My main focus right now is to build my vocabulary.

I used core2k for a while, the version that's divided into steps. By step 3 I encountered quite some vocab that I found a little advanced. Words that you might find in news, but that might not be so important for everyday conversation, like 政府, 発表, 首相 and so on. I have a harder time memorizing some of these.

On the other hand it is currently my impression - while I haven't progressed so far into it - that Nayr's core covers more vocabulary that is perhaps more useful in everyday conversations.

Does anyone share this opinion, or what has been your experience with these decks? Thanks.

I started off using those decks and got itchy feet at about the same point. 政権が交代した was the one that finished me off iirc.

I then switched to Nukemarine's i+1 ordered core6k and never looked back.

Nayr's core5k has many issues which make it unsuitable for beginners imo, though it's great as a listening deck for more advanced learners.

thanks anotherjohn.

good to hear I'm not the one who thought the learning curve wasn't quite that even in the seperated core decks.

and thanks for the advice. I'm not sure which of the core decks is the i+1 deck you mentioned or if it's still up?
Would you mind posting a link if you can?

thanks!

I remember feeling that way too, but that feeling faded after I got better at guessing the meanings of words. You should probably just suspend any words that seem less useful or hard to memorize. There's no reason to learn every word in core immediately and you can always unsuspend them later. The hard words get easier to remember once you gain more and more vocabulary.

The exact title of the deck is "Core 2k/6k Optimized Japanese Vocabulary".

Unfortunately it was removed from the Anki shared decks list in the recent purge, but if you search this forum (e.g. using site:forum.koohii.com Nukemarine "Core 2k/6k Optimized Japanese Vocabulary"), you never know what you might find.

anotherjohn Wrote:The exact title of the deck is "Core 2k/6k Optimized Japanese Vocabulary".

Unfortunately it was removed from the Anki shared decks list in the recent purge, but if you search this forum (e.g. using site:forum.koohii.com Nukemarine "Core 2k/6k Optimized Japanese Vocabulary"), you never know what you might find.

I hadn't heard this, how come?

OzarM Wrote:

anotherjohn Wrote:The exact title of the deck is "Core 2k/6k Optimized Japanese Vocabulary".

Unfortunately it was removed from the Anki shared decks list in the recent purge, but if you search this forum (e.g. using site:forum.koohii.com Nukemarine "Core 2k/6k Optimized Japanese Vocabulary"), you never know what you might find.

I hadn't heard this, how come?

Because the company behind iKnow/smart.fm decided that they were going to bully people into taking down content when they had no right to do so (apparently they regretted using a license that allowed for copying and derivative works after they decided to charge people an arm and a leg for flash card subscriptions and decided they didn't like the legal competition...)
Yeah, never paying them a dime ever again, and I tell people not to use them as well... Seems that stunt worked out quite well for them (lol).

As for that deck, I remember the format not being to my liking, so I changed it (though that might not be the iteration of the deck I'm thinking about). I think it was cloze delete or something... Anyway, I recommend going through Core as a recognition deck.

anotherjohn Wrote:I started off using those decks and got itchy feet at about the same point. 政権が交代した was the one that finished me off iirc.

I then switched to Nukemarine's i+1 ordered core6k and never looked back.

Nayr's core5k has many issues which make it unsuitable for beginners imo, though it's great as a listening deck for more advanced learners.

There's nothing advanced about Nayr's Core 5K. It's essentially just a version of Core2k/6k with a significantly more modern and accurate corpus, and better sentences.

sholum Wrote:Yeah, never paying them a dime ever again, and I tell people not to use them as well... Seems that stunt worked out quite well for them (lol).

I hope you don't actually think your actions are relevant enough for them to care or even be 0.01% affected by.

ryuudou Wrote:

sholum Wrote:Yeah, never paying them a dime ever again, and I tell people not to use them as well... Seems that stunt worked out quite well for them (lol).

I hope you don't actually think your actions are relevant enough for them to care or even be 0.01% affected by.

I don't, but I hope you don't think I care what you think about my purchase choices... Because that's even less of an effect than my actions have on their profits.

ryuudou Wrote:There's nothing advanced about Nayr's Core 5K. It's essentially just a version of Core2k/6k with a significantly more modern and accurate corpus, and better sentences.

I didn't say it was advanced, just that it has issues that make it unsuitable for beginners, so that in my opinion the core2k/6k is the better choice of the two, based on long experience of both. Recommending a deck based on nothing but hearsay is not something I would do.

sholum Wrote:Because the company behind iKnow/smart.fm decided ...

I think Core 2k/6k/10k which were discussed here on the forums came from iKnow and JSensei.
The 6k iKnow Core sentences were probably based on JSensei (which has almost 10k). iKnow brought in different/better audio for the first 6k and also pictures for the first 2k.
iKnow had engaged professional “voice artists”, so the rendering of the first 6k sentences is really good and the remaining (almost) 4k from JSensei are performed significantly less interesting/vivid. But it is less important at that stage.

ryuudou Wrote:There's nothing advanced about Nayr's Core 5K. It's essentially just a version of Core2k/6k with a significantly more modern and accurate corpus, and better sentences.

Nayr's Core 5K is completely independent of Core 2k/6k/10k. It uses a different corpus – not necessarily a better one:
Do you want e.g. を / だ / が / も/ です in a vocab builder? Bit basic, but also for more advanced grammar points I would turn somewhere else.

Is it helpful to (re)learn www, CD or NHK? Well, you can throw these 15 cards out, they do not harm.

Do you need after 一年 also 二年, 三年 and 30 other variations?

ブログ you don’t find in Core. Yes, that is less modern - so what?

アメリカ, ヨーロッパ, インターネット and メール are really important words and they seem to miss in Core – but I do not miss them as an independent vocab entry. They are in Core example sentences. And if they were not, you wouldn’t miss them either.

"Better sentences" is a question of taste and expectation. Here is the sentence for だ:
僕は英語が苦手だ。Not that helpful (too difficult to bring across だ) and also not so interesting.
And here is the sentence for 苦手:
私は英語が苦手です。 That looks familiar [don't worry Core is similar: 私は料理が苦手です] and there is more like this:
私は漢字が苦手です。
私は数学が苦手です。
私は魚の骨を取り除くのが苦手だ。
ドイツ語を話すのは苦手だ。
私は葱が苦手だ。
The only thing interesting in these sentences is the kanji for ねぎ (葱 Heisig 2389). But you find ねぎ also in Core, once without kanji and later also with kanji:
みそ汁にねぎを入れました。Core 6k
みそ汁に葱を入れました。 Core 10k

These are random samples, which make clear that both decks have boring sentences. But I wouldn’t care, mostly they serve their function (providing a very simple example) well and even interesting sentences do get less interesting each time you see them. Just concentrate on the vocab.

anotherjohn Wrote:...in my opinion the core2k/6k is the better choice of the two, based on long experience of both. Recommending a deck based on nothing but hearsay is not something I would do.

Makes sense.

sholum Wrote:

ryuudou Wrote:

sholum Wrote:Yeah, never paying them a dime ever again, and I tell people not to use them as well... Seems that stunt worked out quite well for them (lol).

I hope you don't actually think your actions are relevant enough for them to care or even be 0.01% affected by.

I don't, but I hope you don't think I care what you think about my purchase choices... Because that's even less of an effect than my actions have on their profits.

Just giving you a reality check. How much you care about me doing so is much greater than how much they care about your "disobedience". You're a nothing strictly speaking. You thought you "got" them or something.

sholum Wrote:Yeah, never paying them a dime ever again, and I tell people not to use them as well... Seems that stunt worked out quite well for them (lol).

Not how it works. You're not a hero. This has absolutely no effect on their finances, and you have no concept of scale if you think otherwise.

Matthias Wrote:Nayr's Core 5K is completely independent of Core 2k/6k/10k. It uses a different corpus – not necessarily a better one:

I know it's completely independent. And yes it's a much better one.

ryuudou Wrote:

sholum Wrote:Yeah, never paying them a dime ever again, and I tell people not to use them as well... Seems that stunt worked out quite well for them (lol).

Not how it works. You're not a hero. This has absolutely no effect on their finances, and you have no concept of scale if you think otherwise.

You have a point, but aside from firebombing their offices or hacking their servers, there's not much else you can do to express your disgust with a company. And I do think it has an effect when your cohort acts similarly. I know I won't pay them any money either. I don't think they're loosing sleep over me and sholum, but we're not the only ones who vote with their wallets. And there's some research suggesting that unethical companies do worse than ethical companies. Some call it karma, I call it death by 1,000 cuts.

Sure but in this case the vast majority of people simply don't know or don't care in the slightest, so no 1000 cuts. If anything they're profiting because all of the new people learning Japanese will be paying them and not using their corpus for free on Anki. Him writing a sentence like he saved the world and "made them think twice" on his actions alone was just comical. iknow.jp is 150,000 ranks above this site on Alexa.

I assumed the (lol) part meant he wasn't entirely serious.

[RawToast]

...and let's back on topic

ryuudou Wrote:

Matthias Wrote:Nayr's Core 5K is completely independent of Core 2k/6k/10k. It uses a different corpus – not necessarily a better one:

I know it's completely independent. And yes it's a much better one.

I agree. I found the grammar and sentences in general were a step up in difficulty from Core 2k/6k, but if you're comfortable with them -- I would use Nayr's deck instead.

If you want to use one of those decks to unsuspend words as you learn them, then I think both are ok but obviously core10k has more words so you'll end up with more cards already done for you to use.

But if you want to use one of those decks as a frequency list to study following their order of words, then I think Nayr has a more coherent frequency order.

Just as an example, among the 10k cards of Core10k there isn't one with "口調". I know there isn't a perfect frequency list because it depends on what you read, but overall I think this is a word which deserves to be among the 10k more common Japanese words. (It is also a word which uses the secon on-yomi of "口")

I know it is only an example, I'm sure they are both good decks but it's good to know that both have "faults" (but this is obvious, a perfect deck is unrealistic xD )

cophnia61 Wrote:... among the 10k cards of Core10k there isn't one with "口調".

I think that is a valid point. The most important question is which vocab do you find in the deck. I have some sentences with 口調, so it is definitely a word I miss in Core. And there are many more which are missing in Core. The only thing is, it is worse with Nayr's. It is a simple question of size: 10 vs 5.

With just 5k you need to have an impressingly better vocab selection than Core. That is difficult to achieve. And there is not a lot of room for slack - Nayr's has plenty. I gave some examples, but there are a lot more.

I agree that both are good decks. Just the blunt claim that Nayr's is much better than Core is not backed by a comparison of the data.

Actually I very well understand the appeal Nayr's deck has. I made a thorough check as I was initially not that satisfied with Core and Nayr's seemed very promising. But as a result I had to correct my oponion.

yogert909 Wrote:I assumed the (lol) part meant he wasn't entirely serious.

Indeed. Considering I almost never use 'lol', I figured it'd be a good indicator; guess it wasn't... For future notice, I, sholum, only use 'lol' when I really need to convey that something that lacks sufficient context to have tone is actually not to be taken (completely) seriously, because I think it looks out of place when used with my regular writing style.

On Topic:
As far as corpora go, wasn't there a list based on innocent books? I know there's not a deck made out of it yet (that might be a project for me next year if no one else takes it up by then), but I remember someone saying that the usefulness of the most frequent vocabulary in it far surpassed Core10k (or was it 6k) and Nayr's Core when using it to test the readability of various texts.
Maybe it'd be good to do Core2k, then start mining the most common words from that list...

Nayr's Core vs. Core10k
I did Core10k. I was going to use Nayr's as a listening deck, but quit shortly after starting for some reason... I remember not liking the format that much, but that's not a real problem. And yes, Core6k/10k sucks if you mostly read fiction (especially fantasy), but I think I've seen most of the words in one place or another (though, based on my Anki reviews, there are clearly some that I almost never see).

In the sentences, Core2k uses very basic grammar so would probably be easier for a beginner trying to build up some initial vocab. Nayr's Core has more interesting/natural sentences.

One big difference is that nayr's deck hasn't had the sentences indexed for n+1. I guess it would be trivial to run it through morphman, but for someone starting out, they likely don't want to bother with that, let alone know why it's desirable.

Another difference is that nayr's deck counts words like の、二、＆ ね as words while core does not. All this means is that nayr's deck has maybe 30 less actual words than you might think.

Nayr's Core is better, because it covers a better vocabulary set. Imo, the audio is better quality as well, and there's a lot more variety in the examples sentences too. And both decks are damn near perfect as far as the correctness of the sentences and the translations.

While it's not geared specifically towards beginners, the way Nukemarine's optimized deck is (who did great work with that deck), you CAN use Nayr's deck as a relative beginner. You just have to suspend the difficult cards. It's a very easy thing to do, especially if you learn and use the keyboard shortcuts (which is something everyone should be doing anyways with Anki, and not just Anki; the mouse is an awful invention that has cost the human race billions upon billions of wasted man hours ).

And even if you end up suspending half the deck, it's still only a few seconds per card, so there's a negligible time cost to it that shouldn't influence your decision and cause you to choose the older, less relevant Core2k/6k deck.

yogert909 Wrote:One big difference is that nayr's deck hasn't had the sentences indexed for n+1. I guess it would be trivial to run it through morphman, but for someone starting out, they likely don't want to bother with that, let alone know why it's desirable.

I think making a deck truly i+1 (as per the Krashen definition), which I'm assuming is the intention of the optimized Core deck, can only be done by the person doing the studying anyway. (not to downplay the work that has gone into the optimized Core deck: it's very useful to a lot of beginners, including to me back when I didn't understand the theory behind it - but now, having read Krashen's theory, I can roll my own optimized, personalized deck, with very little effort).

I did some number crunching on the two decks and compared the vocabulary in the example sentences against 4 different corpora. As Matthias predicted, since core has 1k more sentences it's not a surprise it beats nayr's deck by a nose in all categories. It's also worth noting that nayr's deck seems to re-use more words as his sentences contain only 4656 unique words(lemmas) vs 7057(52% more!) in core. Of course this says nothing about the quality of the sentences in either deck.

If I have time, I will crunch the numbers for the first 5k of core for a more even playing field.

core 6k coverage
76.46%: anime
63.67%: drama
84.46%: nhk easy
63.98%: wikipedia

nayr 5k coverage
72.65%: anime
59.47%: drama
81.57%: nhk easy
62.42%: wikipedia

For those curious about the corpora used, here are the sources:
anime: every subtitle from kitsunekko japanese subtitle database
drama: every subtitle from d-addicts japanese subtitle database
nhk easy: about a year's worth of articles from 2013
wikipedia: every article from japanese wikipedia.