5 ms·

Long-term user of Anki ~ 13-14 yrs, with mostly language learning focus. It’s good; but out of the box, I find that most users barely tap into its utility. It’s

Long-term user of Anki ~ 13-14 yrs, with mostly language learning focus. It’s good; but out of the box, I find that most users barely tap into its utility. It’s templating engine, ability to shape card content and formatting via JS and CSS put it in a league above most other SRS applications. (With the caveat that it’s easy to waste a lot of time on inconsequential card-tweaking…) A lot of people complain that it uses an inferior SRS algorithm. Other algorithms can be patched in; but I’ve never bothered because it seems like a hyper-optimization without known real-world outcomes. (i.e. Will I speak better {L2} if I use alternate SRS algorithm?)

subtra3t 3y ago
I think the issue of inferior SRS algorithm is at least partly alleviated by the existence of FSRS.
- OJFord 3y ago
  ... https://github.com/open-spaced-repetition/fsrs4anki/wiki/The-Algorithm https://github.com/open-spaced-repetition/fsrs4anki/wiki/The... ... I'm not sure I believe we understand our own learning/memory anything like enough for this not to be total pseudoscience? Reminds me of A Beautiful Mind.
  - g-w1 3y ago
    I'm not really sure what you mean? It's been empirically validated to predict how you remember better than alternatives (or equal to SM-17) (essentially when it predicts that your % recall is below a certain threshold, it just shows you the card)
  - kouteiheika 3y ago
    It doesn't matter whether this actually models how memory works in reality, as long as it has actual predictive capabilities which correlate to how it works in practice. And this can be easily verified with data and simulations. The algorithm predicts how likely it is that you'll remember a given piece of information (e.g. in three days you'll remember this card with 80% probability), so if you can get a big dataset of reviews and run it through the algorithm you can easily check how accurate it is, e.g by calculating a brier score, or by comparing predicted vs actual recall curves. Source: I've developed an even better algorithm than FSRS (I've directly compared them in the past, although that was quite a while ago so it might have been for one of the older versions of FSRS, so I'm not sure how the newest one compares), and now I'm working on an even better one.
    - jarrett-ye 3y ago
      It's welcome to compare your algorithm with FSRS. Here is the benchmark of FSRS: https://github.com/open-spaced-repetition/fsrs-benchmark https://github.com/open-spaced-repetition/fsrs-benchmark
      - kouteiheika 3y ago
        Yes, I'll probably do that once I'm finished with the new version of my algorithm.

subtra3t 3y ago

I think the issue of inferior SRS algorithm is at least partly alleviated by the existence of FSRS.

OJFord 3y ago
... https://github.com/open-spaced-repetition/fsrs4anki/wiki/The-Algorithm https://github.com/open-spaced-repetition/fsrs4anki/wiki/The... ... I'm not sure I believe we understand our own learning/memory anything like enough for this not to be total pseudoscience? Reminds me of A Beautiful Mind.
- g-w1 3y ago
  I'm not really sure what you mean? It's been empirically validated to predict how you remember better than alternatives (or equal to SM-17) (essentially when it predicts that your % recall is below a certain threshold, it just shows you the card)
- kouteiheika 3y ago
  It doesn't matter whether this actually models how memory works in reality, as long as it has actual predictive capabilities which correlate to how it works in practice. And this can be easily verified with data and simulations. The algorithm predicts how likely it is that you'll remember a given piece of information (e.g. in three days you'll remember this card with 80% probability), so if you can get a big dataset of reviews and run it through the algorithm you can easily check how accurate it is, e.g by calculating a brier score, or by comparing predicted vs actual recall curves. Source: I've developed an even better algorithm than FSRS (I've directly compared them in the past, although that was quite a while ago so it might have been for one of the older versions of FSRS, so I'm not sure how the newest one compares), and now I'm working on an even better one.
  - jarrett-ye 3y ago
    It's welcome to compare your algorithm with FSRS. Here is the benchmark of FSRS: https://github.com/open-spaced-repetition/fsrs-benchmark https://github.com/open-spaced-repetition/fsrs-benchmark
    - kouteiheika 3y ago
      Yes, I'll probably do that once I'm finished with the new version of my algorithm.