Estimate reviews per day


#1

As I reach higher levels, my number of reviews keeps climbing. So far, I’m okay, but I know eventually I’ll reach a level where the daily review count feels like too much and I fall off the wagon.

So I’d like to adjust the rate at which I do lessons or complete levels to keep my daily reviews (excluding apprentice-level items) below a certain threshold - let’s say 300. Is there any good way of figuring out what the right new lesson rate is to achieve this?

(Unfortunately, the amazing https://www.wkstats.com/, which has functionality to simulate all kinds of stuff, doesn’t seem to include this… unless I’m missing it, in which case, tell me!)


#2

You’re looking for [Userscript]: WaniKani Ultimate Timeline for daily review count, and [Userscript] WK Anti-burnout (or [Userscript] WaniKani Lesson Cap) to determine lessons per day to prevent burnout.


#3

Well, most people struggle with just 150/day (not counting apprentice, like you did). I think that you’ll be fine. Just spread the vocab lessons throughout the level.


#4

Ah, I love WaniKani Ultimate Timeline!

That lets me anticipate how many reviews I’ll have coming up when, but it doesn’t let me see how this will change as I complete lessons, does it?


#5

Well, it’s really just 1 -> 1.
1 lesson done -> 1 more review in queue.
The timeline will update as you complete lessons. (Say you complete 25 lessons at 5pm, it’ll say that you have 25 reviews at 9pm.)


#6

Right, so over the long term, reviews will build up as I complete lessons, but they’ll drop off as they age out (i.e., I complete them successfully and the review interval becomes longer).

I’m wondering specifically about how different lesson completion rates will affect the “steady state” peak number of daily reviews.


#7

The script creator for WK Anti-Burnout recommends the Lesson Cap that’s based on theirs, and I agree.


#8

Ask for stats and you shall receive!


Blue: lessons left to do. Yellow: reviews waiting in the queue. Red: reviews calculated for the next day. Green: how many lessons I did the day before.

The higher the green is, the higher the red becomes.


#9

It’s an overall better script. I’m not a coder, I just had an idea for something, but nobody is in the business of realizing other people’s ideas, so I realized it, and hoped someone would improve it. And they did!


#10

Oh wow… this looks so cool! It took me a while to wrap my head around it but it seems very useful.

Reviews calculated for the next day, though - does that consistently match up with the number of reviews you did? Or is it potentially off due to reviewing the same item multiple times in one day?


#11

I think you’re looking for something like this Google spreadsheet.

You can edit the part highlighted in yellow to check things out. It’s based on how long it takes you to level up rather than how many lessons you do a day, so you’ll have to convert from one to the other. For reference, I do 3 kanji lessons a day (plus some vocab of course), which has me usually level up in about 15 days.

Based on this, I used the number 300 in the input (since the spreadsheet says a 2 week level up time is 320 and lower is slower), which says I would have 80 reviews a day. Of course, this includes all reviews, including Apprentice items. Obviously this can’t account for items you get wrong, so the number of reviews you actually have will be higher than what it tells you. For comparison, my daily reviews are usually between 90 and 110, depending on the day, compared to the 80 that it says.


#13

Ha! This is almost exactly what I was looking for, yes.

Unfortunately, I’m a huge tool, so after posting this thread (but before seeing your reply) I ended up creating an Excel simulation of what would happen with different error rates, numbers of lessons per day, etc. I assumed a rough baseline of 150 items per level based on A question to the fast learners ~8 days per level.

My big takeaways were roughly as follows:

  1. My review “pipeline” is already almost full. Even though my number of reviews per day is increasing right now, it will level off in the teens of levels. So I may have been worried for nothing.
  2. The point about not doing your lessons all on one day is really important! I’ve read this mentioned in other threads but I had kind of brushed it off, because, let’s face it, it’s hard not to want to see the cool new radicals/kanji when you first get the chance. However, breaking up the lessons into equal increments over the length of a level gave a maximum reviews-per-day of 259. Doing them all as soon as they became available yielded a maximum reviews-per-day of 455.*
  3. I’m really bad at Excel simulations. (Seriously, I had to rewrite my sheet three times because I realized halfway through that the way I’d set it up was internally inconsistent.)

*I’m making some simplifying assumptions about when they become available. This is also using a dubious “miss” rate - I’m assuming misses are IID, which is probably not correct, and affect 5% of reviews, which is probably too high. I’m also assuming a 7-day level time, which is probably not remotely correct - I average around 8 and expect to get longer. Bottom line: don’t use these numbers for anything important, but they were helpful to me.


#14

I did roughly ~15 to ~20 more than what’s on the graph due to the new lessons I did + the 4 hour + 8 hour items coming back the same day. On average I did 8.5 lessons a day from level 10 to level 16, and you can see how low the numbers are even with the burns coming in. That’s 60 lessons a week, and I had an average of 43 items in Apprentice during that period, 237 Guru, and 371 Master.


#15

If you expect to slow down a bit but still want to level up consistently, this script may be helpful for you. It lets you specify the number of lessons you want to do in a single sitting, broken up by type. That way you can take exactly the number of kanji necessary per day to ensure that you level up when you want, while also allowing you to balance it with vocab lessons. It makes it really easy to spread out your lessons over many days, without having to refresh over and over to get the balance you want.


#16

Because you need to review an item 8 times to burn it, assuming you do the same number of lessons a day, and you get 100% correct, your daily reviews will by 8 times the number of daily lessons (eventually).

if you do 20 lessons a day, eventually youll do 20 burns from 4 months ago, 20 enlightened from 1 month ago… etc and comes out to average 160 reviews a day. of course if you get reviews wrong, youll review those items more often… hard to account for that long term tho


#17

As the reviews have started piling up I find the best way for me not to have to many is to only do new lessons when my apprentice column says 50 or less.

I also stop on every item I get wrong and write it out with the mnemonic. I’ve found this greatly reduces the chance of seeing it again or it becoming a leech thus keeping the total review count down.


#18

Yeah, 8 * number of lessons per day is a pretty good estimate if you have very high accuracy.

If you want to factor in mistakes, you can check out my post:

A closed solution formula to estimate your daily reviews is R = 8 L / (3 A - 2)

  • R is the # of reviews per day
  • L is # of lessons per day
  • A is your absolute % accuracy across reviews

#19

seanblue writes:
If you expect to slow down a bit but still want to level up consistently, this script may be helpful for you. […] [Userscript] WaniKani Lesson Filter

YES! This is awesome. It’s to resist doing all of the new radicals/kanji as soon as they come up, so this makes it possible to (say) do radicals/kanji on day 1 and then space out vocabulary at 15-20/day. Sweet! :smile:

BobaGakusei writes:
A closed solution formula to estimate your daily reviews is R = 8 L / (3 A - 2)

Nice! This is a much better way in the event that one really is doing a constant number of lessons per day. (Which… with the help of seanblue’s plugin, I vow to do from now on, honest.) [Somehow I missed that the drop was 2 levels instead of 1, which also means that my numbers from simulation are all incorrect… again. Sigh.]

tl;dr is: WaniKaniers are really the best.


#20

I was doing 300 a day from whenever that started until level 27, at level 27 I stopped taking on new lessons.

I then got even lazier and stopped doing my reviews because it was relentless for two months after stopping.

Ultimately it’s my laziness, but I’m back since yesterday and have completed 200.

rip social life


#21

I have always wondered what’s so bad about doing 200 reviews in one session. Maybe I just don’t understand it - is it just 200 items listed under the “review” button on the dashboard? if you’re using your mobile phone then good luck for the night… but on your desktop is should take about 20 to 30 minutes. maybe a bit more on the first apprentice levels and a lot less just before guru’ing them.
at least that’s what I’ve had for the last couple of months. each day about 200-300 items to review (sometimes more because of the 4-8 hour intervals) I’ve been steadily accumulating items at apprentice (from <100 to about 250 now) because of leeches, but they generally don’t bother me that much.