Average Card Ease Percentage Incorrect

I noticed this and it seemed like a bug so I wanted to be sure to report it.

Under my card ease tabs, I have 280 cards at 130% ease and 296 cards at 131% ease, but my average ease shows up as 133%. The aforementioned 576 cards make up all of the ones that I have started learning so far, so the average ease percent does not seem correct. Looking at my other decks, I see that the percentages appear to be off as well.

Not sure if there any recommendations on correcting this, but if so please let me know! Thanks for your time.

Also, my version is 24.06.2 on windows

You posted this in the add-ons category, but you don’t say what add-on you think is causing this.

If it’s not related to an add-on, I’d want to see what is actually going on with the underlying cards. What do you see in the Browse window when you click on either of those bars? Do you have any suspended cards? Are there teeny-tiny nearly invisible bars on any of those other numbers?
Ex (from a different graph):

[I’m setting aside my questions about why you have such absurdly low Ease on every card – because you don’t seem concerned about that at all.]

My apologies, I misread the category descriptions when posting this initially. I have changed it now as this is not related to an add on.

When I click on those bars, I get the two attached images. It just shows the cards that I have for review in the future.

I do however have a lot of suspended cards within the deck (i suspended them right away when I got the deck)

I hovered over the invisible columns and all of them say 0.

My other decks that are having the same percentage issue discrepancy have some suspended cards, but no where near as many as this deck.

The reason for the card ease being so low is due to the low key anki setup I am using. I have gone off of what they recommend, so the low ease does not concern me: https://refold.la/roadmap/stage-1/a/anki-setup

All of that information helps. Now search for deck:current is:suspended -is:new or deck:current is:suspended is:review – anything?

Yes, refold has strong feelings about that. If you’re happy with it, and it makes sense to you – then I won’t upset the apple cart.

When I searched for that in the deck above I did not get any results (please see the top photo)

I did check it in the other decks and received results, example of that attached as well.

I appreciate you mentioning at least that the card ease is super low, I doublechecked and made sure my settings matched theirs so it looks like this is how they have intended.

Run that search again on the 1st deck, but choose either one of the options, not both. I want to eliminate the chance that the terms are misgrouped.

deck:current is:suspended -is:new

deck:current is:suspended is:review

[I can’t do much with the 2nd deck’s result, because I don’t think you’ve posted any Stats data for that deck to compare it to, right?]

I tried them again separately but came up with no results on either of them.

(That is correct, I have not. Just wanted to throw it in in the case it was useful info)

Since it’s nothing that’s visible on the surface – it’s possible there are some card-ghosties skulking around in your database. Have you run Tools > Check Database?

I just gave it a go, and it said that my database was rebuilt and optimized. However, it appears that the average card ease percentage is still showing as 133%.

Btw, thanks for all your work as you have been helping me try to figure this out.

Thanks for trying that!

I don’t see anything in the issue tracker similar to this, but let’s see what @dae wants to happen next with this one. I don’t know if he believes in database ghosties, so he may want to grab some of your data to figure out what is averaging incorrectly.

1 Like

If you’re using their recommendation of …

We strongly recommend ONLY using the “again” and “good” buttons, and avoiding the “hard” and “easy” buttons.

Then your ease will inevitably trend to 130% (or whatever your minimum is), since there is no way to raise it.

With FSRS, this recommendation is fine, but with SM2 I think it’s a big mistake, for just this reason.

But the average being outside the 2 values that you seem to have is definitely odd.

I do use their recommendation of only those two buttons, could you clarify what FSRS and SM2 mean?
(Disregard, I found out what they are)

I have the latest version of anki, so is this something that I can see within my settings? Currently, I don’t know which I am using.

You’re definitely using SM-2 (not FSRS) – because you still show Ease in your Stats.

It’s not really a “mistake,” since it’s by-design in the approach that Refold / Low-key takes. They argue that since under the SM-2 algorithm your Ease could potentially bottom-out, you might as well start there and stay there. And then they mess with other settings to “fix” the rest of the algorithm. It’s a system to make Ease meaningless – which makes sense if you think Ease is the root of all evil, but doesn’t if you think Ease is a valuable tool for learning.

FSRS doesn’t use Ease at all – but uses a more complicated (and accurate) system built around the Difficulty of a card.

1 Like

I see, that makes sense. Thank you for the explanation. I want to do more research on all of this before I change anything, this is something I need to think more about.

1 Like

Thanks for the report; this is a known issue. I’ve added it to the issue tracker: Averages in graphs are not completely accurate · Issue #3269 · ankitects/anki · GitHub