Hey everyone. My first post on the forum!

I guess I`ll introduce myself. I`m a third year JET and I`ve been studying pretty dilligently (Around N3 level). Kanji has always been the bane of my existence. After dabbling in many different methods, I`m now 250ish into my second attempt through RTK, the first being when I tried to make my way through the book with physical flash cards (it was brutal). I WILL finish this, as I`ve made it a major life goal of mine.

Anyway, I have a question about reviews. I`m not sure if I`m doing reviews correctly. Basically, I use the site`s built in SRS.

If I can write the kanji without needing the story, I click easy.

If I can`t recall the writing of the kanji, but I can after I read the story, I click yes.

If I can`t recall the kanji at all OR I get it wrong, I click no.

Doing it this way, I`ve been getting 80-90% correct. And every time I do a review, I feel that the kanji is reinforced.

However I`ve been wondering if you guys grade your cards as "yes/easy" ONLY if you can recall the kanji without needing the story.

That`s basically it for now. Hopefully I can get some insights from this great community Thanks so much again in advance.

You shouldn't pass a kanji if you have to read the story to write it. Only if you can recall it with just the keyword.

Thanks. I`ve been searching around the boards and I havent found a conclusive answer to this. It seems some people are failing their cards if they can`t write the kanji from the keyword and some are allowing a pass if they take a peek at the keyword.

for example:
http://forum.koohii.com/showthread.php?tid=9124

Splatted Wrote:.
.
.
7-Allow yourself to look at the stories and pass the card if you can then write it. This may sound too easy, but a lot of the people with high retention rates are doing this. (Also, see below)
8-Be less strict about failing cards. For example, if you get a stroke wrong but tried to write the correct primitives, you can pass the kanji and just practice writing the primitive a few times. Focus on trying to get it right in the next kanji it appears in. (see below)
.
.
.
About 7 and 8: It's debatable how well you need to know each kanji, but if you're planning on SRSing vocab keep in mind that you'll be starting from the smallest intervals, so you're effectively re-studying a kanji every time you learn a word that uses it. This means it doesn't really matter if you've not got it completely perfect, as just being familiar with the kanji will make easy to remember. It's also worth bearing in mind that it will be a long time before you learn words for most of the kanji in the book, so I think it's probably inefficient to spend too long on them. In fact, I now find that it almost doesn't matter if I'm learning a word with new kanji, partly because the primitives are always familiar, but also because the way in which they are arranged often seems the most natural. In some cases I know that they'd never be the other way around.

This logic makes sense to me. I feel like high 80%-90%+ retention rates that people have are only possible with this method. I`m getting a good scene in my head for every kanji I add, and it "feels mine", to the point that when I need to glance at a story, the "memory" of the same story jumps back to me. But often, I can`t jumpstart that memory without taking a peek at the story. I feel like grading "no" on cards I need to look at the story at will make my relearn list way too big and will cut into time on my other reviews.

To anyone reading, PLEASE post a comment on how you do your reviews and what you retention rates are! I`d like to get a general sense of the relationship between retention rates and review methodology.

ryuudou Wrote:You shouldn't pass a kanji if you have to read the story to write it. Only if you can recall it with just the keyword.

"shouldn't" is a powerful word. Many people consider reading the story a fair hint and not a failure, and just grade it as hard to remember.

I do think it's important to take a moment to think deeply about how the keyword leads into the story whenever you can't go from keyword to story, but whether you consider it a harder grade or a fail is up to you, as long as the resulting kanji is correct.

It's also sometimes useful to change your keyword, if only to add a parenthetical comment to remind of -which meaning- of the keyword was meant, as some of them are ambiguous and the meaning associated with the kanji isn't the one that first springs to mind seeing the naked keyword.

I agree with Ryuudou completely--you should fail a card ('no') if you had to read the story.

I know there are people who think otherwise, but I don't see the point of doing it any other way. The story basically tells you 95% of the answer, so if you read the story and then can write the kanji, you barely accomplished anything. You might as well get used to the real world, where you won't have a story, right now.

Apollo87 Wrote:To anyone reading, PLEASE post a comment on how you do your reviews and what you retention rates are! I`d like to get a general sense of the relationship between retention rates and review methodology.

Standard 80-90% retention rates on mature cards were the norm even before "lazy kanji" methods you may read about started getting attention. 60-80% is rather typical for young card retention rate.

If you're getting 90% on young cards by allowing yourself to look at the story then it just proves that method is not properly testing your memory, but feeding you the answer with passive recall.

http://ankisrs.net/docs/manual.html#acti...ll-testing

I actually have no opinion one way or the other, but.

Tzadeck Wrote:You might as well get used to the real world, where you won't have a story, right now.

In the real world, you won't have a keyword trying to write kanji.

blackbrich Wrote:I actually have no opinion one way or the other, but.

Tzadeck Wrote:You might as well get used to the real world, where you won't have a story, right now.

In the real world, you won't have a keyword trying to write kanji.

Using the keyword to recall the kanji strengthens memory of the kanji by using active recall to retrieve that item. Reading the story doesn't do this as too much of the "answer" is given in form that you can read which doesn't test memory.

In otherwords "keyword - kanji" is the most effective review method for performance in real-world situations.

blackbrich Wrote:In the real world, you won't have a keyword trying to write kanji.

When I first finished RTK, I would think things like, "What are the kanji for べんきょう again? Oh yeah, exertion and strong, so it's 勉強." And I could write the kanji from memory, in sort of two steps.

Or, perhaps more often, I could recall the story to a kanji even if I didn't remember the keyword--"Oh yeah, the べん in べんこう is Bugs Bunny getting muscular 勉. It's some kind of strength-related keyword, but I can't remember exactly."

Just seems to me like I wouldn't have been able to write kanji from memory as well if I had allowed myself to read the stories.

Then again, I stopped doing my RTK reviews and now I'm pretty bad at writing kanji from memory. Keep up with your reviews kids. Then again, I'm 800 times better at Japanese than I was back then. I do want to go through RTK again someday though.

(And yeah, I think it tested my memory in a more rigorous way, so it reinforced the characters better in my mind even when it came to simple recognition.)

ryuudou Wrote:Using the keyword to recall the kanji strengthens memory of the kanji by using active recall to retrieve that item. Reading the story doesn't do this as too much of the "answer" is given in form that you can read which doesn't test memory.

In otherwords "keyword - kanji" is the most effective review method for performance in real-world situations.

What exactly is your real world? In mine I read a kanji in a word and go to a meaning. Going from keyword->kanji does not necessarily mean you will easily go from kanji->keyword, there are plenty of threads already about people having trouble already.

If by real world you mean writing however, then you may be right as Tzadeck mentions their story about writing from memory. Though I'm not sure everyone is able to go kana->keywords->kanji,

@Tzadeck
That's really interesting. The fact that you could remember べんきょう and remember the keywords used to then write kanji impresses me more than anything.

blackbrich Wrote:@Tzadeck
That's really interesting. The fact that you could remember べんきょう and remember the keywords used to then write kanji impresses me more than anything.

Ah, I should mention that I had already studied Japanese for two or three years before I did RTK, although almost two years of that was a speaking-centered class. But I actively studied about 500 kanji and had lived in Japan for a year or so before I did RTK, so I already sort of knew a bunch of kanji.

In other words, perhaps I was making some extra connections in my brain as I was going through RTK that you wouldn't be making if you did RTK right at the beginning of your studies. That might have helped me associate the keywords with real words to some extent.

Okay. That makes me want to try RTK again later. Maybe with Japanese keywords maybe. Writing by hand isn't extremely important to me but I'd like to able to write simple things even if it takes a bit of mental gymnastics at the beginning.

blackbrich Wrote:

ryuudou Wrote:In otherwords "keyword - kanji" is the most effective review method for performance in real-world situations.

What exactly is your real world? In mine I read a kanji in a word and go to a meaning.

In mine I read the whole word and go to a reading. Meaning usually comes automatically if I can read the word without slowing down.

Ah mines a bit flipped. I read the whole word and go to a meaning. Reading usually comes automatically if I can read the word without slowing down.

Heh, my "real world" so far is VNs with a text hooker so if I'm the least bit unsure about a reading, I can always just glance down for furigana (not that machine-generated furigana is 100% accurate).

Hmmm. Which text hooker do you use? That would make it a bit more enjoyable for me I think.

But to stay on topic a bit. Just do what you feel. And if you feel the liberal option isn't working switch to the more conservative. Or vice versa.

blackbrich Wrote:Hmmm. Which text hooker do you use? That would make it a bit more enjoyable for me I think.

Hookers make everything more enjoyable. I use VNReader. I prefer it over Translation Aggregator because it can use the Windows IME for more accurate furigana generation. TA's WWWJDIC lookup was nice back when I had to look up a lot of words in every sentence (specifically, so many that I couldn't even remember all of them at once!), but once you already know most of the words, it's a nuisance having to hunt for the one word you don't know, only to see a crappy EDICT definition (or to find out WWWJDIC broke up the text incorrectly).

On topic: if you look at the story and can't immediately picture the correct kanji, I think you should fail the card (and change the story if this happens multiple times). I've only ever done kanji to keyword (the kanji basically contains the story) so there's not much else I can comment on.

I always used ITH. It was alright for when I was doing more sentences in Anki, but not all that great after I stopped. Thanks.

edit: tried it. My word that's great software.

Thanks for the replies everyone. I`m going to switch to grading my cards more harshly for a few days and I will report back on my experiences. Actually, I already started yesterday and it wasn`t so bad but I don`t know what my reviews will be like in a few days or when my cards build up more.

Please continue to post your experiences. I`ll continue to update this thread as I add more cards. Just about to pass 300. Woo!

In the beginning I had a similar issues as Apollo87.

I used to grade my cards to lightly I would accept if I either remember the story or the way that Kanji is written. Now I click on "no" if I don't remember the story or I don't remember how the kanjis is written.

The other issues I had was with the fact that my stories were bad, they were directly takes from the book without making any changes. My native language is Polish, but I used to try to force myself to remember Kanjis using English keywords / stories, and
I noticed that I need a few times more effort to remember them.

I also started to memorize two stories for Kanjis that are harder to remember from this website. So after a few days one of them would usually stick and I would forget the other one.

Nowadays, I translated all keywords into Polish while keeping the English translation on the side as well. When I read stories I also translate them into Polish in my mind as it makes it easier to remember. I gave up on using Heising stories long time ago. I had trouble remembering most of his stories. What I do is I read the top 10 stories posted on http://kanji.koohii.com/ starting from 10th to 1st in order to figure out my own story. Sometimes some stories don't translate well into Polish, so I just ignore them as use ones that make sense for me.

I'm using my own list of cards made in Anki that I generated with my own script, I keep the top 10 translation from kanji.koohi.com there as stories.

I would spend like 30 minutes a day generating cards on my computer and then 2-3 hours at gym learning/reviewing Kanjis on my iPhone while exercising xD.