Everyone from #roamcult who has been thinking about spaced repetition lately (, , , , ...) read this from and tell me want you think.
Conversation
Yes. I want cloze deletions, they are one of the most powerful ways of doing spaced repetition. It's how I taught myself to programm fluently in R (instead of snails pace), wrote about it here:
3
1
I find that cloze deletions are the most natural way to integrate these prompts into prose notes, but I’m conflicted about them. I find cloze prompts usually work less well than good Q&A prompts:
2
6
For me, cloze deletions are best when you want "stimulus - response" without reflection/integration. Learning a poem/script or for learning programming. Agree that they are not particularly great for deep understanding of a concept.
2
2
What inspired this question: is there a 80/20 principle in play. Example, can a simple spaced rep algo do the job. Like advance 1 day, if scoring is good, then 2, then 5 days & etc. SM's promise was algo learns based on use over time & self optimizes. Appealing: but does it work?
1
1
This algorithm demonstrably “works” for many purposes: Leitner boxes popularly deployed such a mechanism for decades. Much opportunity left on the table with simple algos; optimization is potentially valuable, but I’m personally much more interested in other research questions.
Agree. A few important UI and storage choices have to be made first in Roam and then we can play with algorithms.
Though my current plan includes just any algo - you have the history of each prompt at hand with a simple roamAlphaAPI call! :)