I agree with what JamOnDrums wrote, that "repeat period should depend on the user and their previous scores on a particular card. but I definitely think that the durations should not be completely fixed but should adapt to the learner. " Harder cards should be displayed more often than easy cards, as this is how most flashcard learning programs work. Intervals cannot be fixed, as the learning effectiveness will be low. Instead of fixed, uniform intervals for all cards, a single coefficient should be introduced for each card individually, calculating the time of the next repetition. With an incorrect answer, the coefficient decreases (the card is shown more often), with a correct answer, it increases.

It works like this: the previous learning interval multiplied by the coefficient (K) equals the new interval.

Example with correct answer (for a selected card):

for K = 2, n=0.3

If interval >1 day

K=K+n

K=2+03=2.3

1 * 2. 3= 2.3 (days)

correct answer

K=2.3+0.3=2.6

correct answer

K=2.6+0.3=2.9

5.98* 2.9 = 17.34 (days)

correct answer

K=2.9+0.3=3.2

17.34* 3.2 = 55.48 (days)

With an incorrect answer:

We start learning from the beginning.

Intervals:

90 seconds

30 minutes

If intervals >=1 day

incorrect answer

K=3.2-0.3=2.9

1*2.9=2.9 days

incorrect answer

K=2.9-0.3=2.6

1*2.6=2.6 days

incorrect answer

K=2.6-0.3=2.3

1*2.3=2.3 days

This is only example.

Hi Tatiana, thank you for your comment! You can adjust how often the words appear by swiping left instead of right if you want to see the word again sooner.🙌

Other than that, a similar idea is already on Canny, so I’m merging your post with the existing one, 'User can choose the spaced repetition algorithm,' to consolidate votes and keep things organized.✌️Michaela from DuoCards

This sounds like a power-user feature. And Anki is already free

I feel like the repeat period should depend on the user and their previous scores on a particular card. but I definitely think that the durations should not be completely fixed but should adapt to the learner. The user choosing their own intervals might not be the way to go as it could depend on the card you are tying to learn

I really agree because although space repitition is used, I need to be able to choose more regular intervals for those words that Duo thinks I've remembebered long-term and won't show for a very long time, when I'd benefit from another recall perhaps in a month's time. This means my hard work for learning some of the words doesn't effectively carry through for the long-term which is frustrating.

From my experience - I would reduce long 'few-month' period. I am not sure what actual value is - 2 or 3 month but I have forgotten number of cards during this period and suppose shortening this repeat-period would be good. But it is my feeling - it may be different for different people

Hi Mochaela

Reapeting a word and learning it again are two different things . The algoritm should be more optimized for the path of learning or a user should have option to choose intervals. I have read several times on reddit that your algoritm is not perfect and other apps have better one but still wanted to do you a chance. But they are right. With such intervals nobody will learn anything for a long time

I got it, but make cards return after 1 or some years is definitely more convenient btw