I’ve created a card so I can query a particular subject to practice a set of my flashcards:
{{cards [[German Verbs]]}}
However, it worked only the first time I practiced, if I want to practice again it keeps me showing the message:
Congrats, you’ve reviewed all the cards for this query, see you next time!
I also noticed if I add a new card into the query, I can practice only with the new card, and it shows me the message again,
Thoughts?
It seems to be “by design” like that and it’s unfortunate that one can’t “reset” the counters somehow from within Logseq.
I hit the same wall and the only workaround was to use VSCodium to do a global Search&Replace with REGEX, like this:
# Search:
card-next-schedule:: .*$
# Replace With:
card-next-schedule::
This basically gets you back to marking the “Overdue/Total” numbers the same with “Total”, so if you have 100 flashcards, it would write “100/100” and will go down from that as you review cards (“99/100, 98/100, etc)”
Unfortunately I can’t provide a way to only affect a specific wikilink like yours ‘[[German Verbs]]’ and the above will reset ALL your flashcards sets, but for me it doesn’t matter atm. I am still hoping that, because I haven’t touched the other cards-specific properties like for ex:
card-last-interval:: 44.08
card-repeats:: 3
card-ease-factor:: 2.7
card-next-schedule::
card-last-reviewed:: 2024-12-10T07:16:33.027Z
card-last-score:: 5
, it will take into account those numbers and the cards that have a poor score will re-surface faster/sooner than other cards, as well as cards that have fewer repeats should come first but I have no idea about the algorithm used.
Hope it helps.
The Flashcards feature uses an algorithm (SM5 in Logseq Org/MD, replaced by FSRS in Logseq DB). These algorithms are designed to serve the flashcards towards the end of the forgetting curve, so you don’t waste time cramming stuff you weren’t likely to forget anyway.
If you do want to cram, I recommend writing a regular query. Then you can just test yourself by collapsing all the child blocks and expanding each individually to check your answer.
In Logseq DB, the #card page has some tools to show and filter your flashcards which could save you writing a query. It even gives easy access to the Due property and lets you change it:
For me “cramming” should mean “condensing” the time-span the algorithm now hard-codes in days/hours into much smaller chunks of time, dynamically, depending on the speed the user goes through the cards and the number of cards in the set/filter. If I have a set with 100 cards I can easily go through them twice a day but they should still resurface in an order that would bring the ones I forgot or remembered with difficulty faster than the ones I marked as “remembered” so that a forgotten card could come up twice or more times before a “remembered” card comes up again.
I don’t know if such an algorithm exists and don’t want to propose a new one here. With the current Logseq implementation in the non-db version I hope that, if I remove the card-next-schedule:: value, it still takes into account the other properties that dictate what cards should come up faster.
