Concerned with Lessons

Yeah, I was concerned with vocab getting too far away from the relevant Kanji. Right now it is off by about 2 levels, so I will probably have to just put off leveling in order to get caught back up.

3 Likes

Self Study with the leech filter from Additional Filters.

Thank you!

1 Like

Honestly, you should never level up/do kanji if you have vocab lessons at -2 levels.
-1 level is ok, you unlock some new ones just as you level, and can start the kanji first for level up. but DO get them before you level again!

They reinforce the kanji and are a vital part of the process.

With that many I wouldn’t even start the kanji. Focus on getting your vocab lessons to zero before continuing. Getting the lessons down is more important than reaching the next level fast, as you haven’t truly finished the previous level yet.

I recommend reaching Zero lessons at least once per level. Get all of them out!
The easiest way to do this is to do the vocab first, so you are sure to get the zero.
The harder way is to do kanji first, and then get all vocabs out before you guru the kanji and unlock more vocabs in 3.5 days (at best, more if you fail some)

I unlocked so many lvl 50 vocab just before level up, so spent a week doing vocabs before even starting my lvl 51 kanji. I could technically cram this level in 3.5 days (fast level), but will instead spent about 3 weeks, to divide up the vocabs and kanji and get my sweet zero =)

5 Likes

i’d advise doing the 0/0 challenge: at least once per level, get your lessons/reviews down to 0/0. it doesn’t have to be a hard rule, but it’s a simple way to be sure you have done all the vocab lessons from the previous levels.

if you’re re-ordering lessons (with a userscript or one of the phone apps) then you risk building up a backlog of vocab lessons. this ought to be avoided, as the vocab lessons are designed to supplement and reinforce the kanji lessons.

3 Likes

You may also look at Item Inspector.

  • Self Study quiz with the leech filter is for quizzing linke you do in reviews.
  • Item Inspector displays the leeches on the dashboard, You can reveal information on the leech in a popup by moving the mouse over the items. You can use this for studying. You recite the meaning and reading aloud and you move the mouse over the item to see the answer.
2 Likes

I disagree. I randomize my vocabulary so that I don’t get a lot of words that use the same kanji all at once. This helps in several ways.

  • Blocks of words that use kanji in the same way come up less frequently. This reduces the effect of “I only got that right because I saw that other thing earlier”, which I personally hate. Usage spikes like this can cause you to think you’ve learned things better than you have.

  • Words that use kanji in different ways are less likely to come up together and cause leeches as they ‘fight’ each other for headspace.

  • Words that use particular kanji are spread out over a longer period, giving more even exposure to the kanji. You’re less likely to forget a kanji over a long dry period.

To get enough words to reasonably break everything up, I randomize over the last five levels. Level n-5 will mostly be complete, n-4 will be a bit less done, etc. When I level up, I do vocab lessons to finish up lesson n-6, which is usually <10% of the level by that point.

This, however, I completely agree with. Whatever method used, you want to do the vocab to reinforce your understanding of the kanji. It’s the only way to really understand the kanji and start to get a feel for the nuances of how it’s used. I do lessons at a rate to maintain an approximate 3:1 ratio of vocab to kanji, so I’m not leaving vocab on the table.

I find this approach works really well for keeping kanji in active recall and might even be more efficient. I seem to recall reading that what matters for long term recall is number of exposure events. Drilling is less useful. And essentially, that’s what I’ve done, increased the number of small events and decreased the number of larger clumps that become more like drills.

I should mention that I also learn vocab externally with Anki decks which are WK level linked (ie, use kanji I know from WK, but may not be vocab on WK). So there’s some redundancy in the system. I can’t say it would as well for me if that weren’t the case, but it has been working very well for me.

1 Like

Haven’t seen that since level 2 or 3 probably. Last time I had 0 lessons (happened once) I had reviews pending. And when I finished the reviews, I got a pack of new lessons. And a levelup very soon after.

There’s about 100 vocab per level. That is enough to randomise them, no need to leave them 3 weeks behind. And they have a tendency to get mixed, if you get two of same kanji/reading in one go that you didn’t know well enough you might fail the first but get the second since you had the reminder. Problem solved, they are no longer together :wink:
They can now be weeks apart if one Guru’d and the other didn’t (or if it was Guru 1 review it is two weeks apart)
The SRS handles that problem.

If they are left too long the Kanji you learned have faded too much before you get to the vocab, making them more likely to leech at first, as you have to go back and re-read the Kanji mnemonic to remember it.

But as long as they are done, within a reasonable time really. If you mix two levels up you have 200 vocab to mix. No need for -2 levels (3 full levels of vocab). There are usually about 3-5 vocab per kanji, some already distributed in other levels (as it combines with future kanji)

I’ve had 0/0 every single level. To get it (if you want it), calculate when next kanji Guru and try get all lessons out before that time (and make sure to stay zero on reviews as much as possible)

I went too fast, got too many left over vocabs, ended up resetting. I’ve reset twice in fact!
The best advice I can give, with my experience of trying both ways, is aim for that sweet zero. Not only does it help your progress, but also gives a great motivation and boost.

But everyones WK journey is different. How much they knew, how much they do other than WK and just what works for them =)
I’m only sharing my experience and what worked for me and why =)

1 Like

I really don’t think this approach is a good idea. Like Toyger said, there’s plenty of vocab in a single level to still randomize the vocab. I’m not saying your bullet points are invalid. I just don’t think intentionally falling behind by five levels is a good way to deal with those concerns. I think the downsides from falling behind outweigh the concerns you mentioned from keeping up.

2 Likes

It’s not really falling behind by 5 levels, though. Since the furthest back level is like 90% done, the level 4 back is like 75% done, etc. So I’m only ‘behind’ by around 2.5 levels at the end of the day. And since even the closest levels have some exposure, it’s not as bad as falling behind when doing vocab in strict order, where the more recent levels have no exposure.

Really, I know it’s unintuitive, but it is working very well for me. It might be because I have those external sources to help. Or maybe because this is my second time through (I reset from 21 to 5 at the beginning of the year) and I’ll hit a roadblock when I get into the 20’s where it’s totally new material. If so, I’ll change my approach. For now, though, this has made WK way, way, easier for me for the reasons I’ve already mentioned.

Yeah, if it works for you that’s great. I would just be reluctant to recommend this approach to others.

1 Like

100 really isn’t, though, because they don’t unlock all together. If you do your kanji in a block, maybe that would work, but if your kanji are guru-ing one at a time, you will still have clumps, especially for the very common kanji. Initially I started with just one level back, then two. Both had unacceptable spikes. Eventually I let it ride. I settled on 5 because that was where this method naturally seemed to settle, where a level would finish anyway just by random chance.

This would be true if it were just 2. When there are a dozen, the whittling down takes much, much longer. When there is more than 3 or 4, they fall out of sync, for sure, but end up back in sync when I would miss the next one. Note that it’s not just a single review session where clumps are a problem. If you de-clump at Apprentice 4, for example, they’re only apart by a few days, which can feel like much the same problem at Master or Enlightened when you haven’t seen a kanji for 4 weeks and suddenly you get all the words for it showing up together over a week or two. It’s… lumpy.

So, yeah, that would be a worry, but I find that at my normal rate of around 20 lessons a day, A newly guru-d kanji’s vocab will show up within a week of unlock, and generally much less. Since the kanji itself is showing up in a week to Guru2, that seems just about ideal. As soon as the kanji is going into higher intervals, the words start occurring to keep the exposures going. In fact, I think the weakness of this system may be that kanji never get a chance to truly decay with the constant reinforcement because there’s never a dry period where I don’t see it for a few weeks. Quite the opposite of leaving things too long.

Anyway, yeah, I do understand your concern. It is an unintuitive system. I only came up with it after I had to reset at the beginning of the year (Covid, ugh). But it works surprisingly well.

When I was beginning with WK, I read a lot of different approaches here to how people tackled the problem. The perils and pitfalls of each process. We all learn very differently. As a new learner, I really appreciated seeing all the different things I could try, especially the ones that seemed to only work for a few people or were unintuitive. They encouraged me to try and find the process that worked for me. To try things out.

So I do hear what you’re saying. I probably should have caveat-ed my initial message with a ‘this is weird, but it works for me’ message more than I did. But I’d rather post it, have this kind of rational discussion of differing viewpoints about the merits and flaws of the approach, and let the learner decide based upon that whether it’s worth trying based upon that.

2 Likes

There aren’t a dozen. Most are usually 2-3 of same reading. Rarely more (of same kanji/same reading/same or near level)

And after enough reviews there are rarely any left in clumps (if they still are together, it simply means I know them since neither fall down)
There is more than enough time to “forget” an exposure when in Enlightened for 4 months. You don’t get enough exposures from same/similar items in that period to warrant it "seen too much to say you truly remembered)

As for first, yes, if you do them as they unlock and did the kanji spaced you don’t get 100 lessons at once., and thus several vocab from same kanji at once. But you can choose to wait, and let them combine, if this is important to you. You don’t need to purposefully not do items from 3 levels ago to achieve the effect.

A point to consider though. WK has a set order they want to teach you the kanji and vocabs.
If you learn them in the wrong order you will find several “You know this reading already”, when you don’t.
This might be more apparent in the later levels, but even within the same level this happens.

I have given them the suggestion to just add the mnemonic to each, instead of saying “you know this already”, when it happened to me recently. I was unaware that a script I have randomised the order even when I didn’t ask it to do anything.
Since if you made the first one alright, but fail the second, it is better to have the mnemonic readily available anyway.
They have no plans of changing this as of now though.

I’m worried about both the same readings and different readings, which cause leeches. So double your count (at least, since more than 2 readings is also common). Also, since most vocab contain multiple kanji, there is again more than you might think. My ‘dozen’ is probably hyperbole, but it was both noticeable and annoying to me, and getting worse when I got to the 20s. If you didn’t find it a problem, good on you. I did.

We have different ideas of what’s a clump, it seems. In particular, at higher levels, it doesn’t need to be a single session.

I found it (and still find it) very annoying when it happens, but with this method, it doesn’t happen very often. Five levels may be more than necessary to break the effect, but as I’ve seen no negatives that outweigh the positives. If I see them later on, I will shift. One of the great things about WK is the meta-learning it provides. Learning how to learn kanji is, for me, really fun, so experimenting with these different approaches and analyzing what works for me is really part of the goal. I understand that not everyone will get a kick out of it, but I do. And I find the benefits (for now) more than adequate to keep doing it.

So I don’t purposefully wait to do items. I randomly choose items to do. Fate decides. Five levels wasn’t chosen, it evolved based upon how quickly the levels get done. Basically, the n-1 level (90% or more unlocked) has more undone vocab than the n-2 which has more than the n-3. n-5 is almost fully complete by this time, because I’ve been sampling from it for the last 5 levels of progress. So I naturally, by randomly selecting vocab, get mostly stuff from level n, a little less from n-1, less again from n-2, etc. It’s a constantly decreasing tail.

For me, five levels is about 6 weeks (at ~20 items per day), so each level’s vocabulary gets done over a 6 week period that overlaps other levels. Which means that by the time my kanji reach Enlightened, I’ll have done most of their level’s vocab to some level of proficiency.

One last thing. It’s not as if this is different from real reading in the world. In your reading, you get words at random, spaced at various intervals. Sometimes you’ll read something on a specific subject and will get a clump of kanji related to that subject all at once. Sometimes you’ll get isolated words that use a kanji you haven’t seen in awhile. That’s part of reading native material.

I like it. It feels more natural to me. Maybe I’m an outlier. We all learn differently. I have found this process to work way better for a multitude of reasons than doing stuff in strict order. I’ve tried, and possibly failed, to articulate these. Enough that anyone coming to this thread in the future will have a good idea of the method and be able to evaluate whether it is something worth trying for them.

Things that work for you or me or anyone else will not work for everyone. So all we can do is offer up what we’ve found to work and hope people find the method that works for them. That’s why we’re here, isn’t it? To help?

I do appreciate the advice, though I’m not changing my process just yet. You’ve mentioned flaws in the system that you see. I either don’t see them as flaws, or have found the things this system does well outweigh those negatives. I also appreciate the discussion. The best thing we can do here when helping new learners is to discuss methods in both their strengths and their weaknesses so that the learner can figure out what works for them.

tl;dr I’m not changing my system, and I won’t change your mind about it at this point. And that’s ultimately ok. Thanks for the interesting discussion. It’s really great that we could talk, and have such strong opinions about methodology without it getting heated. :wink:

No, 3 is the actual norm, and rarely is more. Look at the kanji page you will find half has only 3 vocab or less, counting all readings, and those with more usually have vocabs unlocked at later levels. Kanji who give more vocab than 3 at unlock are rare, if any. Yes, some kanji has a good number of vocab, but some I haven’t even unlocked yet at my current level when i looked at random kanji at lvl 21 (has some of the numerous kanji).

If you are fine with 3 vocab, with some different readings, then it works as intended. There are no half dozens at unlock that I know of :wink:

No, I have my similar vocabs divided in weeks and months, after failing one but not the other at Guru/master level. I’m not talking hours, or even days.

But honestly, the most important part is to actually do the vocab, late or not. Some try to just do the kanji and think that is enough, just to fail every review later (the vocab does help strengthen the recollection)

I find delaying them even a week from Kanji guru is enough to weaken the recollection for me. The kanji gets strengthened by having the vocab drilled shortly after first guruing it. And then, once both kanji and vocab has moved on to higher areas, they are more likely to stay in long term memory after their exposures.

1 Like

I don’t really care about 0/0 at all. Definitely not slowing down my progress for that. I can’t imagine myself being motivated by losing several days for numbers which only mean you have nothing to do atm. One can get them by just doing lessons at any speed and regularly messing up reviews, thus having 0 or even negative progress. I totally understand people, who are happy about good speed + accuracy though, because that means they are learning fast and well.