Cool project! Go for it!
Just remember that you're gonna get some insane reviews at that tempo! (That is if you're using Anki / this site)

As Zorlee said, if you rush the new cards in the first days you will have a crazy amount of reviews then. In 2 weeks you can see a lot of new cards and you won't have any problem because you will be motivated, but keep in mind that the learning process (=reviews) will ask you also a huge quantity of time after the 2 weeks.

Just be aware of this

Not only is what you're doing entirely possible, I argue that it is even easier than with a slower pace.

I recently did RTK again (after almost 3 years and having forgotten almost every kanji) at around 150 a day and it took me 18 days (my original goal was 14 days as well). Here are the Anki graphs showing my activity during that time:



That massive 2 (and a half?) day dip in the middle was a business trip in which I sadly couldn't do any reviews . So it could have been 16 days!

As you can see, I only reached the full 5 hours you're proposing on two of the days, which happen to be Saturdays. So you don't even need that much time.

In my case, I did both production AND recognition, in that order, and found that having both was a significant contributor to my ability to recall the kanji. I added a priority on production cards to ensure recognition cards only appeared after all the production ones were done, which I suspect also helped. During the first month, I wrote down the kanji for every singe production card (a lot were on the whiteboard in ankidroid, but it still counts as writing).

Here are my deck stats. I like to think this exercise worked out well in the end.

Keep in mind that even once you're "done", you still have a massive amount of reviews to do for the next month. You'll still have around 200 a day in the second month. After the third, it'll slow down to a manageable 30-40 a day, but you'll probably just quit around then (I'm really thinking I should stop around here as well).

I promise this isn't bragging. I really think it's important for people to know that it is perfectly attainable. I pulled it off with a full-time job and no life, so you can too!

EasyJapanezy wrote:

I'm exited knowing I'll conquer kanji in 2 weeks instead of the months it would have taken me at my previous pace.

It's the only way to survive this madness.

Did I hear you say 2500 kanji in the video? Are you planning to do RTK3 as well?

Last edited by netsplitter (2011 December 06, 7:03 am)

netsplitter wrote:

Not only is what you're doing entirely possible, I argue that it is even easier than with a slower pace.

I recently did RTK again (after almost 3 years and having forgotten almost every kanji) at around 150 a day and it took me 18 days (my original goal was 14 days as well). Here are the Anki graphs showing my activity during that time:

As you can see, I only reached the full 5 hours you're proposing on two of the days, which happen to be Saturdays. So you don't even need that much time.

In my case, I did both production AND recognition, in that order, and found that having both was a significant contributor to my ability to recall the kanji. I added a priority on production cards to ensure recognition cards only appeared after all the production ones were done, which I suspect also helped. During the first month, I wrote down the kanji for every singe production card (a lot were on the whiteboard in ankidroid, but it still counts as writing).

EasyJapanezy wrote:

I'm exited knowing I'll conquer kanji in 2 weeks instead of the months it would have taken me at my previous pace.

It's the only way to survive this madness.

Did I hear you say 2500 kanji in the video? Are you planning to do RTK3 as well?

Can you explain the method you used to do this many cards in a short time in more detail? I find 50 cards + ~120 reviews takes me 3-4 hours, so I'm interested how you managed 150. I have a winter break soon and would like to finish RTK then.

I don't see the point in RTK3. You'll be learning characters you won't be seeing for a long time, and likely will never need to write. When one learns words like 蚯蚓、麒麟、蝙蝠、駱駝, etc. (which are likely to be written in kana anyway) one doesn't study the individual kanji, one just recognises the compound.

I'd just do RTK1 for now to reduce reviews / chance of burnout. You don't actually need RTK3, and even if you want to practice writing kanji that it contains, you can just unsuspend them as you encounter them in words that you've learned. They will probably stick better this way, too.

Fallacy wrote:

Can you explain the method you used to do this many cards in a short time in more detail? I find 50 cards + ~120 reviews takes me 3-4 hours, so I'm interested how you managed 150. I have a winter break soon and would like to finish RTK then.

My fear is that I will give you bad advice simply because I feel obligated to give any advice, since you asked, and I'm a nice guy. I don't think there is anything special about what I did beyond what I've already described, but I'll try to expand on it.

I used the shared deck that has the top 2 stories from this site included (and even though it was generated some time ago, most stories in that deck still align with today's top stories). That way, I could progress with new cards using my phone very easily when I'm not home, which is something I did often. I commute by train every day, which is about 90 minutes a day of ankidroidin'. Trains are so boring that even something as mind numbing as plowing though 800 reviews becomes the most exciting thing there is.

Like I mentioned, I did production first and recognition after. Recognition cards were introduced only the day after. I felt that introducing them the same day was a bad idea, since the kanji was still fresh in my mind. But that overnight gap, and something about using a different mechanism to re-build the meaning of the kanji, was a major contributor to solidifying the memory. I once forgot to unsuspend the recognition cards, and on the 3rd day when that batch of production cards trickled in for reviews, I struggled a lot, and even thought I messed up the order somehow since I was pretty sure I hadn't even seen some of them.

Another advantage of sparing recognition cards until the end is a breath of fresh air after 3 hours of drawing squiggles. They are so easy and effortless, and arrive at the same time, so you don't have to context switch your learning mode[citation needed]. You should be able to do 200 in 40 minutes.

Also, it goes without saying, but finish all your reviews before adding new cards.

I did most of my learning within anki. That means I got the new card, showed the answer, read the stories, drew it while embellishing the story in my mind, failed it, and moved on. This needs to happen fast. Don't dwell on a new card. I don't know or care what Heisig says, you can't spend 5 minutes making a theatrical performance in your brain for each kanji. It will return again very soon (and by default, anki maxes out at 20 failed cards, so they won't pile up and become unmanageable). And when it does, you will struggle, and you will fail it. And you'll cycle through this 5 or 6 times if you have to (usually 3, though), for each new card, until it just freakin' sticks. Somehow, it does.

Another thing which might sound dumb (and certainly earns me many strange looks on trains), but I mutter everything I do in anki. Especially in the learning phase, I might repeat the keyword to myself over and over until I finish drawing it. Maybe I'm over-analyzing it, but it's possible that saying the keyword uses an additional part of your brain to build an association with the character. Maybe it even counts as input since you're hearing yourself. (I know nothing about brains, so apply salt as needed).

It's worth mentioning that I've noticed, compared to three years ago, that anki's algorithms and policies are so much better than they used to be, especially for the initial learning period.  I'm very pleased to see it progress, and am confident that it is contributing to my progress a lot more than I give it credit for.

And if you have a job, just do anki instead. They won't even know. There are worse employees than you.

Oh no it's a wall of text ;_;

As for RTK3, I do think it's a bad idea. I even think RTK1 is overkill. Don't do it until you know you need to do it.

Last edited by netsplitter (2011 December 07, 3:53 am)

Thanks for the tips. Could you explain the difference between recognition and production cards? Also, if I've already done 620 on the other RTK deck (and I assume I can't transfer over deck stats), would it be worth just spending a day doing the first 620 on the other RTK deck with stories and progressing from their?

A recognition card is having the kanji in front of you and recognizing correctly which keyword is associated with it.

If you are recognising to Heisig's keyword then this seems just silly to me (unless your objective is to memorise Heisig's keword's, that is...)

OTOH, if you are recognising to some Japanese word (like, 寝 to 寝る let's say) then it makes much more sense to me, although I still would question that approach.

Well, yes. I would agree that testing for Kanji -> Keyword is not only pointless but harmful, as you're creating a very strong mental association that will need to be broken later to create the proper Kanji -> Actual Japanese Vocabulary association that is needed for real reading.

It is true, however, that drilling in both directions makes recalling the production cards easier in the short term, which is why people do it, even though it's likely harmful in the long term. Many people have convinced themselves that drilling in both directions is necessary for their particular learning style. I suppose I can't disprove that either.

netsplitter wrote:

Jarvik7 wrote:

Testing for the keyword makes about as much sense as testing for what frame number the kanji is.

You mean it uniquely identifies them? That makes a lot of sense. I couldn't have explained it better myself. Thanks. As far as I'm concerned, recognizing a character means looking at the components and deciding which character this one is or isn't. As a beginner, an English keyword happens to be that unique identifier. Attaching a useful meaning (and usually correct and common meaning) to the character is simply a bonus. That meaning has helped me a lot. 喫茶店 == consuming tea shop. The underlying meaning is instantly attached. Instant learning is best learning.

Don't cherry pick your examples.
漫喫 = loose consuming?
開発 = open discharge?

Having a uniquely identifying keyword (or frame number) is great and useful, but there is no real value in memorizing it. I wouldn't go so far as to say it's harmful (unless you mistake keywords for a kanji's true and sole meaning), but it's a waste of time to study twice as much material.

Last edited by Jarvik7 (2011 December 08, 3:25 am)

Very well then. But I wouldn't call it "twice as much" material. Not at all. I did mention how quickly I reviewed recognition cards (I wouldn't have done them if they took long), and they made the production cards much easier to learn. It's hard to say whether overall it took less effort because of it, but it definitely took less time, which was my main goal.

And those cherry-picked example occur frequently enough, so I'm content with the baggage of this extra knowledge since it makes at least those words easy. It's not like I can forget them now anyway.

Loose consuming and open discharge might be how I enter the word into my short term memory. I might even come up with a dumb mnemonic like "loose pages of manga where people consumed coffee". These short-term vocab mnemonics have been discussed on this forum before. They're usually involuntary and don't last more than a few days.

It's a fantastic stepping stone. You think it's incredibly stupid and a waste of time, I think it's wonderful.