Anki Users: An alternative to spaced repetiton?

[Foot Fetish]

---

Members don't see this ad.

So, Anki uses "spaced repetition," a learning technique based on the principle that you can more effectively commit material to long term memory if you're exposed to it in increasing intervals of time between subsequent reviews. Basically, Anki has an algorithm that predicts when you will forget a given flash card (based on your own rating of how difficult it was to recall), and it tries to expose you to that card at the approximate time that you would otherwise forget it. In theory, the time it takes for you to forget the card will get longer and longer with each exposure until the card is eventually stored indefinitely in your long term memory, as illustrated in this diagram:

The traditional wisdom is that, when it comes to long term learning, spaced repetition is superior to "massed presentation," a.k.a. "cramming," wherein you expose yourself to the material repeatedly in a short period of time until 100% of it is stored in your short term memory. Cramming is great for test preparation as it allows you juggle the massive amount of info in your short term memory just long enough to dump it on the test. The problem of course, is that you then quickly forget the crammed material after the test.

...But what if you made a comprehensive pass through the crammed material every day after the cram session? Instead of attempting to gradually commit the deck to long term memory over a really long period of time by waiting multiple days to almost forget bits and pieces of it before Anki prompts you to recall it again, why not just cram the whole thing into your short term memory and then force it to stay there by repeating the whole deck every day thereafter until it is converted to long term memory? In terms of the above figure , I think this would essentially look like a straight line maintained at the 100% chance of remembering mark.

The obvious drawback of this technique is that it takes significantly more effort as you build up multiple decks, which you may then be loath to review in their entirety on a daily basis. However, it seems like the more risk-averse approach compared to relying on an imperfect software algorithm. You leave nothing to chance when you brute force it. Plus, you can make it more manageable by being more selective in what information you choose to turn into a card, focusing only on lists and minutiae while learning the more conceptual stuff by simply mulling over didactic texts to truly understanding the logic, which you can then deduce once it finally clicks.

Thoughts?

TLDR: Instead of slowly committing a deck of flashcards to long term memory via spaced repetition, why not just cram the whole deck at once and then maintain the whole thing in your memory by reviewing it daily after the cram session?

 

[ForHumors]

That would be superior for remembering those particular facts. But IMO it's unrealistic because you'd end up spending a lot of time every day reviewing things that, predicted algorithmically, you already know really well. Med school asks you to remember too many things for you to be that inefficient.

 

[eteshoe]

Who the hell would have time to keep up w/ that kind of torture daily? OP if you can do it - more power to you

 

[Donald Juan]

Because eventually your deck is thousands of cards that would take all day to go through...the majority of cards which would seem so basic to you that it would be a complete waste of time.

Even the way it is structured as is (anki or firecracker), spaced repetition often becomes unrealistic for many people as daily reviews become so enormous...and that's with increasing intervals, not your review everything everyday approach.

 

[mehc012]

Even more so, you're actually better off waiting those longer pauses. You will remember a set of facts better if you review it on Day 1 and Day 5 and Day 15 (reviewing the ones you forget) than if you review it on Day 1 and 2 and 3.

 

[Foot Fetish]

Even more so, you're actually better off waiting those longer pauses. You will remember a set of facts better if you review it on Day 1 and Day 5 and Day 15 (reviewing the ones you forget) than if you review it on Day 1 and 2 and 3.

I'm not very familiar with the algorithm...I guess my concern is that it will not force you to review enough..So does Anki make you keep repeating a card until you're able to "get it" at least once that day? In other words, no cards will go to the Day 5 pile unless I get them at least once on Day 1 right? It just doesn't feel like enough to only match the deck once per day or less...

 

[mehc012]

I'm not very familiar with the algorithm...I guess my concern is that it will not force you to review enough..So does Anki make you keep repeating a card until you're able to "get it" at least once that day? In other words, no cards will go to the Day 5 pile unless I get them at least once on Day 1 right? It just doesn't feel like enough to only match the deck once per day or less...

*Note to Anki power users: I know that this simplified example isn't exactly right, in that the Graduating Interval is the same even if you got it wrong a few times, but this was the easiest way to introduce the general idea of Ease as in 'the more you get it wrong, the more you see it' without going into the actual variables.*

Say you make 100 flashcards. You're new to Anki and have no existing load, so you go ahead and learn them all that day.

You will view each of those existing flashcards at least 1x that first day. Let's say that you get 80 of them right, 15 of them wrong, and 5 of them are just the easiest thing you could imagine.
You will, depending on your settings, be shown those 15 cards again IMMEDIATELY, and each one again and again until you get it right.
Later that day (how long depends on your settings) you will see the 80 you got right the first time, and the 15 you got right eventually. You will not see the easy ones.
Day 1:
First pass: 100 cards (15wrong/80right/5easy)
Second pass (or more): 15 cards (the 15w from above)
Third pass: 95 cards (the 80r from above and the 15w)

OK, for Day 2, you've *just* seen all of these cards. You don't need to see them all again. But you had a hard time with some of them, so maybe (depending on your settings) Anki shows those to you today. You get 5 of them wrong again. Anki shows them to you again immediately, and then checks those ones later in the day (just like it did yesterday).
Day 2:
First pass: 15 cards (5wrong/10right)
Second pass (immediate): 5 cards (the 5w from above)
Third pass: 5 cards

On Day 3, you probably just see the cards you *just* got wrong. You get them all right because now you've seen them like 6-7 times in 2d.
Day 3: 5 cards

On Day 4, it has been 3d since you first reviewed your cards. 85 of them you haven't seen in 3d, so you should probably check to see if you remember them. 10 of them you knew just 2d ago, so you can wait until tomorrow. And 5 of them you saw just yesterday, but you got them wrong twice already so you should review them anyways just to make sure you know them. We're going to stop following them at this point because they're going to splinter further as you get different combinations right/wrong, but hopefully this gives you the idea.
Day 4: 90 cards (80 that you got right 2x on Day 1, 5 that you thought were easy on Day 1, 5 that you've been struggling with)
Day 5: 10 cards that you got wrong once on Day 1, but right on the 2nd/3rd pass of Day 1 and then right on Day 2. Plus whatever you get wrong from Day 4.

This is oversimplified, but hopefully you start to get the idea that Anki doesn't just schedule things randomly, it schedules them based on your demonstration of how well you know the information on that card. Why review all 100 cards everyday when you can focus on the ones that you don't know well?

If I simply went through all 100 everyday, I'm not reinforcing the ones I got wrong. I only have time to see it once on a given day, even if I don't know it yet. In the 5d example above, I did only 340 card reviews, yet I saw my problematic cards 6-8x each minimum, possibly more. If I had simply done one pass of all cards each day, I'd have done 500 card reviews and reviewed my problem cards only 5x each. Anki's system is way more efficient. As time goes on, this difference in efficiency will only increase.

 

[mehc012]

Also, for the record, you can always choose to do a Cram deck full-pass session on the days leading up to the exam. So you'd use the SRS system, have a normal load everyday for a month or so, and then when the end of Block comes up you can choose to have a few days where you go through the entire deck without messing up the scheduling.

As an added benefit, once you've gone through the entire deck in your dedicated exam study time, you can then use the "Review missed cards" feature to pull out ALL of the cards that you missed even once during that comprehensive cram session, and just cram those repeatedly until you know them cold. That would be more efficient than doing a full Cram every day in the last few days before an exam...which is itself more efficient than doing a full cram every day the entire Block (god I would die.)