Mastering isn't worth doing until you've added them all since a lot of kanji share primitives, thus you will automatically train lots of kanji extra. However, your fail pile is indeed enormous so I agree you should probably calm down for a few days and work on your failed kanji. Make proper stories, proper vivid images then work on them. When your failed pile is down to 10-20 kanji, start adding new kanji again.

I think if you keep pushing through with such huge fail rates, it's indeed only going to get worse, because as you keep adding more cards, it will just make your fail stacks get bigger and bigger with each passing day. I would take a break from learning new ones, or just drastically cut back the amount of new kanji you learn each day. If you were taking 50 kanji each day, maybe try just 10 kanji per day for a while. As those stacks start getting smaller, ramp it back up.

I was using the book for about 2 months before finding the site.

I second Tobberoth and Zarxrax, and say spend a few days ONLY working on your failed cards.  I'd venture to say that you failed 10-30% of those cards because they are made of other "primitive" kanji that you also failed. (i.e. [#558 reef] is made of [#557 char]).

I cycle through my failed cards every day and return "learned" ones to the normal stack.  At first, I wasn't sure if this was the right approach.  I posted about it previously.  I noticed that LOTS of people have large failed stacks (so don't feel bad!), and so believed that maybe you were supposed to have lots of failed cards. Over time, I became more comfortable with my system of dealing with failed cards, and noticed that I was remembering all of them better.

Frequently re-adding your failed cards to the stack also gives you great feedback about your pass-to-fail ratios.  You notice which cards you've passed numerous times in the past, but failed (signals that your story is getting muddled with other stories).  You notice which cards you've failed more than passed (signals that you need to revise your story).  You notice which cards you've never passed (signals that you need to totally replace your story, and/or revisit primitives).  Use this to nip problems before they get out of control.  If you fail one kanji, any other kanji that use it as a primitive will also be failed, double/triple/quadrupling your failures.

Personally, I study my failed kanji right after my review where they failed. I work on the stories and such. Then I leave for an hour or two, come back and look at them. If I feel they are basically stuck, I add them right then and there. So I've never had a failed pile with more than 10 cards in it. This technique might seem like it will give you your failed cards back every time they come up again since I spent so little extra time on them, but far from it, there's like... 5 cards at most which I've failed more than once.

If your fail pile is that large, you should also consider *how* you're learning these kanji. Are the stories not sticking? Are they not vivid enough? Are you just trying to brute force remember the keyword? Because that high of a fail rate indicates that something isn't working right to me. Take a week or two to sort this bit out, then apply what you've learned to the rest of the book, and you should be fine.

I'd recommend tackling 20-30 cards in the fail pile every day, on top of the other cards that are due for review. That way you'll chop the pile down in ~10 days, and you won't fall behind with your regular reviews.

QuackingShoe wrote:

Wisher wrote:

When you start to read real Japanese, you will only really use less than 800.  (From what I understand)

Woah, woah, where'd you hear that?

I hear it everywhere.  Just about every book on learning Japanese says this. I even had a Japanese person, who grew up in Japan, tell me this yesterday. 

That is where one of the arguements against Heisig come from.  "Why dont we only learn the most common 800 Kanji since that is what we mostly use anyway?" Then Heisig says that it is important to learn all 2048 first because of the method of learning, not pratical usage.

Come to think of it, I just read a manga it it seemed to only use about 50 of the same Kanji over and over.  I still do not know what it says, but I practice my Kana and learn to recognize the Kanji I know.

I'm sorry, but I'll have to warn you that you can learn japanese without knowing any kanji at all.

But for the sake of reading japanese, 1000 are very few, you need much, much more.

I read very little japanese and I see non-RTK kanji everyday. This makes me think that even the 2042 in RTK are few.

There are many kanjis that are in RTK that are very rare, but those are the realy easy to learn, like "prosperous".

With a failed stack that size, I think you should stop and review before adding any more cards.  It's better to have a good solid grasp of the the ones you've read before progressing.  If you have failed/can't remember them, then those are the ones you need to study first: you're somewhat familiar with them but don't quite have them nailed down yet.   

Look at it as a positive thing: the failed stack is a very accurate picture of the cards you will benefit the most from reviewing, and you'll remember them more quickly than any new cards, since you've already begun the process of studying them and have read the story at least once.  I would study just the failed cards (while maintaining your normal reviews) until you eliminate them before adding more.

Once you start using this site consistently (i.e., daily) the spaced repetition of the reviews really helps move the kanji from your short-term memory into your long-term memory.  You came in at 800 without the benefit of these reviews, so the kanji may not have "stuck" as well as they will when you review them now.  Once you get rid of your failed stack and have a managable number of daily reviews, you can move on to adding new cards.