Hi folks,

I've been lurking this fantastic site for a year and if I may, I'm going to bother some of the anki geniuses here with a question.

Up to now a significant part of my study time has been put into working through a large JLPT vocab deck one level at a time.

This strategy helped me breeze the early JLPT tests, which I needed to do to enact the second part of my plan - postgrad study in Japan. I've covered about 2000 of the 8000 word list and my comprehension is definitely improving.

Now I have secured a place and have just received notice that I'll be going to Japan in two months and I want to change tactics slightly. I want Anki to deliver the remaining vocab by order of frequency in natural language rather than the random JLPT based approach.

What would the best way to do this be? Can I tag the cards sequentially using one of those big frequency lists? I don't want to have to start from scratch but rather just dole out the remaining cards in a more logical and useful order. Suggestions? Thanks in advance!

Jamie

This is what I came up with regarding vocab learning :
JLPT4 (easy, quick and basic vocab) -> JLPT3 (continuation) -> Core2000 -> Core6000

If you use smart.fm and import into Anki it should be painless to create such deck without duplicates. First step is a warm up with around 700 words, second part is another 700-800 that adds up to around 1500.
Now it starts getting interesting for you since Core's are based on some occurrence statistics. Some vocab will overlap with what you already have so adding Core2000 to that should (I haven't tried this yet so numbers could be different) give you another step of around 700. Once you've cleared this you have covered 2k most common words plus some JLPT vocab and now you can start working on big 6k

i would recommend the Core 2000 full vocabulary deck or the one I made combining that with the core 6000 spreadsheets from Group Study. They are both on Anki's shared decks. These have one vocabulary word for each sentence and the vocab words and meanings are in separate fields so that you can make vocab cards rather than sentence cards. The sentence is still there so you can have it as part of the answer field as well.

I leave in about 2 months also, so I am in a similar situation. My goal is to finish Core 2000 vocab, then 6000, then switch to a monolingual dictionary and just use native sources until I leave.

Thanks folks,

I took a look and downloaded the core 2000/6000 deck on your advice.

I'm glad I ditched the most of the remaining JLPT deck - way too much obscure / low frequency stuff at this stage compared with the smart.fm list.

I imported the new cards, mucked around to get the old ones into the new model then spend an excruciating day consolidating duplicates.

Now I have 4500 new cards, in order of usefulness, mixed in with the 2000 previously learned JLPT vocab cards to review, which should keep me pretty busy until April. Plus I still have a few hundred to finish with RTK.

So ... how close does core 6000 come to covering JLPT2? 70-80%?

Cheers,
Jim

@jimbojones:

The core 2000/6000 lists are no particular order of usefulness and aren't particularly selected based on their frequency. That was the intention, but the frequency list was derived from a financial newspaper so you learn words like "finance minister" before "mom".

Out of all the words in the JLPT1/2/3/4 lists, I only have yet to encounter about 500 (out of 8000 or so).

jimbojones99 Wrote:Thanks folks,

I took a look and downloaded the core 2000/6000 deck on your advice.

I'm glad I ditched the most of the remaining JLPT deck - way too much obscure / low frequency stuff at this stage compared with the smart.fm list.

I imported the new cards, mucked around to get the old ones into the new model then spend an excruciating day consolidating duplicates.

Now I have 4500 new cards, in order of usefulness, mixed in with the 2000 previously learned JLPT vocab cards to review, which should keep me pretty busy until April. Plus I still have a few hundred to finish with RTK.

So ... how close does core 6000 come to covering JLPT2? 70-80%?

Cheers,
Jim

I think the Core lessons are good for practicing your listening skills (that's what I used them for, anyway), but if you are wanting to focus on word frequency or natural Japanese, I would actually suggest mining sentences from blogs or TV shows. It's not an exact science, but you are more likely to encounter a variety of commonly used words by getting materials from natural resources, rather than using a bunch of pre-made sentences.

To answer your question, however, I think there was a thread about this awhile back that said Core 2000+6000 altogether covered maybe... 50? 60? percent of JLPT2.

Jarvik7 Wrote:The core 2000/6000 lists are no particular order of usefulness and aren't particularly selected based on their frequency. That was the intention, but the frequency list was derived from a financial newspaper so you learn words like "finance minister" before "mom".

LOL this is .......... sadly true. Although the first Core 2000 lesson isn't actually bad, if I recall correctly. I think it makes use of only very basic words.

Well, there is a reason there's a thread about the sorted Core list. Only Core 2k has been done so far though. Kanji Word Progression Thread

Helps solve a lot of issues with word order. Good response from those that try it.