how far out does the app schedule?
I've had some cards that haven't popped up for a long time, seems like over a year. But maybe I'm wrong. If you score a card right, every time, does it eventually stop scheduling them forever? Assuming that you do answer a card right each time so that it gets out to a one year interval for its next scheduling, if you answer it right yet again, what is the next interval beyond 365 days? Is there a way to cap it?
Thanks
Zeppo
Comments are currently closed for this discussion. You can start a new one.
Keyboard shortcuts
Generic
| ? | Show this help |
|---|---|
| ESC | Blurs the current field |
Comment Form
| r | Focus the comment reply box |
|---|---|
| ^ + ↩ | Submit the comment |
You can use Command ⌘ instead of Control ^ on Mac
1 Posted by support on 05 Nov, 2014 06:57 AM
Hi Zeppo,
The details of the learning schedules are complex, and we deliberately don't go into great detail in the help pages. We want the tool to be accessible to everyone, not just people who know a lot about flashcard systems.
But if you are into flashcard systems, here are some details about what we are doing behind the scenes...
The 'Cram' schedule is based on the Leitner system, which uses a series of boxes. There are 5 boxes. Cards start in the first box. If you get them right, they move to the next box. Get one wrong and it goes back to the first box. The system makes the notes due that are in the lowest occupied box.
Cram is the only system that is not time based. You can just keep refreshing your due notes to make more due when you finish a slideshow. This schedule is best when you are in the last days before a test or exam and you just have to keep going over your notes until they stick.
The other systems all have a time element. The standard 'Long Term Learning' schedule is based on one of Supermemo's algorithms. The schedule begins with a sequence of increasing time intervals. Assuming you keep grading a note correct, the note will first become due again after 1 day, then 6 days later, 16 days later, and so forth. Get the note wrong, and the note starts the schedule again.
(Even this is a simplification of the actual algorithm, because the intervals are also adjusted based on how well you answer a particular note. Answer wrong a few times, and the intervals for that note shorten to help you out.)
The 'Date-Targeted Schedule' is unique to Mental Case. It is basically the Supermemo algorithm, but it dynamically adapts to try to ensure a certain likelihood of recall on the day you target (eg exam). The intervals are increasing just like the long-term schedule, but can be dynamically contracted or extended to try to ensure you know the information on the day you need it. The schedule thus combines the efficiency of the long-term schedule, with the pragmatic need to pass an exam. It works best when your target date is a few months away.
Best,
Sam R. — The Mental Faculty support
support closed this discussion on 05 Nov, 2014 06:57 AM.
zeppo re-opened this discussion on 05 Nov, 2014 03:45 PM
2 Posted by zeppo on 05 Nov, 2014 03:45 PM
Yes, but my question was a very specific one. Assuming you answer a card as correct each time using the Long Term Learning mode, what happens? Does it eventually stop scheduling cards for good? If not, once the interval gets out to around 1 year (assuming it does), what is the next interval once you answer a card correct?
In other words, what are the intervals between 16 days and around two years time?
Is there a way to see when a card is next due to be studied? (For instance, can I check a card and see that it is not due to be tested again for another 67 days and counting, etc?)
Support Staff 3 Posted by drewmccormack on 06 Nov, 2014 08:51 AM
Hi Zeppo,
We use a long-term algorithm very similar to what is described in the link to supermemo. We don’t do anything special, like stop the schedule after a certain time. It goes on forever.
If you look closely at the linked page, you can see that each interval is multiplied by a so-called easiness factor (E) to get the next interval. The E value can go up and down dynamically. If you get a particular note wrong a lot, E will be smaller, causing the intervals to shorten. If you always get it right, the interval will be longer. The largest E value is 2.5, so if you keep getting a note right, E will be 2.5.
What that means is that if the time interval is 1 year, and you study the note and get it right, the next time interval will be 2.5 years. It goes on like that forever, unless you mark the note wrong, at which point it starts the whole schedule again at 1 day.
You can see when a card is next due. On the Mac, double click the note, and then click the button bottom-right to show note info. The ‘Next Scheduled’ date is when it is next eligible to become due for study. (It may not actually become due, depending on other settings, like whether you have set a limit on the number of due notes allowed.)
On iOS, tap a note to show the details, then swipe to the very bottom and go into Study History to see the next due date.
Hope that helps.
Kind regards,
Drew
4 Posted by zeppo on 13 Nov, 2014 12:15 AM
Thanks for your reply. I appreciate you figuring that out for me. Helps a lot. Seems like quite a leap to go from a one year interval to 2.5 years. But I guess if you get it correct it the first time you are tested after not studying it for a year, it is probably pretty solid in your long term memory.
If I understand the Efactor thing then, I guess if someone gets a card wrong and the card gets set back to a 1 day interval, it is not exactly the same as getting set back to "square one" (i.e., a card being studied for the first time). The E Factor of a brand new card is 2.5 by default. But the card that gets set back to a 1 day interval will have an Efactor less than 2.5 due to it having been missed on the last attempt. Or does the Efactor get set back to 2.5 as well?
Zeppo
Support Staff 5 Posted by drewmccormack on 13 Nov, 2014 06:51 AM
Yes, the efactor will be less if your schedule is restarted, but can build up again with a few correct responses.
The intervals are scientifically based, including 2.5 years. Indeed, if you answer right every time, and even after 1 year, you probably know that info pretty well, and can handle 2.5 years.
Kind regards,
Drew