Anki for exams

The computer science exams I take usually heavily rely on understanding. Learning for those exams means doing tasks that require applying the knowledge acquired. Memorizing facts almost never plays a big role. Quite a few exams are even "open book" — bring as much non digital material as you want.

Last semester I took two psychology classes with exams that were more or less the opposite. Being used to memorize stuff (kanji) with Anki, I thought it might be most efficient to use it for exam preparation. Because I was curious how much time I'd spend I recorded the whole process. Here's how to pass a 5th semester psychology class (Occupational and Organizational Psychology) as a CS major:

The challenge
The lecture consists of 386 slides. The exam can be passed with a very good grade solely based on the information given on the slides plus what is being said during the lecture, but students are advised to also look into referenced books and papers (I didn't). The exam consists of (iirc) 50 multiple choice questions, each of which has 5 answers of which between 0 and 5 are correct. Choosing a correct answer or not choosing a wrong answer results in 1 point. Not choosing a correct answer or choosing a wrong answer results in -1 point. The overall points for one question are reset to 0 in case they're negative. Example questions given for the exams go as deep as asking for names of authors or effect sizes of specific studies.

The solution

  • Attend all the lectures, annotate the slides (yellow boxes in the image above)
  • Start preparing for the exam about 3 weeks in advance, within these 3 weeks:
  • Spend a total of 14 hours to write a summary of the lecture's content (25 pages)
  • Spend a total of 9 hours to create Anki cards (477 cards)
  • Spend a total of 12 hours reviewing these Anki cards (2775 reviews)
  • Squeeze in 5 days of preparation for another exam with less material to learn
  • Pass with a not so great result, having spent a total of 35 hours on exam preparation (not counting lecture attendance)

Anki stats


28 hour day

I recently stumbled upon xkcd's comic on the idea of a 28-Hour Day. Having been interested in polyphasic sleep for a long time (without ever trying) and this being similar, I checked whether or not it would work out with my current situation of temporally fixed obligations throughout the week. Turned out that there would be very little overlap (blue=sleep, green=obligation). And the two cases where a lecture falls at the very end of a day ... just schedule two/three hours of sleep before that, right?

Wrong. Instead of extending my day by 4 hours per sleep/wake cycle I ended up with 3.5, 0.5, 2.5 ... -6. Friday, 1:30 a.m. I was super tired after having been awake for a mere 7 hours, went to bed and slept for 8 hours straight. This is how the whole week turned out. More like a rubber band that I stretched a bit than the intended wrapping around over the course of a week.

Lesson learned: if you sleep a lot anyways (8-9 hours per night in my case), the 28 hour day may not be for you. ^_^


漢字 — 2000+

99 kanji, 575 words, 489 days; Anki stats, kanji so far.

Whew... Haven't done one of these in a loong time. Who would've tought I'd actually reach 2000 ... In my first kanji progress post, 100+, I wrote “there will — hopefully — be a total of 21 of these in the end” — I'm really not sure I'll ever get to 2100. Since realizing at 1800+ that new kanji as a main focus isn't worth it anymore, it took me 115 days to get the first 100 on top of that and now 489 days for the second. If factor of increase in time stayed the same I'd reach 2100 in August 2022. ^_^
Quite the deviation from my extrapolations at 1000+, where I took a closer look at what I had leared so far. Here's the same thing for the present situation:

Additional data:   [values from 1000+]


  • Grade 1: 80 of 80 (100.0%)   [100.0%]
  • Grade 2: 160 of 160 (100.0%)   [100.0%]
  • Grade 3: 200 of 200 (100.0%)   [88.0%]
  • Grade 4: 200 of 200 (100.0%)   [64.0%]
  • Grade 5: 185 of 185 (100.0%)   [55.1%]
  • Grade 6: 178 of 181 (98.3%)   [51.4%]
  • JuniorHS: 775 of 934 (83.0%)   [21.6%]


  • 2000 total unique kanji   [1002]
  • Jōyō: 1910 of 2136 (89.4%)   [45.7%]
  • Jinmeiyō (regular): 66 of 641 (10.3%)   [3.1%]
  • Jinmeiyō (variant): 1 of 145 (0.7%)   [0%]
  • 23 non-jōyō kanji   [5]



27  26  25  24  23  22  21  20  19  18  17  16  15  14  13  12  11  10  9  8  7  6  5  4  3  2  1