Best way to review before due date

Hmm, I think the safer option would just be raising the desired retention to 99% if Advance is not reversible. Problem is that my deck has a lot of other subdecks, each with their own preset. :disappointed:

Is there like a debug console code with could raise or lower desired retention rates for a bunch presets at the same time. :question:

For today I mean. Do advance cards, that have not been reviewed for the lack of time, go back to their original memory state and scheduling after one day has passed :question:

The memory state (DSR values) doesn’t change if you change the due date. It only changes if you change parameters or do a review. And no, the due dates of cards don’t change once you have advanced them, unless you click Undo.

Well it seems then the advance feature is not as good as I thought even for a one time review thingy since the cards will stay like they are. I guess at this point it would be for the best if I just make a custom study session with filtered decks (scheduling enabled).

How bad do custom study sessions mess with scheduling tho :question:

Or should I rather disable the scheduling from the custom study session.:question:

Because about a year ago, I did a very brain-dead thing with Anki where I put the max interval limit at 2 and it has really made FSRS get a messed up idea of the actual difficulty of the cards.

They don’t “mess with” scheduling. But reviewing ahead has some special logic, and it is recommended to be used sparingly – Filtered Decks - Anki Manual . If you still want to use it, you can obviously filter the cards to get to ones that will reduce the downside/time-waste – e.g. cards that are due soon, cards with longer intervals, etc.

That just makes the studying a waste of time. :person_shrugging:t4:

Is this exactly true for FSRS? According to FSRS logic, no special algorithm is needed to process preliminary reviews, since it takes into account the value of “R” to build a new interval.

Do you recommend saving the parameters somewhere in case the filtered deck with the scheduling does end up skewing the scheduling in some odd suboptimal way as. My exam is exactly in a month from now, so I figured I wanted to start from now, since a lot is on stake.

I have 10000 cards for the deck in question and my current average predicted retention is 98.5%ish but I still do feel like I am forgetting a lot continuously and that I actually have forgotten a lot more than what this percentage suggests.

@Expertium Would that mean that FSRS doesn’t really care if you review before optimal time or not, because it will just calculate new values based on the custom session :question:

FSRS doesn’t separate cards into two distinct categories, like “due” and “overdue”, or “young” and “mature”. In fact, FSRS doesn’t care about anything other than interval lengths and grades (well, “Compute optimal retention” uses additional info, like how long it takes you to press the answer button, but whatever). The increase in memory stability depends on R, the probability of recall at the time of the review. If R is low, stability will increase more. This means that if you reviewed your card later, you will receive a “bonus” to your next interval.

As far as I can tell, the FSRS formula has similar logic and downsides built into it. To compute R, you need t.

t is the amount of time elapsed since the last review

Like in a completely overkill saves-every-set-of-my-old-parameters-in-a-spreadsheet-for-later-study-like-a-weirdo sort of way … sure I can endorse that behavior! But I don’t see how that would help you here?

While you are studying those cards in that filtered deck, the exact same parameters and FSRS formulas are applying to them as when you study them in a regular deck. There’s no difference.

Wait. But I thought the parameters will change once I optimize, since I would have turn on scheduling change in the custom session :question:

Well, sure, I suppose if you re-optimize, and the preset has new parameters, cards whose home decks use that preset will start using those parameters. But why are you now talking about re-optimizing your parameters in the middle of studying in your filtered deck? :person_shrugging:t4: Just leave your parameters alone and get your studying done.

1 Like

Huh. So maybe it is not so bad after all reviewing before the original due dates, I assume if I understood correctly :question:

For example:

What would FSRS would do to a card that shows a 1.1 month interval if I press good in a custom session, AND what would happen to the same card if i press good once more in a different custom session the day after :question: Would it still be 1.1 months or some astronomical number

So like this:

  1. You review a card after 1.1 months
  2. You then review it on the next day

Your R will be close to 100%, and as a result SInc (increase in stability) will be close to 1 aka no increase, since S*1=S. Therefore, your next interval will be only marginally longer.

1 Like

Okeeee…so this does clear things up a bit. Is this a reason why doing custom sessions repeatedly is not recommended :question:

Because if I were to do it repeatedly, this basically nullifies FSRS and spaced repeition all together, not that there is something wrong with custom study sessions themselves, no :question:

I’ve never used custom sessions or filtered decks or whatever myself.

1 Like

But theoretically, the worst that could happen (and generally the reason why doing custom sessions repeatedly is discouraged) is that the intervals would not change and because of that one is putting in a lot more extra work but getting little benefit. But if one later wanted to return to normal, there would be no consequences.

Have I understood that correctly :question: