Spaced Repetition

Spaced repetition is a practice I only just recently learned about. Don’t worry if you don’t know what it is because I’m about to explain it in depth. This is something that I wish I had known about long ago and I’m still surprised how obscure it is. None of my friends or family had ever heard about it which is a shame because it’s extremely useful and easy (and not nearly as boring as it sounds, I promise!)

But first, a picture:


This image shows your chances of remembering something later after learning it. You can recall the information trivially after immediately learning it, but this memory rapidly decays, to the point where after a day you have a mere 33% retention rate.

This is terrible! We spend all this time learning and forget 80% of it by a month. If only there was a better way.. Well, there are many better ways. For example, in the book “Moonwalking with Einstein”, the author discusses various ways of using mnemonics to encode information efficiently in your brain. For example, immediately upon learning the Memory Palace technique, you will be able to store a list of 50 words in your head without much effort, and remember them for weeks to come.

The problem with these techniques is that they are often optimized for specific forms of information. As just mentioned, the memory palace is very good for remembering lists. The Mnemonic major system is very good at memorizing numbers. Then there are various ad hoc types of mnemonics for individualized pieces of information (e.g. ROY G BIV for remembering the colors of the rainbow and a jingle for memorizing the alphabet). (A future post will likely be dedicated to the various types of mnemonics.)

Another problem with these devices are the high startup costs. Which type of mnemonic is appropriate for this sort of information? Can I make it rhyme in any way? Maybe the first letter of each word spells out something intelligible? Oh, maybe I can associate it with an image! Make sure to add some smells and slimy textures to it to make it extra impactful! Try not to think about how silly this all is or question how weird it is that forcing yourself to remember extra information allows you to remember the thing you care about more easily.

Eventually you may begin to wonder the point of using all this cognitive energy to memorize shopping lists and years of famous battles when you can just look it all up in your phone. Also, exactly which mnemonic technique should I be using to memorize these five paragraphs of this legal case for my Law final..? Surely there’s got to be a better and more general method? And yes, there is! Have another picture:


This image shows how you can avoid the usual rapid memory decay (dashed line) by occasionally reinforcing the information (solid red line). Eventually it gets to the point where the memory retention rate will stick to around 80% for months or even years after first reviewing it several times at spaced intervals.

This is the spaced repetition that the title of this post refers to. Research has found that the ideal time to review something is right before you are about to forget it. Through this technique one can maintain excellent retention even years down the line!

However, isn’t this all obvious? Everyone already knows that spacing out your studying is better than cramming it all in if you want to remember something long term. Also, how exactly are you supposed to remember to review something right as you’re about to forget it? It’s almost paradoxical! Lastly, some things are easier to remember than others. What, are you supposed to keep track of each item’s individual difficulty and manually set alarms to review based on their difficulty? This sounds just as difficult as the aforementioned mnemonics!

Thankfully, all of these issues are completely addressed. There exist software programs (such as Mnemosyne and Anki) that have taken decades of research on spaced repetition (creation and storage of information cards, ideal times to review, denoting individual difficulty, etc.) and internalized it all within a single software package. As a result, these programs can prompt you to review a piece of information right as you were about to forget it, leading to ideal retention!

I’ll give an example of how it generally works. You first launch the program and then create a new flash card. A prompt appears that asks you for the front and back of the card. On the front you type “abrogate” and on the back you type “revoke formally.” Later on when you check the program it will remind you to review the card. The front appears:

Card Front: “Abrogate”


At this point you are supposed to think of the answer because you can’t yet see the back of the flash card. Memory researchers find this form of active recall to be extremely efficient and effective for learning. After you think you have the answer or give up, you click the button to show the back of the card.

Card Back: “revoke formally”


There are then usually around five buttons you can click, respectively labeled 0 through 5. You click 0 if you couldn’t think of it at all and you click 5 if it was the easiest thing in the world and you doubt you’ll ever forget it. Likewise, I’m sure you can imagine what 1, 2, 3, and 4 correspond to. Then, the next review time will be algorithmically calculated based on how many prior reviews you had of this card, the difficulty you denoted, and the average human memory decay rate.

The reason why I’m so excited by this is because it is quite rare for the solution to a problem (remembering information long term) to be the most simple, easy, efficient and most effective solution. Also, this sort of solution had been unavailable for the majority of human history, only made recently made possible by the advent of computers. Lastly, due to the spacing effect, even if one has a card database of tens of thousands of cards, you may very well only need to review 30 or so per day because many will be scheduled for review months down the line.

In doing this, for the majority of people it will only take a couple minutes per day to maintain maximum retention of all the information they care to commit to their database of cards. It is also quite pleasing to occasionally scroll through this personal knowledge library, each item bringing back memories of the context under which you added that item to your knowledge pool. Most programs also allow you to tag your cards so you can instantly access all your accumulated knowledge in whatever subject.

However, what about when I said earlier that this sort of remembering is pointless because you can just look anything up nowadays? At this point this post turns a bit philosophical because this is true to an extent. However, it is not the full story. How many times would a piece of knowledge you had forgotten been an appropriate response to a situation? How would you know? You don’t know what you don’t know.

As a reductio ad absurdum, try imagining your life if you couldn’t form any new memories. Would it not matter much because you could just look everything up anyway? Memories allow us to make connections to disparate topics real-time with whomever or whatever we are interacting with. Spaced repetition gives you access to a larger repertoire and range of relevant responses in reaction to recognition of statements, patterns, and situations.

Okay, but if spaced repetition truly is a panacea, why does hardly anyone use it (other than people who set Jeopardy records)? Why has no one heard of it? It’s because while it may be optimal in long-term, cramming still wins out in the short-term. In this trade-off you get a strong memory now for a weak memory later, thus the old adage about forgetting everything after a test. And since school tests generally consist primarily of new information, this strategy pays off.

Also, since spaced repetition involves proper time management and discipline at a distance, cramming is typically the inevitable outcome for most students. Also, as previously mentioned, forgetting something is tautologically the exact opposite of something you notice. It is difficult to students and teachers to notice the rapid decay of all this knowledge since the system is designed around measuring performance in an environment conducive to cramming.

People I explain this to often ask me what sort of cards they should even make. “How do I know what is important enough to make a card? What sorts of topics should be made cards?” My advice is not to think too hard about it. Making a card takes around 20 seconds usually, and reviewing it around the same amount of time. If I’m surfing the internet and learn something new and interesting and feel like it, I’ll make a card for it. Basically if something interests you and want to increase your chances at bringing it back up in relevant situations in the future, just make a card for it. You can always delete it later if it turns out you don’t actually care about it that much or think that it’s the sort of thing you think you’d know to look up if necessary and you don’t mind doing so.

Well, I think that’s about it for now. Personally I use Mnemosyne because I prefer simple, but Anki is probably better and more feature-complete. I’m interested in seeing what kinds of cards you all come up with.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s