Before the recent Daily Lessons change, I had two rules of thumb about lessons: I’d only do lessons unless/until
- I had less then 110 items in Apprentice, and
- According to Ultimate Timeline userscript, what I thought of as my “break-even” point — which was at what point in the future the number of predicted reviews per day was less than 100 — was less than 4 days.
(That is, with my 5-item lesson batch size, if I had ≤ 105 items in Apprentice and < 375 forecast items at 3.75 days, I was good, if I had 450 at 4 days, I’d stop lessons until the situation resolved.)
This was less about tackling specific problems (I’ve become more and more Zen about that as I’ve gotten past Level 30; now I’m just ruthless with marking things wrong via the Double-Check script, even if I guess right, and let SRS magic take care of it) than about dealing with overload.
Since the Daily Lessons feature arrived, I’ve just been using it (for me, that means 15 lessons a day) without modification. I was skeptical, and assumed I’d at least want to make sure I get new-level Radicals out of the way faster than its algorithm does—but it seems to work much better than I’d have predicted. It has caused my queue to bop up above 120 Apprentice items from time to time, but my discipline before sometimes resulted in “wasted days” when I had as few as 60-ish Apprentice item in the queue.
I think I skipped the Daily Lessons on three separate days when I had particularly bad performance with lots of stuff slipping.
Next month I go on my first totally-offline vacation since starting WaniKani (thanks, pandemic), and I’m trying to decide whether to just keep going normally until I go offline or to ease off beforehand. I think it’ll depend on my precise level progress as I get close (right now it’s taking me about 20 days a level, which seems about right for a somewhat <15-a-day new-item average).