Filling In The Blanks With MCDs — 24 Comments

  1. It’s worth nothing that the google search “MCDs Anki” provides some excellent results!

    (I might be a little biased though.)

    • There are definitely a lot of good resources out there to learn MCDs (hence why I was able to get information to write this article) but I felt that a straight-forward explanation and recommendation was difficult to find. But yeah, I actually use your plugin to create MCDs, so I appreciate all of your work!

      • You are absolutely right about the lack of clear examples, explanations and recommendations. I do have a lot more thoughts I’ll share later once more comments are in. I haven’t used sentence cards in over three years so I’ll share some things that aren’t immediately obvious especially if you never manage to get over the transition period.

  2. First, I want to say you did a much better job explaining MCDs than Khatz did on his site. The signal-to-noise ratio with his style of writing made it very difficult for me to parse what exactly MCDs were about without considerable work…whereas your explanation was elegant and straightforward, and I appreciate that.

    I tried MCDs but ultimately became frustrated with them because, due to the “one card per Kanji/particle/word/etc.” I would end up seeing the same content over and over and over again…this would lead to me getting so, so, so very bored.

    Then again, I was also just starting out and finding my way. Now that I’m heading into Intermediate country, your article has me considering giving MCDs another shot…for particles, if nothing else.

    • I really appreciate a lot of what Khatz does. It was from his website that I learned about immersion, SRS, and yes, MCDs. He does have his own unique writing style, and it seems to be working for him. But I was definitely hoping to give a more straight-forward explanation, so I appreciate the compliment!

      You bring up a very interesting point, though. I never even considered how boring it can be to see the same content over and over; that is definitely something to take into consideration. But that’s the beauty of not marrying yourself to “only” one method. It sounds like you could use MCDs for those tough 3+ kanji words and use normal sentences for the rest.

    • If you use anki and use the cloze deletion card style, each cloze becomes a “note” within each card. Once you do one of these notes, it buries the other notes for the rest of the day. This way you dont see the same card multiple times a day and/or read through the same sentence with that particular cloze in it very often.

    • In Anki, you are able to Cloze multiple words in a single card. Using the above article’s example, you could have the following as a single card:


      The background ‘code’ for the card would look like this:


      The idea is to have both cloze’d texts having the same number. In this case they are both ‘c1’.

      This example shows two particles both being omitted and thus, makes it way less boring to see the same cards. In theory, you can Cloze as many parts of a card as you’d like.

      This is an excerpt from the Anki Manual which explains this same point as well as a keyboard shortcut. やった!

      “You can also elide multiple sections on the same card. In the above example, if you change c2 to c1, only one card would be created, with both Canberra and 1913 hidden. If you hold down alt (option on a Mac) while creating a cloze, Anki will automatically use the same number instead of incrementing it.”

      More information about Cloze Deletion in Anki can be found in the online Anki Manual:

  3. Nice article.

    I also tried MCD’s for a while, just wasn’t my cup of tea. Pretty much just as StereotypeA said, it just got too boring and repetitive for me.

    I did it for about 500 sentences, which blew out to 1000 cards once I converted them to MCD’s.

    Seemed like I was seeing the same content everyday. The time I spent learning 500 extra cards made from MCD’s could have been better spent learning 500 new cards in my opinion.

    Everyone is different though, things work differently for others. I definitely did know those 500 cards back to front by the end of it.

    • The repetitive content seems to be a consensus on here. Maybe I never noticed it as a problem because I was too lazy to actually read any of the content on my cards, hence why I listed skimming as a con q:

      The reviews can certainly pile up. But I think that having more cards that are bite size can sometimes be more beneficial than a sentence card with a tough word. To extend the food metaphor, sometimes a whole sentence with a long, confusing new word can be too much to swallow, so you need to cut up your food into smaller MCDs. Man, I need to go eat breakfast.

      But I agree that everyone is different. I know many who have had a good amount of success with MCDs. I don’t think I will personally ever shift exclusively to MCDs, but if it works it works.

  4. I tried using MCCD to learn vocabulary, but it’s too slow to review. Now I mainly use it for grammar related things. Basically, I write a Journal at lang-8 and create clozed-delection cards with every correction I get.

    • Interesting. Clozes would be a perfect way to remember your corrections from Lang-8. I might steal this idea lol.

  5. I’ve been using MCDs somewhat (my MCD deck has 1367 notes, 4674 cards) as a supplement to my main sentence deck (now over 10000 J-J sentences).

    The first time I heard about MCDs was right when I started learning (two and a half years ago), before I even found JaLUp, but since the article about them seemed more interested in telling me just how awesome they were rather than what, exactly, they were, I ended up going with the far better detailed sentence method found here in this website.

    I then eventually decided to try them out a little over a year ago when someone (not sure who anymore, maybe Jan) mentioned them and Mike’s plugin around here. Basically I was trying to find a way to counteract stuff I was having trouble with in my sentence deck (about 3000-4000 J-J sentences in), such as mixing up readings for similar kanji.

    The way I tend to use MCDs is the following: take a paragraph (generally three sentences) whose middle sentence (usually) contains words (/particle) I am already familiar with but for which I’m still unsure about the kanji/reading (/usage in the case of the particle) and make closes for each kanji of those (/particle).

    In my personal experience MCDs excel at the following:
    – anything grammar related, like particles or conjugations. It’s quite easy to be nearly blind to stuff such as transitive/intransitive verbs when doing sentences, and indeed that was my case;
    – compounds of three or more kanji (partially because they can’t be confused with anything else. see below);
    – making sure you remember any specific differences in usage that you identify with reasonable certainty and would like to recall.

    On the other hand, I often found it problematic to try to use MCDs to tackle compounds with two kanji whose meaning is vague and abstract in nature.
    To see what the problem is, here is one of the first examples of this happening to me: at some point I made a card which had the word 思想 (“thought” in the J-E dic) in it with a close for each kanji. This seemed fine when I added it, except then I learned the word 思考 (also “thought” in the J-E dic, with the J-J dic not being much more helpful in distinguishing the words, particularly while you are still at an intermediate level). This means that my card I had made with the close around 想, which initially was just testing whether I knew the second kanji for “thought” was now actually also testing whether I knew the (presumably subtle) difference between 思想 and 思考, something even now I still don’t have a full feel for, much less back then.

    I find scenarios along the lines of that above to be reasonably frequent with my use of MCDs. Sometimes it does turn out that when I do take a good look at the two words confusing me and their definitions I do now see the difference, gaining a better understanding of both words in the process, but sometimes I don’t, and am just left wondering what to do, as presumably there is some nuance separating whatever words are confusing me, but I have no good way of knowing that the card I chose to use is at all representative of that nuance.

    Scenarios such as above are something I think any intermediate learner trying to go exclusively with MCDs would need a good strategy to be able to deal with, and I’m not entirely sure what that would be.

    • Yeah, I agree there are definitely some things that need to be worked on. If the context isn’t helping you distinguish the difference between two words, then it can be particularly difficult to make those distinctions at a good stage.

      For anyone with questions like this, I think a good person to ask is Mike. Maybe I should do a follow up Q&A with him for an article?

      • Although a single concept like MCDs will never perfectly solve every problem, I definitely have some suggestions for most of the issues brought up here. Adshap has my email address, and he absolutely loves when I bother him to give it to other people, so let’s do it.

        • That sounds like a great premise for an article that I would definitely be looking forward to. In fact, if somehow Bailey finds himself short on the questions, I’d definitely contribute some.

          And let me maybe add a few extra thoughts I didn’t have time for yesterday.

          Something I find understated in the (admittedly not that many) articles I’ve read about MCDs is that it isn’t just the review process which requires more active knowledge, but that in fact so does creating effective cards.
          With sentences it is perfectly possible to pick up a sentence that is a little over your level but for which you have decent guesses as to the meaning of and make a successful card out of it, even if it might then take some time (i.e. reviews) to fully digest. In this sense sentences are quite permissive with regards to “only half understanding something”.
          With MCDs on the other hand, one needs to chose what to cloze. Hence every time there is an implicit assumption (or at least a bet) that the kanji/particle being clozed is the only (or at least noticeably the best) way to fill the hole. This means that making clozes of stuff you only half understand is much riskier.

          Because of this, I’ve always felt that the stuff I was comfortable tackling with MCDs was of a somewhat lower level (at least reading-wise) than the stuff I was comfortable tackling with sentences (though I do feel that a lot less now that I’m at a reasonably higher level).

        • I’d like to see this article as well. I’ve seen Mike (aka Tokyostyle?) on AJATT+ a few times and he does present good ideas. An article on improving MCD’s may give me some good ideas on how to perfect my personal style.

  6. Honestly, AJATT is just one giant mess of a website. It’s a wonder to me that anyone actually goes there to try to get information about anything. I would sum up his website as follows: A lot of ranting/ rambling and self-sensationalized writing with a dash of perplexing humor that I’m pretty sure only he finds funny. The epitome of TL;DR.

    I’m sure that this post of yours will help a lot of people that need more succinct information.

    • Are you serious??? That site is a great place! Language learners from all over gained great knowledge from there! granted more so from his older articles. But a mess? I guess to each his own. If anything even this site may quote or use some things from AJATT because what Khatz says just makes sense to the average language learner.

    • I agree so, so, so much with this statement. I think his techniques are sound, but his presentation is appalling. I also took exception to reading through several paragraphs of rambling only to realize that the “article” was just an attempt to sell me something. That lack of transparency really irked me.

      I feel like Adshap has taken the lessons from AJATT and added his own insights and then presented them in a clear, concise, organized way. I feel like JALUP is “AJATT done right”.

    • I feel like a second counter argument should be inserted here to defend AJATT at least a bit. I am not blind to the fact that the website does get quite a bit of criticism from people, even on the website itself. However, Khatz even says himself that its a bad idea to only use his website due to the fact that he doesnt know everything, his presentation style is rambling and ambiguous, and he doesnt present everything all at once like most people like. However, if you take the time to read through his stuff and filter out the fluff (he even says skim through the articles and read the bolded and highlighted stuff to avoid fluff), you have sound ideas. I personally started with AJATT and it hit home with me. The only reason I ever found out about JALUP is because I read an article on his website saying you should look for other resources.

      Again, everyone sees things in a different light, but rather than bringing down another website, just say it wasnt for you. For me, I like both AJATT and JALUP. ja・ミニマル also has some good things to offer when it comes to simplicity of study and using actual studies on language learning to improve your individual learning style.

    • AJATT is what honestly started so much of this site. If you go there you will find that Adshap himself had a success story that Khatz posted. Adshap followed AJATT for awhile and then branched off on his own to make this site. It has a ton of information that is useful because the guy took a year and a half long experiment with fantastic results and then shared it so people who had doubts (like myself) could learn a language in an extremely FUN way. Keep looking and reading, especially his main articles in the table of contents and you will find boatloads of helpful information similar to what you can find here. Even though the writing styles are completely different.

Leave a Reply

Your email address will not be published. Required fields are marked *

HTML tags allowed in your comment: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>