If you are worried about this large interval, you can increase your requested retention.
Hi Sherlock, so I’ve been using your custom algorithm for a week.Recently, I’ve noticed some irregular intervals. New intervals should be at least 15 days/22 days, but these are only 2day/3day. I also optimised the w=…. Again, and it changed the values. Is there anything wrong? (Umm, I’ve added ~ 100 cards few days ago)
How often do I need to use the optimiser?
Do you use learning steps longer than one day? Please remove them. And graduating interval only works in the default algorithm.
My learning steps are 15 min, 1 day, 6 day
What does it mean?
So I need to remove the 15 min, 6 day, 15 day ?
I recommend removing 1 day and 6 day.
It means the graduating interval is controlled by FSRS, instead of the value in the deck options.
So what’s your expected behavior? If you think the interval is too short, you can decrease your requested retention.
1 Like
Can’t do anything without compromising retention?
@Nishat, the issue persists because you had 1d and 6d in the learning steps when you last studied this card. To resolve this, rate this card without thinking about the intervals. Then, use the FSRS helper add-on to reschedule the cards. This would correct the interval.
For more information, read the first answer in the FAQs.
I am sorry, I don’t have any computer. I use Anki on iPad. From the 1st FAQ, it says to keep the graduating interval low. Mine is 15 days, so should I chance it ?
(Previously my learning steps were 15 min, 1 day, 6 day, now they are only 15 min)
Not sure how easy this question would be to answer, but If I have a sizable amount of reviews and temporarily use load balancer would it have an adverse affect once I deactivate it? Having to redo some cards currently so I was wondering if temporarily using it would be an effective way to get them back up to speed without having to review them all at once. Would they naturally become more dispersed than they currently are once I deactivated the balancer? My idea is pretty much just to have them spread apart from each other after a little while, without having to temporarily remove them from the srs, and then deactivate the balancer. I know this would happen naturally if I just kept doing them without the balancer too, but perhaps in a less efficient manner being that I would have to straight up grind them down that way. I guess it might also have the downside of slightly messing with intervals of cards that don’t need to be redone.
Load balancer respect the fuzz range, so the interval would not deviate too much from its original schedule.
So the cards would still progress and spread out the same way they would naturally? Trying make sure I don’t get bombarded with a significant amount of reviews on a random day in the future once I do deactivate. You said it respects the fuzz range so does that mean It won’t harm my retention if I just leave it on too?
Yes. But fuzz doesn’t ensure a significant amount of reviews on a random day would never happen. So I still recommend using load balance.
Of course, it doesn’t harm your retention.
It seems load balancing is giving me an odd spread of intervals on new cards. I understand the point is to try and even out the amount due per day, but it wants to give me reviews for brand new cards anywhere from 1-5 days as a graduating interval which seems a little odd. Is there any way to force new cards to be reviewed the next day/2 days and have it balance everything else?
Could you show me some affected cards in your collection?
I can, I would have to search but I think I remember at least a few that were affected, and I can always just do a few more and note the affected ones. How should I show them to you though?
You can share the cards info here.
