# Ordering Request: Reverse Relative Overdueness
**URL:** https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051
**Category:** Suggestions
**Created:** [September 28, 2024, 8:10pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051 "2024-09-28T20:10:52Z")
**Posts on this page:** 20
**Page:** 14
### Author:  [@Expertium](https://forums.ankiweb.net/u/Expertium)
#### Post date: [October 1, 2024, 12:35pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/267 "2024-10-01T12:35:10Z")
Alright, here are the results
Default FSRS-5 parameters
```auto
desired_retention = 0.9
deck_size = 20000
learn_limit_perday = 20
review_limit_perday = 200
learn_span = int(deck_size / learn_limit_perday)
```
As you can see, without a backlog there isn’t much of a difference
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 12:56pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/268 "2024-10-01T12:56:39Z")
@sorata Here is with DR = 0.9, and 20/200 review limits.
The other graphs didn’t show much difference among them at all except this one.
@Expertium Can you explain what True Retention per Day is telling us?
---
### Author:  [@Keks](https://forums.ankiweb.net/u/Keks)
#### Post date: [October 1, 2024, 1:08pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/269 "2024-10-01T13:08:19Z")
> Desired retention is what you want. It’s your way of telling the algorithm “I want to successfully recall x% of cards **when they are due** ” (that’s an important nuance).
>
> True retention (download the Helper add-on and Shift + Left Mouse Click on Stats) is measured from your review history. Ideally, it should be close to the desired retention. If it deviates from desired retention a lot, there isn’t much you can do about it.
>
> Basically, desired retention is what you want, and true retention is what you get. The closer they are, the better.
[https://www.reddit.com/r/Anki/comments/1anfmcw/you\_dont\_understand\_retention\_in\_fsrs/](https://www.reddit.com/r/Anki/comments/1anfmcw/you_dont_understand_retention_in_fsrs/)
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 1:15pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/270 "2024-10-01T13:15:04Z")
Ok, so PRL does seem to maintain better proximity to the DR on the True Retention per Day, but it’s not much better on the other metrics.
---
### Author:  [@Expertium](https://forums.ankiweb.net/u/Expertium)
#### Post date: [October 1, 2024, 1:15pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/271 "2024-10-01T13:15:12Z")
What Keks said. However, in Jarrett’s code it’s a bit different. Actually, very different. It’s the average retrievability of cards reviewed today. It’s not based on binary 0-1 pass-fail outcomes, like the True Retention in the Helper add-on and now in the Anki beta. It’s based on retrievability.
Well, in this simulation it doesn’t matter, since in this simulation retrievability cannot possibly be inaccurate, but in real life it matters, since in real life retrievability can be inaccurate and the user can recall 90% of cards even when average retrievability was 70% today.
In fact, the current approach is _better_ for the simulation, because if the number reviews is low, average retrievability can be quite different from average of binary outcomes.
Example: 80% average retrievability. Both cards have been recalled, so both outcomes are equal to 1. The average is (1+1)/2=1.
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 1:27pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/272 "2024-10-01T13:27:06Z")
> [@Expertium](#):
>
> However, in Jarrett’s code it’s a bit different. Actually, very different. It’s the average retrievability of cards reviewed today.
This is what was confusing me. I read the code and thought it was dividing the entire sum, for the whole deck, retrievability by the reviews for today. Makes much more sense now. Thanks.
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 1:48pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/273 "2024-10-01T13:48:27Z")
> [@rich70521](#):
>
> I’m also gonna add the ability to make the cards studied per day resemble your personal study habits.
Running this now with my historical study habits and I probably wouldn’t get through 20k cards in a lifetime 😂
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 1:57pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/274 "2024-10-01T13:57:03Z")
With `difficulty_asc` it’d take me 36 years to get through them
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 2:01pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/275 "2024-10-01T14:01:16Z")
I can run someone else’s if you want. Just need your percent of days studied and your average reviews for days studied.
---
### Author:  [@Expertium](https://forums.ankiweb.net/u/Expertium)
#### Post date: [October 1, 2024, 2:08pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/276 "2024-10-01T14:08:58Z")
I just got a really cool idea - iterating over different review card limit/new card limit ratios at different desired retentions to find out the ratio that doesn’t cause a backlog at every level of DR. Though, of course, this will be based on default parameters, but still. It will be better than the current rule of thumb “You will have 10x more reviews than the number of new cards”.
Will take forever, though.
EDIT: wait, I’m dumb. I don’t need to iterate over ratios, I just need to uncap review limit and see how many reviews the simulated user gets, on average
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 2:59pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/277 "2024-10-01T14:59:05Z")
The results after inputting my study habits in for reviews/day:
As annoying as it is because the formula isn’t really intuitive, I might just have to admit `difficulty_asc` is the best and start using that. Minimizing reviews is what I want to do, and it’s been winning on that front pretty much every time.
---
### Author:  [@Expertium](https://forums.ankiweb.net/u/Expertium)
#### Post date: [October 1, 2024, 3:00pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/278 "2024-10-01T15:00:50Z")
To make it accurate you would need
1. Your own FSRS-5 parameters, which you can’t get right now because the optimizer in Anki 24.10 beta isn’t working properly
2. This stuff
Which you can’t access, Anki calculates these in the background when you run CMRR
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 3:03pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/279 "2024-10-01T15:03:05Z")
Yeah, I’m ok with not getting it that exact. Not sure it would make a huge difference with which perform better. Seems like no matter what variables I change, the rank of sorts stays pretty consistent.
---
### Author:  [@sorata](https://forums.ankiweb.net/u/sorata)
#### Post date: [October 1, 2024, 4:58pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/280 "2024-10-01T16:58:08Z")
The biggest takeaway for me is that `random` is actually performing really well without a backlog. It starts performing really bad only when you have a backlog. Assuming this simulation captures at least some of reality, I see no reason why `random` cannot become the default sort order. It’s not as good with maintaining a constant retention though, that I see.
Or if not, what about `difficulty_desc`? For some reason it seems we’re instead suggesting dae to make `least overdue first` the default. Any reason why?
---
### Author:  [@sorata](https://forums.ankiweb.net/u/sorata)
#### Post date: [October 1, 2024, 5:02pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/281 "2024-10-01T17:02:53Z")
Another thing that I dont get is why `difficulty_asc` suddenly performs really bad when we stop learning any new cards in the previous sim. I would like to see what a simulation looks like when youre not learning any new material and just maintaining previously learnt stuff.
---
### Author:  [@Expertium](https://forums.ankiweb.net/u/Expertium)
#### Post date: [October 1, 2024, 5:20pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/282 "2024-10-01T17:20:33Z")
Done
---
### Author:  [@Expertium](https://forums.ankiweb.net/u/Expertium)
#### Post date: [October 1, 2024, 5:24pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/283 "2024-10-01T17:24:05Z")
Retrievability descending performs well both with and without a backlog, random performs well only without a backlog, and even then R desc. performs better. Why would random be the default?
> For some reason it seems we’re instead suggesting dae to make `least overdue first` the default. Any reason why?
Retrievability is less arbitrary and more interpretable, and will likely work better in edge cases.
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 5:30pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/284 "2024-10-01T17:30:02Z")
I tried to switch it up and put a 500 day gap at the 2500 day point to see if the consistency was affecting anything. It’s not rare for me to take a year off. `difficulty_asc` is still basically the best 🙄
I just don’t want to base my sorting on something that isn’t dynamically changing with each passing day. Difficulty only changes when you answer a card, and it’s only affected by what answer you give it, it doesn’t factor Retrievability at all.
---
### Author:  [@sorata](https://forums.ankiweb.net/u/sorata)
#### Post date: [October 1, 2024, 5:32pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/285 "2024-10-01T17:32:12Z")
Even though most or your cards have very high D values, possibly with a lot of similar numbers? Its surprising how is that happening.
---
### Author:  [@rich70521](https://forums.ankiweb.net/u/rich70521)
#### Post date: [October 1, 2024, 5:34pm UTC](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051/286 "2024-10-01T17:34:33Z")
> [@sorata](#):
>
> Even though most or your cards have very high D values, possibly with a lot of similar numbers?
The simulation isn’t using my actual deck, so I have no idea what the D distribution is during the sim. Maybe it’s reflecting my deck somewhat because I changed the daily reviews, but that’s not guaranteed. But yeah, I am surprised.
[Previous page](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051.md?page=13)
[Next page](https://forums.ankiweb.net/t/ordering-request-reverse-relative-overdueness/50051.md?page=15)