open plank
Hi, I’m new to this. I last reviewed this card on September 12, and today is October 1.
I only reviewed it once (on September 12), and now when I review it today and hit “Easy,” it says the next review is in 2 months.
Is this how the algorithm is supposed to work, or did I mess up my settings?
i kinda tried FSRS but i dont understand it so i reverted it back. not sure if that has an effect
for reference these are my settings attached
spark oasis
Show the card information for the card you're talking about.
Find it in the browser and press ctrl + shift + i to pull the information up.
I suspect that you studied this card well after its initial due date had passed. When you do that, Anki (rightly) assumes that you know the card pretty well, and increases the subsequent interval.
That being said:
I'm not sure why you would hit Easy and not expect it to go far into the future. The point of SRS is to not waste time on things that you already know.
In general, if you're rating the answers correctly, you can trust what Anki wants to do for you.
Don't worry about the displayed intervals, and I generally recommend that people turn off the ability to see them.
FSRS is even more trustworthy than the default algorithm. You should switch it on in your settings. The only additional complication is that you have to make sure you're using the buttons correctly, and you need to press the Optimize All Presets buttons once a month in your deck options.
open plank
here is the information.
would like to talk about more about FSRS regarding my reviews after this is resolved
spark oasis
Thank you. It's exactly what I thought.
The card was initially due on 9/15. You didn't study it until October 1st. That's a greater than 2 week delay.
Despite the delay, you still thought the card was Easy. So Anki assumes (correctly) that it must be firmly fixed in your head and sends it even farther out into the future than it would have if the previous interval had been set to 2 weeks and you succeeded.
You also would have rated the card Easy both times. That tells Anki that the card is easy for you to remember, so you don't need to see it as often.
open plank
I see that cleared my doubts! Thank you
right soo about fsrs
I’m using the default algorithm and created a spreadsheet to calculate how many reviews per day I need in order to cover my cards within the number of days left before my exam. This approach worked for me during college, and now I’m applying it while preparing for my licensure exams.
When I tried FSRS and used the reschedule button, it ended up messing up my new and due cards, so I’m not sure if I did it correctly. I’m still new to scheduling methods, so I’m figuring things out as I go.
kinda like this lol
spark oasis
it ended up messing up my new and due cards
Rescheduling won't do anything to your new cards.
Rescheduling can either create a large backlog of review cards or reduce your daily load, depending on how much worse / better you're doing relative to your desired retention.
--
There's no way to guarantee how many reviews you have per day, because that wholly depends on how often you fail your cards.
open plank
i made a habit of doing it to reduce my bombarding myself with 400 equations per day are there more efficient ways to measure this
open plank
> it ended up messing up my new and due cards
Rescheduling won't do anything to...
alright ill give it a shot again
spark oasis
i made a habit of doing it to reduce my bombarding myself with 400 equations per...
In general, your daily review load will end up being 7-10x your daily amount of new cards.
But, if your material is unusually difficult or easy, it could fall out of that range.
The traditional lever for controlling your daily review load is controlling the amount of new cards you put in.
You can use the FSRS simulator to more accurately predict how your review load will grow and shrink based on your review history + settings for learning new cards going forward.
spark oasis
e.g.
open plank
"But, if your material is unusually difficult or easy, it could fall out of that range."
about this yea when i find something easy i just tend to suspend it like if i really know it
spark oasis
If you have review history, and it's good (you always press Again when you get something wrong), you should also just go ahead and optimize your parameters.
open plank
damn im having dilemma
open plank
If you have review history, and it's good (you always press Again when you get s...
ohhh i have one like 100 days streak will that help
but its in a different account ill transfer all my deck to that account now and try
spark oasis
Uh. So, to be clear, the data being good is just a matter of you rating the cards honestly. It doesn't matter if it's a streak or if you got a lot wrong or if you got them all right.
open plank
Ohhh okok so can simulate this deck without using that old account
spark oasis
As a general rule, more data is better. So pulling your old history in would be a good idea.
But, if you've changed your rating habits, then it'll do more harm than good.
spark oasis
Ohhh okok so can simulate this deck without using that old account
FSRS represents its understanding of your memory using the parameters. It generates that understanding by looking at your card history. Without any card history, it uses a set of default parameters that represent the average Anki user's memory. Those are fine.
Personalized parameters are almost always significantly better, which will result in a better prediction.
open plank
im sorry but what does tell me exactly in terms of optimization
sorry about that heatmap i just finished creating all my decks
spark oasis
im sorry but what does tell me exactly in terms of optimization
I'm not clear on the question.
open plank
like what does the graph tell me srry im stupid
spark oasis
The simulator shows you that once you get through your current backlog of cards, you'll have about 100 cards to do for this deck every day.
open plank
alright sec leme just think about that for a min brain is not braining
leme just put this here and use my brain on what this mean
say my exam is in 75 days i have 94 cards left? XD
spark oasis
The simulator is predicting your reviews.
New cards are blue. When you see them, they become learn cards (red). When you graduate from learn, they become reviews (greens).
Reviews are where Anki tries to make sure you're using your time efficiently to review the information and not forget it.
The simulator predicts that with the current amount of new cards you have, at the current rate of daily new card introduction (12) and desired retention (90%) you'll end up with about 100 daily reviews in this deck for the remainder of your prep period.
open plank
correct me if im wrong
so out of 700 cards all i retain is 103 cards at the end of 77 days?
OHHH
so i gotta fiddle with the new cards in the simulator to know if i will surely retain 700+ cards
spark oasis
...
open plank
im sorry
spark oasis
The fundamental point of Anki is to only show you cards when it thinks you're going to forget them, so you don't waste time.
The cards that become green each day are those cards and those cards only.
With default FSRS settings, those are the cards that Anki thinks you have a 90% chance of getting right today.
That means that all of the cards that aren't due today have a > 90% chance of being correct.
The simulator is showing you that on the last day of exam prep you will have 90 cards with <= 90% chance of retention (they would need to be reviewed that day) and 610 cards with a > 90% chance of being correct, if you had to answer them today.
open plank
if u were me what would your settings loook like assuming prep period of 75 days
open plank
The simulator is showing you that on the last day of exam prep you will have 90 ...
i gotta read more onto this
spark oasis
Here are two things to make things more complicated for you:
FSRS only learns about your memory for each preset (set of deck options). So, what it learns about your memory from this deck will not apply to other decks unless they belong to the same preset, which it looks like it doesn't.
We know that FSRS works better with more data, and we don't know for sure when it's best to split different types of cards into different presets. Because of that, my recommendation to you would be to make sure that all of your decks are using the same preset, so that FSRS can learn from all of your data.
You can use the Deck toggle to set different new limits while sharing the same preset.
spark oasis
If you only have 700 cards total and Anki predicts you'll need to review 100 of them a day, that's pretty crazy.
Your cards may be badly formulated. Here's a long page with advice.
You can still succeed with badly formulated cards, but just pointing that out in case you make future cards.
open plank
## Cards might be too hard?
If you only have 700 cards total and Anki predicts y...
wait how do you know it predicts that i need 100 a day
thats what i want
spark oasis
empty crest
i might pass of FSRS
anki without fsrs also behaves this way
open plank
dont get me wrong im trying to understand buy my poopie brain gotta read more to it
I believe it will highly optimize my self made new cards calculator lmao
balmy moth
so you would go review 9999 in a day?
The review limit is a limit. It's not saying "show me 9999 cards every day", it's saying "Never show me more than 9999 cards in a single day, no matter how many you think are due". That's why it's recommended to leave high; you control how many cards you see by how many actually due.
open plank
that is noted
But i just really wanna know how do i calculate how much new cards i need for my prep period with the help of FSRS so i can sleep peacefully
thats all i wanna understand and im out of ur hair guys
balmy moth
Two main ways: total cards you expect to need by your exam date / days until the exam, or total workload you can handle a day / 7–10. If those two ways of estimating new cards/day are out of alignment, then you'll need to decide what's going to work for you.
spark oasis
dont get me wrong im trying to understand buy my poopie brain gotta read more to...
open plank
yea thats what i did here
open plank
1. What is SRS and how does it work?
[Introduction from the Anki Manual](https:...
thanks man ill read into it!
yo i actually been doing the -14 days thingy thats sick never knew it was a widely accepted thing
spark oasis
yo i actually been doing the -14 days thingy thats sick never knew it was a wide...
Widely accepted in my heart.
It feels long enough that you can course correct if you realize during cramming you don't know enough, but short enough that you're mostly using Anki efficiently leading up to it.
open plank
It *feels* long enough that you can course correct if you realize during crammin...
bottom line i dont actually need to change my settings
just turn on FSRS
if i understand the guide correctly
open plank
yea thats what i did here
because i already scheduled it from hre
spark oasis
On a basic level, FSRS is not complicated. Just turn it on, press the optimize button sometimes.
I do think you'll get better results if you move all of your decks to use the same preset (which means manually putting the new card limit per deck), because that'll give FSRS more data to understand you.
open plank
i did that yesss!!
wait
am i doing it correctly while i think im wrong
lol
righty so here are my settings for each subject hope i did it correctly based on my calculated new cards from spreadsheet
from here
does this fuck it up
spark oasis
Math is wrong. Should be 8 not 6.
open plank
oh good catch thank you !
spark oasis
does this fuck it up
How could it?
open plank
like since urs in on a preset
aaaa nevermind Alrighty ill get off ur hair now thanks a lot man
spark oasis
If you have a question in the future, just create a new thread in #1266615749779390474 and someone who's bored will get to it.
Good luck!
open plank
Thank you monami
nocturne oak
righty so here are my settings for each subject hope i did it correctly based on...
You can set "This deck" new card limits and still leave the max review limits set to follow the Preset of "9999." There's no need to switch a deck to a "This deck" limit unless you want it to do something different than the rest of the decks using that preset.
open plank
just a follow up. Im suspending cards that i know at the back of my head while using FSRS. Are there any issues that will happen regarding suspending cards
Or do i jst delete them
balmy moth
The default settings ignore suspended cards when calculating FSRS parameters, on the theory that you weren't really studying them or noticed a problem, and they're not representative of your active cards. So with that setting, there's really no difference as far as FSRS is concerned between deleting and suspending. With suspending, you can change that if you do want to count those toward scheduling.
spark oasis
just a follow up. Im suspending cards that i know at the back of my head while u...
If it's something you didn't know and now you know super well, my advice would be to just be honest when rating and press Easy. It's good to let FSRS know about how your memory progresses with cards that become easy (because that'll let it bump up the intervals more quickly for future cards with a similar review pattern, saving you time with such cards in the future), and the amount of time you waste studying them should be minimal (because the intervals get big and you can answer quickly).
If you're studying from a premade deck that has stuff in it you already know super well, then I think suspending is fine. It would also be fine to be honest and just hit Easy. The interval will quickly grow.
open plank
If it's something you didn't know and now you know super well, my advice would b...
thanks
open plank
If it's something you didn't know and now you know super well, my advice would b...
Hi, I just wanted to expand on my previous question.
How can I estimate—using FSRS/Anki metrics—how many cards out of a fixed set (e.g., 500 cards) I’m likely to recall by a given date? I’m currently limiting my daily reviews to around 60 cards due to time constraints, but I’d also like to compare this to the scenario where I raise the limit back to 9,999. I know the higher limit will obviously help, but I’d like some numbers or metrics so I can have a bit more confidence in how much recall to expect under each setting. and are there like a metrics for the overall forgetting curve of a given deck?
I’m trying to identify which cards are most likely to be weak so I can focus on them during a planned 2-week cram period based on the scheduling I set up.
spark oasis
Use the FSRS Simulator built into all modern versions of Anki.
I’m trying to identify which cards are most likely to be weak so I can focus on them during a planned 2-week cram period based on the scheduling I set up.
The easiest way to do this is to create a filtered deck using the "Ascending Retrievability" card order. This will pull cards in order of most likely to forget to least likely to forget.
open plank
im quite not sure how to access the FSRS simulator while on a filtered deck
spark oasis
You can't. Do it from your regular deck.
(The simulator technically works via the preset. It will run the simulation on all of the cards/decks assigned that preset.)
open plank
I only have about 500 on my deck why does it tell me its about a thousand
spark oasis
The simulator technically works via the preset. It will run the simulation on all of the cards/decks assigned that preset
open plank
here is the settings
open plank
> The simulator technically works via the preset. It will run the simulation on ...
Ohh i see
spark oasis
Unfortunately, the only way to run the simulator on just that deck would be to clone the preset for only that deck (so it's just the one deck in the new cloned preset), save, then run the simulator, then reassign the original preset back to the deck (so you benefit from more training data).
open plank
Right ill try it out one second
spark oasis
(Also, it seems that you've already figured out that, when maintaining a permanent backlog, if your goal is to have the highest memorized count, Ascending Retrievability is the best sort order.)
open plank
Just cloned my original preset . Then set it to my deck that i wanna analyze
open plank
Unfortunately, the only way to run the simulator on just that deck would be to c...
So i must train it first before i can get data? is that you meant
spark oasis
No. You don't have to optimize or anything.
If you exit out of the options screen and re-enter, is PIPE TEST REVIEWER CLONE still assigned to the deck?
open plank
oh i should probably set it to the subdecks as well?
spark oasis
Yes. That would explain it.
Your parent deck probably has nothing in it, while your subdecks have all of the cards.
open plank
Ohhh that clearly fixed it! Thank you so much obe your the best
spark oasis
Now reward me by showing me a screenshot of how much you'll miss out on if you limit your daily reviews 😉
open plank
Right so my total cards are 459
I think i can get away with 60reviews per day assuming im doing math decks which takes me around 1-2mins
spark oasis
Looks like it won't hurt you!
Remember to reassign your Default preset to all of your decks again!
open plank
Oh right! Thank you again so much!
Damn that metrics motivated the hell out of me. That feels so good
U the goat sir
open plank
Hi obe its me again this would be probably my last question since i have 14days remaining before judgement day. Currently I identified my "Weak" topics outside anki by doing a pretest on all of my decks, which are these attached below.
I was wondering if there is a way to filter this out without affecting the algorithm like basically cram it until i get it / active recall kinda sutff.
open plank
Hi, I’m new to this. I last reviewed this card on September 12, and today is Oct...
damn time flies this thread was 2months ago 😭
spark oasis
Hi obe its me again this would be probably my last question since i have 14days ...
You can schedule your cards in a custom filtered deck and uncheck "reschedule cards based on my answers in this deck"
Unless you're going to run through the deck multiple times in a single day, I think it's fine to leave "Reschedule cards based on my answers in this deck" on.
Multiple reviews in a day can confuse the current implementation of FSRS, but otherwise it'll probably be fine.
open plank
You can schedule your cards in a custom filtered deck and uncheck "reschedule ca...
I see.
darn cant thank you enough 🙁
spark oasis
Pass your test. Become super wealthy. Donate to... Ankidroid, I guess. (Regular Anki doesn't accept donations.)
open plank
Pass your test. Become super wealthy. Donate to... Ankidroid, I guess. (Regular ...
Will do, sir. The passing rate is only 0.13%, but it doesn’t feel impossible anymore—unlike two months ago when I had no clue FSRS even existed. Not gonna lie, there were some burnout moments, but I stuck with it. I’ll keep you posted. Thank you so much. 👑
