). These things can all help you. If after that you still can’t get it you should see what the answer is and give yourself time to try and find a better connection to the kanji or word or whatever it is. If you just look at the answer and move on you are likely to find yourself in the same spot over and over again, which means whatever you are using to remember the word/kanji isn’t strong enough and you need something stronger.