Anki SM2 beating FSRS in FSRS Simulator

My workload increased significantly after using FSRS. I blamed this not on FSRS but because of the fact that I was creating many organic flashcards a day given how fast my syllabus was rushing.

The FSRS schedular says that the “suggested retention” is 0.78, if I recall correctly. My current retention is 0.8, which I had downgraded begrudgingly as I was not able to keep up with so many reviews. I have changed and played around with the settings often, such as changing my requested retention, updating my w variable every month or sometimes even in weeks, which might be confusing the simulator. I use custom study sessions heavily, often just before my several exams.

The simulator seems to be showing that Anki’s SM2 beats FSRS on almost every metric, but I do not understand why. It has more “learned cards”, less “time in minutes”, as well as less number of leeches… and the graphs tell a similar story as well, which are self explanatory…

Does this mean that SM2 would be better than FSRS for me? Or is the simulator just confused?

Remembered cards of FSRS are more than the one of Anki in the simulation result.

What is the difference between learned cards and remembered cards?

Besides, while FSRS does have 355 more remembered cards, it also has 1064.7 more time in minutes (time taken to review all the cards FSRS scheduled, I presume?); maybe if I spent that extra time with the SM2 algorithm I would have more remembered cards…

Also, it just occurred to me that perhaps these are the wrong metrics; What does the “time per remembered card” mean? Does it mean the time in minutes it takes me to learn/remember one card for some set amount of time?

For example, if you never review your cards and learn 100 new cards per day, the number of learned cards will be very large, but the number of remembered cards will be very small.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.