Anki and the Easy button
Oh boy! This is going to be controversial. Well, as controversial as any topic about SRS based flash card application can get.
Anyway. Use the Easy button. Religiously!
Probably the most common sentiment I read is that people using Anki should almost never, if ever, hit the Easy button, and instead should lean towards the Hard and Ok buttons.
I disagree with this. With caveats. 3 of them.
The first is when you're first using Anki. Memorization is a skill. It takes time to get the brain up to speed. In those early days it might make a bit more sense to punish yourself. Especially if you have a need to get this stuff committed to memory.
The second is if you have a VERY small set of cards. Say, less than 500 total. Or a lot of free time. With smaller decks or a lot of time to dedicate to memorizing things, you can afford to be inefficient.
And lastly, if you need to maximize retention over volume. For instance, if you have a test and the only cards in your Anki are for that test. Obviously, in that case, you want to prioritize accuracy over efficiency.
Efficiency was mentioned in both of the latter cases. The root of my assertion that the Easy button is your friend revolves around two concepts. Memory fatigue and time. And efficiency for me is determined by minimizing time required to maximizing the number of flash cards committed to memory.
"Memory Fatigue" which is just a term I'm choosing to describe the observation that the more things you try to remember in a given period of time, the more likely you are to fail to advance any particular item further in your memory.
Remembering one new word a day is easy. Thus, remembering 100 new words in a 100 days is likely pretty easy as well. But, remembering 100 new words in one day, studied in a 1 hour session? That is pretty hard. And, in my experience it is non-linear. Memory fatigue grows faster than the number of things you try to commit to memory.
Those pessimistic choices in Anki then, mean that you'll see each new card more often and also that your daily reviews will grow more quickly. Which in turn means it increases the odds that memory fatigue will become a significant factor in how well you absorb the material if you're not also inflating your time investment accordingly.
And if you have large decks and are adding in new items every day, that just raises the ceiling on how many daily reviews you might need to contend with.
Sure, you CAN limit daily reviews. But all that is doing is artificially doing the same thing as using more generous intervals. You're kicking cards further down the road than they were scheduled originally. And worse, they're being held back by other cards which probably showed up sooner than they truly needed too. It just makes sense to adjust your intervals and button choices over limiting reviews.
The next metric to look at is accuracy. A lot of people strive for 100% accuracy. But, again, unless you're studying for a test with a limited number of flash cards, 100% just means that it is almost certain that you could have handled more.
While memory fatigue is bad, having it affect your studies is paradoxically a good sign. It means that you can be sure you are giving yourself at least as many reviews as you can handle. And so, generally speaking, you want a high, but sub 100% accuracy to maximize your performance. My personal suggested target is 85-95%.
There is another benefit as well to the "Easy" method. Frustration. The pessimistic approach favours brute forcing things into memory over establishing better habits and techniques. By maximizing the frequency at which you see the same cards, you reduce your throughput in exchange for the easiest path forward. It is certainly more pleasant. But, the frustration from challenging yourself can actually encourage you to adopt better practices.
If I find some cards stump me more often than others, I stop and spend more time looking at and thinking about them. I may even try to come up with a story or mnemonic device of some sort.
It is this combination of both pushing your natural memorization abilities and forcing yourself to spend your time better which has resulted in me bringing down my daily review count from 400 to under 200 even though I increased my new card count by 5 and I'm not done my decks yet. My accuracy rating is still regularly 90-95%. I have almost 17k mature cards and I've been using this strategy for several months now. If it weren't working, I'd know.
My accuracy is more or less unchanged from the days when I was being pessimistic with myself. But, as I said, this shouldn't come as a surprise. While I am increasing the interval between the times when I see a card again, which should decrease my accuracy. It is then offset to a large degree, if not entirely, by virtue of the fact that I experience less memory fatigue.
Efficiency.
Are there conditions under which this technique will fail? Absolutely.
If your accuracy is too low, it will likely make things worse. I would say that under 75% would be dangerously low.
But, then I would still choose other ways to fix the problem. Such as adjusting the interval percentages or dropping the number of new daily cards or even just spending more time.
And I suggest those because Anki offers too few buckets to dump cards into. This makes the app more user friendly. But, reviews can quickly get too "lumpy".
Getting the cards scheduled to come back at the right time to keep them in your memory requires nuance. And, if you tell yourself that you can't hit the Easy button, you've robbed yourself of one of those buckets, made things less nuanced and the results lumpier. Worse, you'll be adding to that memory fatigue by trying to recall some cards well before you needed to see them again.
I use all of the buttons. There is no button which is taboo for me to choose. I define a set of rules for myself based on the type of card which I think gives me the most effective way to group them together.
For instance, with audio only cards. If I can get the sentence in one pass, it is Easy. If I need to listen again, or make non-critical mistakes, it is Ok. If I need to listen more than that, or make a more critical mistake (which doesn't affect the target word or rough sentence meaning) then I will go Hard. Otherwise, Again. And, when in doubt, I round up.
If I'm going to get lumpy results anyway because I only have 3 buckets to dump things in, this seems like a reasonable way to group them. If I can translate the sentence more or less perfectly in a single pass, I don't want it coming back to me at the same frequency as ones I struggled enough with to play a second time. And likewise, even if I had to listen twice or made a minor mistake, I don't want those ones popping up at the same frequency as cards where I nearly got it totally wrong or had to listen repeatedly.
As I hope you can see... three buckets for getting past a card is the barest of minimums. I could easily split the Ok bucket in two, and the Hard bucket into at least 3 more groups.
2 buckets would mean eliminating one of those three and merging it's contents into the remaining ones.
Do yourself a favour. Don't try to solve your problems by reducing your options. It may work by simplifying things for a time. But, it will come back to haunt you.
Anyway. Use the Easy button. Religiously!
Probably the most common sentiment I read is that people using Anki should almost never, if ever, hit the Easy button, and instead should lean towards the Hard and Ok buttons.
I disagree with this. With caveats. 3 of them.
The first is when you're first using Anki. Memorization is a skill. It takes time to get the brain up to speed. In those early days it might make a bit more sense to punish yourself. Especially if you have a need to get this stuff committed to memory.
The second is if you have a VERY small set of cards. Say, less than 500 total. Or a lot of free time. With smaller decks or a lot of time to dedicate to memorizing things, you can afford to be inefficient.
And lastly, if you need to maximize retention over volume. For instance, if you have a test and the only cards in your Anki are for that test. Obviously, in that case, you want to prioritize accuracy over efficiency.
Efficiency was mentioned in both of the latter cases. The root of my assertion that the Easy button is your friend revolves around two concepts. Memory fatigue and time. And efficiency for me is determined by minimizing time required to maximizing the number of flash cards committed to memory.
"Memory Fatigue" which is just a term I'm choosing to describe the observation that the more things you try to remember in a given period of time, the more likely you are to fail to advance any particular item further in your memory.
Remembering one new word a day is easy. Thus, remembering 100 new words in a 100 days is likely pretty easy as well. But, remembering 100 new words in one day, studied in a 1 hour session? That is pretty hard. And, in my experience it is non-linear. Memory fatigue grows faster than the number of things you try to commit to memory.
Those pessimistic choices in Anki then, mean that you'll see each new card more often and also that your daily reviews will grow more quickly. Which in turn means it increases the odds that memory fatigue will become a significant factor in how well you absorb the material if you're not also inflating your time investment accordingly.
And if you have large decks and are adding in new items every day, that just raises the ceiling on how many daily reviews you might need to contend with.
Sure, you CAN limit daily reviews. But all that is doing is artificially doing the same thing as using more generous intervals. You're kicking cards further down the road than they were scheduled originally. And worse, they're being held back by other cards which probably showed up sooner than they truly needed too. It just makes sense to adjust your intervals and button choices over limiting reviews.
The next metric to look at is accuracy. A lot of people strive for 100% accuracy. But, again, unless you're studying for a test with a limited number of flash cards, 100% just means that it is almost certain that you could have handled more.
While memory fatigue is bad, having it affect your studies is paradoxically a good sign. It means that you can be sure you are giving yourself at least as many reviews as you can handle. And so, generally speaking, you want a high, but sub 100% accuracy to maximize your performance. My personal suggested target is 85-95%.
There is another benefit as well to the "Easy" method. Frustration. The pessimistic approach favours brute forcing things into memory over establishing better habits and techniques. By maximizing the frequency at which you see the same cards, you reduce your throughput in exchange for the easiest path forward. It is certainly more pleasant. But, the frustration from challenging yourself can actually encourage you to adopt better practices.
If I find some cards stump me more often than others, I stop and spend more time looking at and thinking about them. I may even try to come up with a story or mnemonic device of some sort.
It is this combination of both pushing your natural memorization abilities and forcing yourself to spend your time better which has resulted in me bringing down my daily review count from 400 to under 200 even though I increased my new card count by 5 and I'm not done my decks yet. My accuracy rating is still regularly 90-95%. I have almost 17k mature cards and I've been using this strategy for several months now. If it weren't working, I'd know.
My accuracy is more or less unchanged from the days when I was being pessimistic with myself. But, as I said, this shouldn't come as a surprise. While I am increasing the interval between the times when I see a card again, which should decrease my accuracy. It is then offset to a large degree, if not entirely, by virtue of the fact that I experience less memory fatigue.
Efficiency.
Are there conditions under which this technique will fail? Absolutely.
If your accuracy is too low, it will likely make things worse. I would say that under 75% would be dangerously low.
But, then I would still choose other ways to fix the problem. Such as adjusting the interval percentages or dropping the number of new daily cards or even just spending more time.
And I suggest those because Anki offers too few buckets to dump cards into. This makes the app more user friendly. But, reviews can quickly get too "lumpy".
Getting the cards scheduled to come back at the right time to keep them in your memory requires nuance. And, if you tell yourself that you can't hit the Easy button, you've robbed yourself of one of those buckets, made things less nuanced and the results lumpier. Worse, you'll be adding to that memory fatigue by trying to recall some cards well before you needed to see them again.
I use all of the buttons. There is no button which is taboo for me to choose. I define a set of rules for myself based on the type of card which I think gives me the most effective way to group them together.
For instance, with audio only cards. If I can get the sentence in one pass, it is Easy. If I need to listen again, or make non-critical mistakes, it is Ok. If I need to listen more than that, or make a more critical mistake (which doesn't affect the target word or rough sentence meaning) then I will go Hard. Otherwise, Again. And, when in doubt, I round up.
If I'm going to get lumpy results anyway because I only have 3 buckets to dump things in, this seems like a reasonable way to group them. If I can translate the sentence more or less perfectly in a single pass, I don't want it coming back to me at the same frequency as ones I struggled enough with to play a second time. And likewise, even if I had to listen twice or made a minor mistake, I don't want those ones popping up at the same frequency as cards where I nearly got it totally wrong or had to listen repeatedly.
As I hope you can see... three buckets for getting past a card is the barest of minimums. I could easily split the Ok bucket in two, and the Hard bucket into at least 3 more groups.
2 buckets would mean eliminating one of those three and merging it's contents into the remaining ones.
Do yourself a favour. Don't try to solve your problems by reducing your options. It may work by simplifying things for a time. But, it will come back to haunt you.
Comments
Post a Comment