Is there a way you could add the selection of different algorithms? This old obsidian-recall plugin allowed this.
Another person in this topic #51 asked if a newer algorithm is possible to implement, while I prefer the algorythm that Mnemosyne utilizes. It may be old, but it's way, way better for learning languages in comparison to Anki's, in my experience 😅
The other way to achieve that would be to allow a user to adjust that manually...? There are not enough tweaks for now.
For example, to achieve Mnemosyne-like algorithm, it lacks:
-
Randomization (“fuzz”) — Mnemosyne adds a small ±5% variation to intervals so reviews don’t all land on the same day. Anki’s SM‑2 (and Aosr’s default) is deterministic unless you code in that randomness.
-
Early/Late review scaling — Mnemosyne adjusts the next interval proportionally if you review early or late. Anki’s SM‑2 ignores this and just applies the scheduled interval regardless of when you actually review.
But, still, obsidian-recall already has some ready-made solutions and algorithm ts files.
Is there a way you could add the selection of different algorithms? This old obsidian-recall plugin allowed this.
Another person in this topic #51 asked if a newer algorithm is possible to implement, while I prefer the algorythm that Mnemosyne utilizes. It may be old, but it's way, way better for learning languages in comparison to Anki's, in my experience 😅
The other way to achieve that would be to allow a user to adjust that manually...? There are not enough tweaks for now.
For example, to achieve Mnemosyne-like algorithm, it lacks:
Randomization (“fuzz”) — Mnemosyne adds a small ±5% variation to intervals so reviews don’t all land on the same day. Anki’s SM‑2 (and Aosr’s default) is deterministic unless you code in that randomness.
Early/Late review scaling — Mnemosyne adjusts the next interval proportionally if you review early or late. Anki’s SM‑2 ignores this and just applies the scheduled interval regardless of when you actually review.
But, still, obsidian-recall already has some ready-made solutions and algorithm ts files.