I have a deck with about 1800 cards. I completed it a few days ago and decided to try a new algorithm for future notes - FSRS. I have been reviewing about 70 cards per day with the original algorithm. However, after using FSRS, I have a few concerns:

Why does the number of FSRS review sessions fluctuate like this? 0 cards on the first day, then 90, then reduced to 2,… Is this normal when using FSRS? I’m doubting its effectiveness.

Can FSRS be trained with data from the old algorithm? Or do I need to review about 2000 times with FSRS before training it with new data?

In the worst case scenario, if FSRS isn’t suitable and I want to go back, do I just need to remove the custom scheduler? I tried this and it doesn’t seem to give the desired results

P/s: I’m using the settings as instructed on GitHub, and I haven’t made any changes.

For 3.
Q6: Once I started using this FSRS algorithm on my existing deck, if I ever have to go back to using Anki’s built-in algorithm for the same deck, would that still be possible?

A6: The ease factor is modified by the v3 scheduler as usual, and FSRS4Anki doesn’t intervene in it. FSRS4Anki only modifies the interval of cards. So you can go back without any problem.

The Anki simulator still use the Anki built-in algorithm to simulate your future review. So the graph is inaccurate to predict the review scheduled by FSRS.

Thank you. I’m also confused about the accuracy of that chart. I apologize for not being able to describe the issue specifically, and I don’t understand the mathematical explanations in the instructions either. But is there a concise explanation you think I can understand for the significant decrease in the number of review cards?