I think making MCD is not worth the effort. I've been timing myself and I'm reviewing sentences at a rate of over 300/hr. Most of my sentences have various degrees of English in the answer field, ranging from definitions to full translations. However, when I review, I simply glance over the answer field. If anything I look at the readings and that's it.

Besides Khatzu doesn't believe in using vocab cards so perhaps this method works for him. I use vocab cards to learn words, and sentence-cards to reinforce their meanings, learn their context, and improve recognition.

well..I dont' think you should do MCD till you go monolingual with your cards anyway.
You shouldn't go monolingual just so you can do MCD. I feel like once you get used to Japanese (have some amount of japanese intuition within yourself), MCDs are REALLY useful. It involves outputting so it only makes that you have some level of japanese intuition and so doing monolingual cards is a must but of course...

Like I said before, I love doing my MCD Deck. It's easy and fast. I chose what I want to put in. I know I can remember it and I know I want to know it and that's why i put it in. I mean sometimes I do forget cards in my MCD deck but usually I just re-learn it.... but even that's easier than my normal vocab/sentence deck.

the more this thread gets bumped, the more i want a big mac. lol

Oh and for the au thing. I know what you mean by that even then I think that the verb au is just so common (even with all the kanji variations you mentioned... how one is usually seen in lyrics if it's emotional or if you come across a bear etc) that there's no point making a card for it in any format. At least I've never made a card for it and I know all the variations in the kanji, and when to use it... just because it comes up b.c. it's common.

Well it is your anki deck. you can do what you want. just my opinion up there.

There's merit to MCD, but here's been my experience:

- Don't use these on vocabulary cards such as from iKnow. The sentences are far to simple. If you must, try English word to Kanji word on those with english sentence and clozed kanji sentence as support context.
- Works pretty good with grammar type sentences, but you're limited in application
- Works GREAT with Subs2SRS type decks with context options.

It seems that MCD is what you want to do when you start digesting real world material. You got a paragraph for a good story or article. You have a word or phrase being used that the context surrounding it seems to support. Intuitively, if you can fill in the blanks in that situation, you can fill in the blanks in real life. A pass is not about understanding the context, just filling in the blanks.

Turns passive review (understanding the sentence is needed to pass) into a more active review. You have only one point of failure. You can make multiple cards where other parts of the paragraph are blanked.

One thing I'd recommend is not to look for paragraphs to teach you words. You want to learn words so that you understand the paragraph (and in turn, the entire story). Only reason I take that point of view is the success from doing Subs2SRS even without cloze deletion.

Nukemarine - I don't see how any of these merits don't apply equally to normal cloze deletion cards with normal sentences and sentence fragments.

You do suggest that you're developing the skill of 'filling in the blanks in a paragraph' that will apply in the real world - but in the real world, there are no blanks unless you've been badly mistreating your books.

If it's about filling in the blanks in production, I don't see how a paragraph matters - sentences are, I think, the basic unit of language, and sufficient for quizzing.

Not trying to rain on anyone's parade, I'm ready to try MCDs if I can conceive of a reason, but honestly, I think MCDs solve one problem: If you take material from books and anime, you start to forget the context and the out of context sentences become difficult to understand. This problem can be solved equally well by taking sentences from dictionaries and grammar books, or being very selective about taking sentences from native material and looking for ones that are coherent without much context.

Quote:You do suggest that you're developing the skill of 'filling in the blanks in a paragraph' that will apply in the real world - but in the real world, there are no blanks unless (.....) been badly mistreating your books.

So, as a native speaker, you can fill in the above blank with "you have" or "you've". It's possible to be "I have" or "someone has". There's likely others. However, since this paragraph was drawn from a source near and dear to you, you know from the context that "you've" goes in the blank.

It's all made up anyway, so you can do as you like. Just remember that fill in the blank is difficult if you don't have enough context to tell you what specifically should go in that blank. The benefit of native material is the assumption that what goes in the blank is a native level answer. The context can be an English word. The context could be the grammar section in question with just filling in how to modify the word with the grammar point.

In a touchy feely sort of way, I think MCDs are more like how we approach song lyrics. Here a couple of versus and the rest flows out of you. Tapping into that with a lot of MCD sentences, you get more native like in your production without leaning heavily on grammar rules.

Nukemarine Wrote:It's all made up anyway, so you can do as you like. ...
In a touchy feely sort of way, .... Here a couple of versus ...

Owww. 'hear a couple of verses', please. Sorry, ah...
Ahem. Not here for English, this is about Japanese, right...

As I said, I'm not trying to rain on anyone's parade, but I'm looking for something a little more concrete than 'If you love the context, it will magically come to you' as a reason to go into MCDs. I find that my love of the context has no impact by the third review or so of any given card. By then it's been a week or so (1+2+4 days roughly in Anki scheduling) and if I'm even still in the same series I'm in a far removed point in that series. That's why I stopped using what I was reading and started using dictionary sentences.

Also, your example fails - as a native English speaker, the only difference I see between filling the blank with "you've" and "you have" is really such a slight difference of tone that I could have gone either way (and 'someone has' is still viable, actually. 'I have'/'I've' ... not so much.) Even though it's my own words, I doubt I could reliably make the 'right' (ie, the same) choice. (Unless, of course, I know it must be one word in which case the contracted form does become the only choice.)

I'm with you Chris, sentences that I've added weeks or months ago have long lost their context.

Back when I had a few hundred sentence in total, I still remembered context, took my time to analyze the senteces, carefully looked at answer fields, etc. At a few thousand sentences in, all that went out the window. Now, I read the sentence, glance at the anwer (if that) and move on.

Maybe it's just me but I don't have 30 seconds to spend on an individual sentence, and I'd rather zip threw my reviews and move on.

I'm still confused as to the difference between normal cloze deletion and MCD. For example, I have subs2srs cards like the following on the front:

[image of scene]
context sentence -2
context sentence -1
sentence with cloze
context sentence +1
context sentence +2

It has some pretty basic context (just the surrounding 4 lines from the show via subs2srs's context options) and a cloze deletion. Does this count as an MCD and if not, what would need to change to make it one?

I would call that an advanced MCD-actually.... or a super-contextual MCD... not only do you have the text as context for the word, but you can hear it as well...

now perhaps that's actually too much context... we don't want this to become too easy right? basically that card is a listening comprehension card with textual clues as well. I like the idea of breaking down a listening comp card into multiple small parts because trying to pull 12 kanji characters out of an audio clip without writing tools is pretty tough... But plain MCDs are good for learning to predict which particles fit in situations, via grammar clozes, and also for learning chinese-derivative multiple-kanji words by closing one kanji at a time out of a jukugo.

datach has a good sence of the essential idea behind the MCD tho... that you can take on native material with awesome compelling content and still break it down into micro-chunks so that on a daily SRS-ing experience it is really easy. I think you guys that have lost the context were never really THAT into the particular context in the first place. I still enjoy MCD cards that I made about sketchy news stories from Roppongi 2 years ago.

Hey datach, how do you make cloze deletion cards from subs2srs? do you just pick the most important point from each card or are you able to somehow choose multiple kanji to cloze, making several cards from one audio clip...?

dtcamero Wrote:Hey datach, how do you make cloze deletion cards from subs2srs? do you just pick the most important point from each card or are you able to somehow choose multiple kanji to cloze, making several cards from one audio clip...?

Nothing fancy. I use morph man and just do i+1 sentences to focus on building vocab for the most part, but every so often I look at my i+0 cards and manually pick out a useful thing to cloze and unsuspend them.

datach Wrote:I'm still confused as to the difference between normal cloze deletion and MCD. For example, I have subs2srs cards like the following on the front:

[image of scene]
context sentence -2
context sentence -1
sentence with cloze
context sentence +1
context sentence +2

It has some pretty basic context (just the surrounding 4 lines from the show via subs2srs's context options) and a cloze deletion. Does this count as an MCD and if not, what would need to change to make it one?

Looks good to me. If you think the context is not enough to help you fill in the blank, either add clues or just delete the card.

Maybe I should clarify I haven't read any of Khatz's posts on MCD other than the recent one's on his blog. However, what you're showing is what I think about when I'm talking about MCD's using SRS. Something's blanked out and there's some clue on how to fill in the blank. What many of us did prior to cloze deletion was use the kana version as the clue or the English translation. Heck, even RTK is cloze deletion if you think about it. You're just using the keyword and any additional info as the context.

Nukemarine Wrote:However, what you're showing is what I think about when I'm talking about MCD's using SRS. Something's blanked out and there's some clue on how to fill in the blank.

I just don't understand the "massive context" part. All good sentence cards should have at least some context (surrounding dialogue and/or image for subs2srs, surrounding lines and maybe the topic for sentences from an article, etc) so obviously any good cloze deletion cards should have some context. What exactly brings it from cloze deletion with standard amounts of context to "massive" context?

Or am I over thinking this and Khatz's MCD vs CD is basically CD vs bad CD?

i think the idea is generally that by using 3 paragraphs as opposed to 1 sentence, you can make the content more interesting, which makes the card more interesting

But another important point is that a cloze can be a bit ambiguous, and not always direct you to the answer very well, even if you know that word. If you have more context you have more contextual clues leading one towards that answer.

It's a more deductive form of game, really...not just rote memorization. This aids the formation of predictive abilities, like for grammar as an example...

I'd say that I've also noticed a boost in my ability to skim over text more quickly, looking for the contextual clues required to find the answer...

My understanding of what a MCD is, is that you take a single card, and make multiple cards where as many possible words are clozed (one at a time) Some people also will cloze every kanji, so that you only have to remember half of a compound's pronunciation.

So for example the sentence:

このゲームは基本的にキーボードとゲームパッドで動くようになっていますがコンフィグだけはマウスしか操作できません。

Would generate around 10-30 cards, each clozing a different word, or a single character in extreme cases. You then study the cards, seeing the same material over and over again, but with a different goal to identify, supposedly making it less boring and giving you a lot more exposure to each sentence.

It sounds like a lot of work to make these cards, but surusu has a built in script to automate the process, and there have been anki plugins written as well.

I never really used them enough to give any sort of opinion on how it works.

wow up to 30 cards? that's insane.

I would make 2 cards... MAX lol.

So many cards for that little sentence is crazy though.

From what I read on MCD it does seem really useful but one has to have a good level of Japanese behind them to understand what the context is giving. I plan to just add MCD-style cards from now on because I feel sentences by themselves lack context to what it really means(studying grammar is definitely done in short sentences but MCD can be used there to)

It is an insane amount of cards, but I guess the whole point is to make the learning more active. For me, with sentence cards, once I see the first few words of the sentence I recognize it and know what the answer is. If I had a random word blocked out and had to fill it in, it would force me to pay attention to the whole thing.

syntoad Wrote:It is an insane amount of cards, but I guess the whole point is to make the learning more active. For me, with sentence cards, once I see the first few words of the sentence I recognize it and know what the answer is. If I had a random word blocked out and had to fill it in, it would force me to pay attention to the whole thing.

well that's a good point, but
-first it is only that many cards if you have that many unknowns in the sentence
-and second, the point is also to make the question as simple as possible, so that doing reps is easy like popping bubblewrap, and not this draining belabored activity. if it's easy you can do it longer and longevity is the key to success...not getting burned out.

I used MCDs last year (had an ajatt+ membership at the time) and made a webapp for myself to be able to make new cards for Anki quickly.

I've moved on to other deck styles, but maybe some of you find it useful. I uploaded it here: http://www.zaoyin.eu/chinese/Chinese-MCD-Generator/
(I'm learning Chinese, but it works just as well for Japanese).

can you describe how this differs from tokyostyle's mcd plugin?

The main advantage of the web app is that you don't have to manually copy the characters you want to cloze into a text field but instead use your keyboard for everthing. This makes the creation of new cards very fast. You can try this by following the link, scrolling down to the light red colored "Cloze codes" input box and inputting a few codes like "fa" and "fb".

Its main disadvantage over the Anki plugin is that you have to manually import the cards into Anki. This doesn't take long but it may be a bit intimidating for people who just started using Anki. Another limitation is that you can only cloze one character at a time.

I actually quite struggle with grammar these days, and I never learnt grammar really after Genki 2 and tend to forget these "rules" (never past tense before と or 前・まで etc.) and would like to have a run through all the grammar points at least from DoBJG.. I haven't comprehended the concept of MCD yet, however, as my grammar skills are so horrendous, I would try also something 'new', so... Advice?