[Feature Request] Adaption to exam learning about fsrs4anki

[L-M-Sherlock]

If we can set a clear goal for the short-term learning, it is possible to design a new algorithm for optimization. I need the goal to be formulated in math.

[galantra]

Just an idea, which can be done with the current FSRS4Anki tools:

The parameter that you might want to adjust is requestRetention. The FSRS4Anki Optimizer suggests the optimal target retention for you; this is the most time-saving target. However, on the date you write the exam, you might want a higher probability of recall.

What you could do is use the suggested requestRetention (e.g., 0.85) up until a short time (like, a week) before the exam. At this point you change the requestRetention value to something higher, e.g. 0.9. Then you reschedule all cards with the FSRS4Anki Helper add-on.

I imagine (haven't simulated it) that this could result in a lot more reviews shortly before the exams – so you'll have less time to learn new cards. However, these additional reviews will be exactly those cards where your probability of recall is the lowest (i.e. 0.85, 0.86, 0.87, 0.88, and 0.89).

[L-M-Sherlock]

What you could do is use the suggested requestRetention (e.g., 0.85) up until a short time (like, a week) before the exam. At this point you change the requestRetention value to something higher, e.g. 0.9. Then you reschedule all cards with the FSRS4Anki Helper add-on.

Thanks for your solution! It inspired me to develop the Postpone & Advance features for the Helper add-on: https://ankiweb.net/shared/info/759844606

[cryptocoinserver]

After giving it some thoughts, it shouldn't differ for exam learning? The memory functions the same way when learning for exams, so your formulas don't need modification. Lowering requestRetention and (maybe) lowering maximumInterval should be enough. Maybe you want to add that to the FAQ or simulate it.

[L-M-Sherlock]

Need more constructive advice.