What’s your opinion of a reasonable number of guru items to have at any one time? I have 304 right now.
I’ve heard 100 apprentice, and that’s working well for me, but I think my gurus are currently bloated with leeches.
5 Likes
I currently have 620 in Guru but I don’t know how much of that is leeches to be honest 
I have no “right number” but if hearing 620 makes you think “hell no” then I would say something like “up to 619 max”
8 Likes
im on around 800 guru right now and have consistently been over 600-900 guru for at least 20 levels
2 Likes
I have 300-500 gurus most of the time. You can bring it down but you’ll get returning items failing back to guru plus apprentice items moving up.
So it’s not that easy to maintain
2 Likes
I usually have 500-600 guru. Don’t think I’ve been below 500 for a long time!
Edit: I currently have 624. That’s probably not unusual.
2 Likes
I think that, as long as you don’t get them wrong, Guru count doesn’t really matter. Apprentice count is the only important one.
If you care about the math behind it and the load of each SRS stage, here is a mathematical explanation of why Guru itself doesn’t matter:
"Long mathematical details, you've been warned
Edited out because no one ever cared about the math. Basically one item more in Appr 3 is 14 times heavier than one more in Guru 2.
TL;DR Guru leeches are a problem because they will fall back to Apprentice, enlarging the Apprentice count in the near future (possibly multiple times). The Guru count itself is not important. The leech count is.
I haven’t found a way around leeches myself 
, but I think studying them in sentences outside WK may be the only real way to squash them for good.
4 Likes
Finally reached 0/0 again after three weeks of struggling! At this point I just want to reach the magical lvl 30 and then hold off on new lessons for a whole month or so for my daily review amounts to calm down so I can find some time to do more leech review sessions and focus more on bunpro.
5 Likes
(25%) (25%) (25%) 
… I care about that math but if it’s too much pain in the neck to retype your explanation then thank you for the figures!
2 Likes
Me: Uses reorder script to isolate radicals because radicals are the absolute most important thing that I can never get wrong in order to maximize level up time.
Also me:
5 Likes
Ghost here, checking in! Can someone please edit the board for me? I am currently level 23 and my current speed is 8 days 15 hours. You can find my username hanging out at level 12, whoops!
4 Likes
Done, congratulations on you multiple level-up!
P.S. Not sure, but it seems some rules have changed and basic members can now modify the wiki as well. If you manage to do it, tell us and we’ll change the first post (at least for the new users).
5 Likes
I can’t find the post or I’d quote it, but one of the mods did say awhile back that they discussed the rules and decided to allow basic users to edit wikis. So yes, anyone with the basic badge and up can edit 
Edit: found it!
5 Likes
I’m trying to keep Guru under 400. Which doesn’t always work, but they do take up a lot of time, and a lot of those are leeches for me. Sometimes I even fail them a bunch of times on purpose to make them go back to apprentice 1. That way I can start all over on that item, and relearn it.
I went from almost 500 guru to almost 400 in about a month or so. Something seems to work.
I should do some more leech training regardless.
3 Likes
Not a problem, I deleted it also because I was unsatisfied by it and because it lacked a real conclusion. Here comes an improved version.
I feel you underestimated what long meant, anyway…here goes:
First of all a review on the SRS stages and their timings:
Let's give them easier names:
Apprentice 1 —> A1
Apprentice 2 —> A2
Apprentice 3 —> A3
Apprentice 4 —> A4
Guru 1 —> G1
Guru 2 —> G2
Master —> M
Enlightened —> E
Burned —> B
How do you estimate the daily review load of each stage?
We are talking about the average here, so we do just that.
You divide the item count of the SRS stage you are considering by the timing of that stage:
A3 / 1
A4 / 2
G1 / 7
G2 / 14
M / 28 [or maybe 30…unsure of what WK considers a “month” here]
E / 112 [or maybe 120, as above]
Apprentice 1 and Apprentice 2 are special:
You may be tempted to divide A1 and A2 as well:
A1/0.17 and A2/0.33 [i.e. 6*A1 and 3*A2]
But if you do your reviews, A1 and A2 are always close to zero (no one does 20 reviews every 4 hours). It may go high because of just finished lessons. Basically the number you see fluctuates too much. If you see items constantly at A1 and A2, they are leeches/mistakes (more about that near the end).
But you do have A1 and A2 reviews, so we need to consider them (but only once per day).
Formula for Daily Average Review Load:
A1 + A2 + A3 + A4 /2 + G1 /7 + G2 /14 + M /28 + E /112 [rev/day]
Let's take a real example (mine):
A1: right now negligible (1)
A2: right now negligible (2)
A3: 137
A4: 144
G1: 286
G2: 213
M: 381
E: 1714
Load = 137 + 144/2 + 286/7 + 213/14 + 381/28 + 1714/112
= 137 + 72 + 41 + 15 + 14 + 15 rev/day
= 294 rev/day
Guru Load in Percentage = 56 / 294 = 19%
The average Wanikanians has less Apprentices so, relatively Guru will be slightly more important than it is in my situation.
Enlightened shold be higher (but I haven’t started burning yet, so I’m still in the transitional phase)
Master should be higher, but I did skip vocab for a month…
I also put A1 and A2 as zero, that’s a mistake, putting the number of lessons planned for today would have been better, but I didn’t want to invent numbers.
I suggest you do a similar calculation to have an idea of the relative weight of Guru in your particular case.
Now let's compare the load of Apprentice 3 and Guru 2:
Looking at the load again, you divide A3 by one, G2 by 14. (That’s where the 14 comes from: 14/1)
The explanation is also easy: doing one more lesson every day means increasing the A3 count by one. If A3 is always +1 for 14 consectuive days, then (once those items reach G2) you will have G2 + 14.
In the same way, having A4 + 10 is the same as G1 + 35 (7 days /2 days = 3.5)
Let’s take two extreme cases:
Guru 900 (300 Guru 1, 600 Guru 2)
Load (guru only) : 86 rev/day
Apprentice 90-150 (depends on the hour of the day) (30 A3, 60 A4)
Load (Apprentice 3 & 4 only) : 60 rev/day
[But what about A1 and A2? If we have 30 A1 —> 30 lessons every day —> every day, at some point (for a few hours) you’ll have 30 A1 reviews and 30 A2 reviews]
Load (all Apprentice stages) : 30 + 30 + 60 = 120 rev/day
Note 1: mistakes/leeches are not considered, read the end.
Note 2: 30 lessons per day is higher than allowed (you’ll end up hitting 0 lessons at some point).
A constraint on Apprentice already limits the Guru count, and the equivalent of 100 Apprentices is 420 Guru:
Let’s say you do 20 lessons every day. (In an ideal situation i.e. no mistakes)
You’ll have on average (steady state):
A3 = 20
A4 = 40
Total Apprentice count =~ 100 (20 + 20 + 20 + 40)
G1 = 20x7 = 140
G2 = 20x14 = 280
Total Guru count =~ 420
[Again, even if you do 20 lessons per day (you’ll probably do more at the fast levels), Apprentice will not be 100 but higher, and Guru won’t be 420 but higher because of mistakes]
An overrall simplified view of everything:
Let’s take the life of a solitary review.
You’ll meet her 4 times when she’s Apprentice. 2 times when she’s Guru. Once at Master. Once at Enlightened.
The load of Guru is only 2/8 = 25 %
We have 8000 lessons, but the percentage in an ideal case does not change. If it does it’s only because of mistakes.
Without mistakes, 12.5% of your daily reviews are due to A1, 12.5% are due to A2,…12.5% are due to G1 and so on.
Mistakes are a mess:
Mistake break the flow. They increase Guru count, and even more so the Apprentice count. Thus making those slightly “heavier” then the last stages. They are also too hard to analyze, more on it at the end.
The Highway of Durtle-drawn carriages.
Imagine a highway. It’s divided in 8 sections of different lengths. At the end is an eco-friendly incinerator.
The first two sections are very short. The third is 1 km. The fourth is 2 km. Number 5 is 7 km, it’s called G1. Number 6 is 14 km (G2). Number 7 is 28 km. The last stretch is 112 km.
(Our durtles go at 1 km/day)
At the end of each section is the Police Crab. They are responsible for a Quality Check. If it’s below standard, the carriages are sent back at the beginning of the section (or several sections before this one).
If 20 carriages enter the highway everyday, the Police at A1 and A2 on average will see 20 carriages per day. The same goes for A3 and A4. The Police Crab at G1 is responsible for a very long 7 km section, he has 140 carriages on it, but he only checks who exits. He too will see only 20 carriages every day. The same goes for G2, M and E, even though the number of carriages in their section is huge.
Real numbers and Conclusion using Plantron's stolen data
In short, the count of reviews on a specific stage is not important. The outflux (i.e. the Review moment) matters. But that depends on the influx (equal to the successful rev/day of the previous stage). So, unless you do a crazy amount of lessons, and Apprentice count goes sky high, Guru will also be reasonable.
As for the load, Guru is responsible for 25 % of the total reviews (if you have reached the burning point). Apprentice (all four) is responsible for 50 % of the reviews.
Qualitatively I can say that Guru will go a little bit higher than 25 % (maybe in the 28-33% range, but it depends on you accuracy in the specific stages) and the same for Apprentice (somewhere in the 52-55%).
This stats by Cryptomilk tell me that 33% of his/her total vocab reviews are Guru.
https://community.wanikani.com/t/guru-1-interval-too-long/41770/56
Using Plantron’s data instead, I get (9215 + 8883)/63335 = 28 % (close to the “ideal” 25%) for Guru Vocabulary.
Again using Plantron’s data for Apprentice Vocabulary, the result is 53 %.
You can use the same tool and see exactly how much the load of Apprentice measures against Guru in your specific case, but even if you are not Plantron, it should be close to those numbers.
Finally, don’t be scared by the high number in Guru as their load is low. And don’t put a constraint on it, because limiting Apprentices will automatically limit Guru as well. So instead of using two constraints (one for Apprentice and one for Guru), just set the Apprentice constraint lower (90, 80 etc…) if the number of daily reviews is unbearable.
The author apologizes
“I never meant for it to get that long”, he says. But witnesses claim they saw him smiling maniacally while writing at lightning speed. It doesn’t bode well for his upcoming trial.
6 Likes
You know you can check their edit history, right?
Basically, they had said that Apprentice III takes one day, but Guru II takes fourteen days, in five thousand words.
2 Likes
:DDD omg THANK YOUUUUUU!!! I know that I didn’t comprehend a lot of it and I was feeling really guilty until you broke out your simplifications and analogies!!! Magnificent and very very kind of you! Thank you very very very much!!! And thanks also to plantron and cryptomilk…too shy to tag them…
Daaannnnnng awesome!! 
2 Likes
wut
Edit history?!
o _ O
1 Like
Awesome! Where are you taking yours?
I need to start up on Bunpro. I did the trial earlier in the year and got distracted. I know that I need work on reading/grammar for sure, so getting Bunpro going is a big goal of mine.