When I try to optimise my FSRS parameters it seems to be missing alot of my reviews and I'm unsure as to whether I should keep this settings

(put here) AnkiDroid Version = 2.20.0 (d41d16a45c00b0bdbd3b13ac988143db5ef8ae90)

Hi, I couldn’t say everything I needed to in the title, but I’ve recently just tampered with my settings in the advanced

Maximum interval

36500

Historical retention

0.83

Ignore cards reviewed before

16/05/2025

Because around that time I completely changed my way of doing anki cards, started using fsrs and everything.

But when I optimise fsrs it says” ⁨1,147⁩ reviews” beneath it. And I’ve done my than that today alone. Ever since 16/05/2025 I’ve been doing like roughly 2,000 a day.

With my desired retention at 0.95, but I want to lower it to 0.90. Just in the last three months for my main deck it says I’ve done “⁨245,868⁩ reviews”

And if I just use the default settings for advanced

Maximum interval

36500

Historical retention

0.90

Ignore cards reviewed before

01/01/1970

And click optimise then it shows” ⁨1,163,460⁩ reviews”

Which roughly alligns to how many reviews Ive done for that deck 1,126,686⁩.

So what I want to know is why is it only saying only ⁨1,147⁩ reviews.

I have really bad memory if you can tell. Thanks for reading this far and I’d appreciate any help given.

That setting has FSRS ignore the entire review history – past and future – for any cards introduced before that date.

See: Deck Options - Anki Manual

Thanks for the response man, sorry to ask the obvious, but just wanted to make it clear that you mean any cards that I’ve reviewed before 16/05/2025 get ignored even if I’ve reviewed them after that date? If so that does make sense. And assuming that’s the case, should I optimise ignoring any cards before 16/05/2025 or add more cards regularly and then do it. Because 1,000 cards sounds like too little data to go on. I’ve already gone down from 0.95 desired retention to 0.90 and optimised that with default advanced settings. And I can just stick with that for a while and see if that helps. Because I’ve not been doing any new cards at all and am still stuck at 2,000 + cards a day.

Because I was never seeing any reduce in amount in my cards for the past 6 months.

Thanks for the help again.

For purposes of optimizing your parameters – correct.

That’s not 1000 cards FSRS is considering – that’s 1000 reps it’s considering, likely on much fewer than 1000 cards. It’s how many cards you’ve introduced after that date. You’re right, that’s not very much review history.

Ignore-before is a pretty extreme solution for handling some specific situations. What makes you think it’s necessary in your case? Simply enabling FSRS isn’t enough, or else we’d all have to use it! Were you grading your cards inaccurately? Were you mis-using Hard for answers you actually got wrong?

Your DR isn’t part of optimization – but 95% DR is too high for most users, especially if that’s a lot higher than your actual retention results.

Hey there, yeah I had before that been not fully trying to recall meaning (I do japanese vocab cards and try to recall the meaning and prounounication) of the cards and just passed if the meaning was roughly the same, of course this wasn’t for all cards but for a big chunk of them. I have cards that I’ve reviewed over 100 times and still don’t remember. Also when I was starting out I kinda abused hard button for a short while, but not long in terms of total amount of time on anki.

These 6 months have been quite long, double the time I spended usually” 2,000+ 2+ hours everyday “ (and because of that I never could do new cards otherwise I’d be spending alot more time in anki that’s why there’s only 1000+ cards introduced since 6 months ago), so hopefully It’s more than enough to essentially ignore those reviews. (Yeah probably not I’d have to maybe another half a year like this before those reviews become a high percentage of my total amount of reviews done.)

Thanks again.

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