3 ms·

Piotr Wozniak, inventor and author of spaced repetition algorithms (invented the SM-2 algorithm used by Anki around 1986), states quite clearly in his research

Piotr Wozniak, inventor and author of spaced repetition algorithms (invented the SM-2 algorithm used by Anki around 1986), states quite clearly in his research that these algorithms are not designed for children under about 7 (of course super high IQ children's brains are different). He reviewed a vast amount of performance data to conclude that, and hypothesized that children's brain before that age is not designed for remembering vast amounts of data, but instead learn through exploration and other natural, non-academic means.

iwsk 3y ago
Regardless of that hypothesis, this experiment was a huge success.
surfmike 3y ago
Also there's a lot of advantages of play-based open-ended learning especially at that age. That's what we've focused on instead of trying to put our kid through a curriculum or stuffing facts in their brain. (That said... he does seem to have the ability to memorize a lot of random dinosaur, insect, and LEGO facts anyway)
keiferski 3y ago
There is a false assumption here though, which is that spaced repetition/flashcards = learning vast amounts of data. It may simply be the case that learning to read via a small number of efficient flashcards is more effective than other learning methods.
rsanek 3y ago
In the latest version, Anki has started experimenting with a new algorithm to replace SM-2, called FSRS. More info in the manual https://docs.ankiweb.net/deck-options.html#fsrs https://docs.ankiweb.net/deck-options.html#fsrs
- hereme888 3y ago
  Yes! I've read the research for FSRS. Been meaning to try it out, but put it off because the github instructions seemed complicated. But it's the logical evolution SR software. IMO SuperMemo has no future unless it goes the open-source route. It's already partially broken due to its dependence on IE. Sad because I have so much data in SM.
  - jarrett-ye 3y ago
    [dead]

iwsk 3y ago

Regardless of that hypothesis, this experiment was a huge success.

surfmike 3y ago

Also there's a lot of advantages of play-based open-ended learning especially at that age. That's what we've focused on instead of trying to put our kid through a curriculum or stuffing facts in their brain. (That said... he does seem to have the ability to memorize a lot of random dinosaur, insect, and LEGO facts anyway)

keiferski 3y ago

There is a false assumption here though, which is that spaced repetition/flashcards = learning vast amounts of data. It may simply be the case that learning to read via a small number of efficient flashcards is more effective than other learning methods.

rsanek 3y ago

In the latest version, Anki has started experimenting with a new algorithm to replace SM-2, called FSRS. More info in the manual https://docs.ankiweb.net/deck-options.html#fsrs https://docs.ankiweb.net/deck-options.html#fsrs

hereme888 3y ago
Yes! I've read the research for FSRS. Been meaning to try it out, but put it off because the github instructions seemed complicated. But it's the logical evolution SR software. IMO SuperMemo has no future unless it goes the open-source route. It's already partially broken due to its dependence on IE. Sad because I have so much data in SM.
- jarrett-ye 3y ago
  [dead]