spaced repetition system algorithm

If I were to build my own spaced repetition system…

sm-2

Priority System

Urgent:

I(1) = 10 min

I(2) = 1 day

High:

I(1) = 1 day

I(2) = 2 day

Medium ( sm-2 default):

I(1) = 1 day

I(2) = 6 day

Low:

I(1) = 1 day

I(2) = 7 day

These still need adjusting

Review Grading

All grading is from 0 to 1. Can be adjusted by the interface however needed, but always normalized to percent.

Other Considerations

What about Pimsleur’s graduated-interval recall?

References

A reddit thread that goes over some ideas (use guidance from SM algorithms and create your own vs try to hook into a potential supermemo API).

Linked from above thread, someone theorized a process here.