So one of you guys asked me how I go about my Anki workflow and what my entire process looks like. Here it is.
The way I see it, pretty much every method you learn should always end up in a workflow or else you won’t do it anyway. I’ve learned a lot about how to learn and memorize faster without using Anki so you can say that Anki’s just something that improves my efficiency a lot.
That basically means I can do well even without Anki; I can always create a workflow using the principles behind Anki’s effectiveness, anyway.
It’s been a while since I last used Anki, so I might be missing some details in this post. It’s not a big deal, though–every guide I post is mostly evergreen and doesn’t change with Anki’s version or how old your dog is.
In this post, you’re going to learn some of the things that aren’t talked about a lot in the Anki world much less the learning world–a complete learning workflow based on principles. I’ll throw in some tactics here and there, but I hope you take away some lessons that would forever change how you make your cards.
My 3-Step Workflow Before Card Creation
When I’m studying and using Anki, I figured I always use three steps before creating a question.In my eyes, what happens before you even answer a flashcard is at least as important as finishing all due cards every day. And I’m not kidding–just a simple tweak can easily half your learning time for some cards. You’ll see that for yourself.
It’s funny because when I realized I can do all of these in a single process, I literally got fired up, deleted all my cards, and started from scratch. Now, I won’t recommend you do that. Instead, you can just modify the cards you already have as you go. (I have guidelines in Step 3.)
Author’s Note: Use this guide as a reference material rather than a “3 steps” article that gets forgotten after 10 seconds of leaving the site. In fact, I want you to create Anki cards to test your knowledge on this article. Besides that, enjoy.
Step 1. Inspectional Reading
To determine what’s really important in a chapter, I look at three areas:
- Chapter objectives
- Subheadings within the chapter
- End-chapter questions
Each has a purpose–they’re the author’s way of telling you what’s important. But if they’re not available for some reason, I just base it off the length of discussion.
Good textbooks use these elements to emphasize the important points; they’re good shortcuts to determine which areas to actually put your attention to. This type of reading, by the way, is called systematic skimming. I didn’t do it the way author Mortimer Adler does it mainly because I haven’t read How to Read a Book yet when I was using Anki. You can learn more about that here.
Step 2. Feynman Technique
While I’m reading, I always use the Feynman Technique (in my head) to simplify the concepts. A lot of technical jargon can be found in books, but the reality is you can simplify them into everyday words and use analogies to describe them. This way, complex concepts become easy to understand and remember.
By the way, the Feynman Technique is otherwise known as elaboration in the realm of learning psychology.1 Honestly, i don’t know why it’s “elaboration” because it’s more like mental compression. The point is, if you can turn multiple ideas into a single memorable concept, it’s guaranteed that you already understand it. (Remembering it is a different story.)
Otherwise, if you just re-read the material without using Feynman technique, active recall methods, or memory techniques, then it’s only a matter of time before you forget it.
Step 3. Creating High-Quality Questions
Something that isn’t talked too much is that creating high-quality questions is a skill. I’ve recently pondered upon my criteria of creating good questions by asking, “what allowed me to answer a high volume of cards in a short amount of time?” Common sense says: because I make questions that are quick to answer. And for that to happen, I figured the questions you create must meet these three criteria: High specificity, Future-proof, Atomic.
- High specificity. Ambiguity slows you down. If there are two reasons why X is Y, then you should make a card that asks for “the two reasons why X is Y”! Don’t make questions like “What are the reasons why X is Y?”. Instead, make it more specific by adding some context. This is one of the most powerful ways to speed up your total review time.
- Future-proof. If you think the question would be hard to understand in the future, you can add even more context and make the question even more specific. This is one way of future-proofing. Another way is to add reference material to the card–which means adding a screenshot of where you got it. If you forget the card in the future, there’s always an excerpt available to re-learn it. No need to go back to your material. More on this later.
- Atomic. Complex cards should be broken down as much as possible. This follows the “minimum information principle“. Concepts are like molecules–sub-concepts are the atoms. For example, in Electronics Engineering, there are lots of output variables that change due to a single input variable. Let’s say Temperature increases. Junction Resistance decreases, Beta increases, and so on. Instead of creating a single card that asks for “every variable that changes”, I create multiple cards that are way faster to answer. “What happens to resistance when temperature increases?” “What happens to Beta when temperature decreases?” And as you might have noticed, I emphasize the words that might cause a bit of confusion at first glance.
When at least one of these criteria isn’t met, you probably won’t see much of a difference in one card, but it sure makes a whole lot of difference when it scales up.
Let’s say I have a concept card that has an ambiguous question and a lengthy answer. It takes me at least 5-10 seconds to answer it in my head because it didn’t pop up instantly. But when I broke it down into three questions, it only takes a maximum of 2 seconds to answer each card in my head.
By the way, there’s no need to verbally answer, really. The brain is where the activity is at. Verbally answering doesn’t make a difference. It’s the recall that matters and that doesn’t happen in your mouth.
Anyway, by breaking down that specific card, I just cut my answer time by more than half. Just imagine if you did that right from the start–your 120 minutes answer time could, give or take, only take up to 60 minutes. This is the power of creating better questions. It’s something that isn’t taught in the Anki manual, though. (I’m always trying to debunk the idea that “Anki manual is all you need”) In addition, these 3 criteria are more of a supplement to the 20 Rules of Formulating Knowledge, which you should definitely read if you haven’t already.
How do your cards look like?
I’ll show you some of the images of my better cards, but I hope I’ve gotten my points clear in the last section. Anyway, my cards have this 3-part format: Question, Answer, Reference.
The first two should be self-explanatory, but “Reference” isn’t the tedious bibliography shit you’ve learned in school. I’m referring to where you learned the material and in what context. The trouble is, over time, you’re going to totally forget some of the cards especially if you’ve missed a session. (Avoid that by building an Anki habit.) It’s either the card was poorly designed, or the facts inside aren’t memorable. Not totally your fault, though, the process should be mentally taxing. Either way, you don’t want to go back to the actual book or notes before you can re-learn it, do you?
That’s where the context comes in. When creating cards, take a screenshot of the material you got it from. The sweet spot is making a paragraph highlight (perhaps on Evernote) and then capturing that. To do that, I use Greenshot for Windows. Mac already has a screenshot tool built-in, though. Here’s an example.
While the Feynman Technique helps make your cards atomic, context makes your cards future-proof. It could easily be a diagram, or just a highlighted passage from your book or notes like shown above.
Should I take notes before creating Anki cards?
Honestly, if you’re just a beginner, I think you should.
At my skill level of creating questions, I find that I can create questions directly while reading. That means I can perform all of the steps above in a snap–filtering out the irrelevant, simplifying the concept quickly, and creating good questions right away. This takes practice, obviously.
Using the criteria above and keeping some of the 20 Rules of Formulating Knowledge (don’t use lists, build upon the basics, etc.), you’ll eventually be able to create high-quality questions directly into your Anki decks.
Just to be clear, simply creating questions directly doesn’t mean it’s already high-quality. Even in my case, I tend to create poorly designed questions and sometimes forget the context. When that happens and I encounter the card, I edit it right away. I’ve wasted a lot of time doing editing before since I didn’t have any criteria to base off my “question design”. And then I got out of the sunk costs and just started creating cards from scratch. Man, I can’t tell you how dumb I felt creating cards about laws of motion.
What I’d Do Differently
While these processes got refined over time, there are still some things I’d do differently. In particular, I want to point out some of the mistakes I did when I started using this software.
Take brief notes in lectures to simplify the material
I mainly study from textbooks because I treat lectures (both physical and virtual) as “bonuses”—meaning, I don’t take notes at all unless there’s a formula derivation involved and had to take side comments. My justification is “I listen better when not taking notes”. That’s totally bogus. Writing allows you to externalize your thoughts and clarify your thinking. That’s the first thing I’d do differently.
Anki Reviews >> Learning New Material
Back then, I treated Anki like some sort of “reminder” material. It’s just some tool used before taking an exam, I thought. Back then, I didn’t know that recall is more important than exposure.
When I started going up the hierarchy of reading lists–from easy texts to harder ones–I noticed I’ve already forgot the material I’ve previously learned. Let me tell you: these are the same material I believed I understood perfectly. When I first learned them, the images in my head were so vivid it felt like I was hallucinating. But here I was, totally blank, learning a new chapter as if I didn’t read the last one.
And that should be a lesson for you, too. Seriously, it doesn’t matter how much information you consume if you can’t remember any of them. Now, before I learn new material, I empty all my due reviews first. It’s your choice, though, if you want to do that before or after you learn new shit. My advice is to review all due cards first because it takes the most willpower. Prioritize harder stuff–they’re the things you’re most likely to procrastinate on. It’s smart to do them when you still have mental resources, not when you’re out of gas.
Anyway, I’m not reviewing my past cards anymore because they’re not needed in Grad School. What I do now, though, is try to connect new knowledge to old ones through my note-taking process. I think I’m also going to post about that soon.
Some questions before you leave
Just for fun, let me test how much you’ve taken away from this post.
What are the three steps you take when creating cards from your material?
What’s the single reason why putting reference material is so important?
What are the three criteria for creating high-quality cards?
Leave a comment down below if you can recall it. No cheating. 🙂