Anki Forums

Anki v3 scheduler (2021) review number mixed up

Hey everyone,

the new scheduler v3(2021) finally implemented a function I was waiting for an eternity (randomizing new cards from subdecks). Unfortunately, I now have a problem regarding the amount of reviews.My main deck “JLPT” (with it own option setting) (pic 1&2) consists of 5 subdecks (3 of them deactivated). The subdecks “JLPT 1-5” also have their own shared option setting (pic 3&4).What I dont understand is the amout of cards which are queued as review cards in the main menu are set on 200 (pic 5) (like I set it in the options), but when actually starting to study (parent deck or subdeck) the number of cards to review is decreased… (pic 6&7). I understand that I dont get new cards when the number of cards to review reaches the set limit of 200 (it is well described in the option settings) but I have no clue why the actual review number is lower. It has to do with the v3 (2021) scheduler since I did not encouter this problem prior to updating or in ankidroid (which still uses the v2.1 scheduler).

Thank you in advance for your answers!



Damien, thank you very much like usual!

So you are telling me that in the old scheduler version it maybe has displayed 200 cards to review but sometimes it would have just skipped from e.g. 150 to 148 (without me ever noticing this?).
And for the new scheduler/or in generell: does this mean when I use the burying function i cannot review the amount of cards that I picked (e.g. 200)? Or how do I get 200 cards after burying was applied? Since I want to review 200 cards, I “lose” cards, since several cards are getting pushed to next day as a result of burying.

If you have exactly 200 cards waiting for review, and some of them are buried, then you’ll get less than 200. If you have more than the daily limit waiting, then extras will be taken to make up for the buried ones. In the former case, the old scheduler reduced the number as you reviewed, whereas now it’s happening up front.

OK, I am sorry for asking this somehow again. I checked today: The old scheduler would have given me 200 review cards while the 2021 scheduler’s number was about ~140. So you are saying that the difference of 60 cards in the old scheduler is getting filled with other cards (you called them “extras”), since I did not see that the counter of review cards miraculously skipped a number of cards. Where do these extra cards are coming from then? And why does the new scheduler not grab this extra cards aswell? If the new version does not fill up the difference of review limit and actual review cards (after sorting out the buried one), is this a a step back? Is the old version not more useful since I want to learn the about cards that I selected…

And maybe hitting in a similar notch:
About the new Intraday learning. It says: “Learning cards that cross a day boundary are now included in the review count, instead of the learning count. They will now be affected by the daily review limit, and can be automatically buried like review cards.”

My intervals for learning lapses are 3m 10m 1d. So in the last step the learning card will automatically transferred to the review stack? Are those cards still treated as a learning card in terms of pressing the again button will not alter the ease? And if it is still treated as a review card how can I see the card’s status while studying? How do I know for sure if I am just studying a learning or a review card? Will there be an option in the future to deactivate this, since I do not want my learning cards to affect the numbers of reviewing cards.

By “extras”, I was referring to any review cards that were still waiting. If you don’t have a backlog, then fewer than 200 cards would be shown. The v2 scheduler would start at 200, but when cards are buried the remaining count would be reduced by more than one internally, and at one point the number would jump down.

Learning cards with 1+ day intervals appear in the review count, but are not assigned an ease until they graduate. You’ll know it’s a learning card because Hard would repeat the 1 day interval. There is no option to exclude interday learning cards from the review count, but I will keep an eye on demand.

Over the last days I checked my decks while studying: Between the v2 and the v2021 scheduler there always was a significant number of differences between the two scheduler’s review cards. So if I understand you correctly: while doing the review cards (the sibling burrying of course is activated) in the v2 scheduler, cards that get burried should simply getting skipped while studying meaning the review counter would skip some numbers e.g. 150 → 148? However, I never saw that happen…

And does the v2021 use the backlog function aswell? Meaning that if there is a difference in the schedulers versions’ review cards, it is simply due to fact that there are no cards waiting (“in the backlog”?). But again this would mean, since the differences between the schedulers review cards (which was in my case always >50), would be only due to burrying. But again, I did not see any “skipping” of cards

May I ask, why the change with the learning cards (>intraday interval) getting shifted to reviewing was implemented?

I was waiting like crazy for the randomazation of new cards from subdecks function. Now it has finally arrived and unfortunately some of the options, as u can see in my weird questions, are a little bit confusing me… Asking as someone who is not experienced in programming, how complex is it to implement optional settings for these new implementations e.g. the learning cards getting moved to review?

In v2, the numbers will jump when a new batch of cards is fetched - every 50 or so. You can confirm how many cards were buried by searching for is:buried at the end of the review.

Because when they have intervals of a day or more, they are effectively reviews. Intraday learning cards are excluded from the limits because they are time sensitive, and may appear multiple times in one session.

And will the learnings cards that are getting put into the review deck also be taken into the review limit or are they are running independently?