What was the last Anki version that had no fuzz?

Anki Help
1

I cannot find this information on the github releases page or the Changes portion of the website.

1 Like
2

I think fuzz is as old as Anki is itself so you’ll probably need to check very old versions to find that out. Any reason why you’re doing this?

1 Like
3

It’s completely useless to me, my retention has tanked since I upgraded to fuzzy Anki versions, I see no way to schedule cards around it, and I’m done wasting time trying to work with it.

I’m fine with using old Anki versions. My previous version was 2.1.15, so I know that fuzz came after that.

1 Like
4

This is likely due to the V3 scheduler, which has been introduced 2021:

The small random delay added to intervals is reflected on the answer buttons, instead of only being applied when answering.

1 Like
5

You seem to be mistaken. Version 2.1.15 came out in 2019. As you can see in those changelogs, and the subsequent Release Notes, fuzz was not introduced in the years since then.

In fact, I’ve searched through the archives on the old support site, and I can tell you that fuzz was alive and well and documented for years before that. The earliest references I found on a quick search were from 2017 (pre-Anki-2.1) - and that wasn’t it being introduced as new!

As ferophilia said, you’re just noticing it now because of the v3 scheduler [which you’ve complained about before, and you seemed to have successfully turned it off already]. Fuzz has been applied all along, and it never hurt you before, so I don’t see why it would start now. We might be in an XY-problem situation – you’re trying to cure fuzz, but fuzz is not new, so it can’t be the cause of your new issues.

From your other post –

linux-user:
I see no reason why fuzz should be mandatory, and seeing cards in the same order and/or close together doesn’t necessarily negatively affect memorization or recall; however, the fuzz factor itself can negatively affect retention by causing difficult material to be scheduled later than it needs to be seen to be remembered.
… I’ve wasted hours trying to learn, failing, and relearning cards since upgrading to Anki versions with V3 and fuzz.
[/quote]

I think you mistake the amount of impact fuzz can have on scheduling for any single card. Other than fundamentally disagreeing with the idea of fuzz, if you want to say more about what problems you think it is causing you, and how your retention has changed, perhaps folks will have some helpful suggestions.

1 Like
6

The fuzz is very, very unlikely to be the cause of sudden retention fluctuations (or, really, nearly any problem).

As others have noted, there are many other more likely causes for which folks are happy to help given more context.

4 Likes
7

Indeed, and in fact, getting rid of the fuzz would potentially act against your goal, so negatively impacting the retention overall due to predictable sequence. Having a better recall short-term due to predictability is just harming yourself in the long-term.

Perhaps the cards are too hard, not well formulated, or you didn’t do them (share decks, AI, where some knowledge gaps exist)

Btw, I think Anki 1.2. already implemented some degree of fuzziness. That was like 12-15 years ago.

3 Likes
8

IIRC in v2 scheduler fuzz wasn’t reflected in the intervals above the answer buttons, so the intervals appeared to be deterministic but actually weren’t, and then in v3 the intervals above the answer buttons started showing fuzzy intervals.

Also, here’s the exact amount of fuzz that Anki uses:

static FUZZ_RANGES: [FuzzRange; 3] = [
    FuzzRange {
        start: 2.5,
        end: 7.0,
        factor: 0.15,
    },
    FuzzRange {
        start: 7.0,
        end: 20.0,
        factor: 0.1,
    },
    FuzzRange {
        start: 20.0,
        end: f32::MAX,
        factor: 0.05,
    },
];

If the calculated interval (before rounding) is less than 2.5 days, there is no fuzz.

If the calculated interval is between 2.5 days and 7 days, the fuzz is 15%, meaning that the interval length will be changed by at most 15%.

If the calculated interval is between 7 days and 20 days, the fuzz is 10%.

If the calculated interval is longer than 20 days, the fuzz is 5%.

4 Likes
9

I want nothing more than an answer to my question.

@dae please help me.

10

As it was mentioned earlier: Fuzz has been around for at least a decade.

From the archives (10 years ago, https://web.archive.org/web/20141129141938/http://ankisrs.net/docs/manual.html):

After you select an ease button, Anki also applies a small amount of random “fuzz” to prevent cards that were introduced at the same time and given the same ratings from sticking together and always coming up for review on the same day. This fuzz does not appear on the interval buttons, so if you’re noticing a slight discrepancy between what you select and the intervals your cards actually get, this is probably the cause.

4 Likes
11

Perhaps Fuzz is embedded in the scheduler. There are 3 types of schedulers: V1, V2, and V3, the latest version of Anki defaults to V3, for Anki 2.1.15 the scheduler you were using is either V1 or V2.

These Anki versions allow you to use older schedulers (you need to change the scheduler in the preferences):

To completely disable Fuzz you need use V1 or V2 and use the Fuzz disable add-on. Almost all add-ons for Fuzz have been deleted but there is one source code available on Github (not sure if it works. Download the src file and put it in the addons21 folder and set manifest.json to use it).

I think it would be ideal to disable Fuzz by add-ons in the latest Anki and V3 scheduler, but that may not be technically possible to develop by add-ons. (not sure exactly)

3 Likes