these buttons are basically useless now that we have FSRS and are potentially even harmful. I do not ever intentionally hit these buttons. On desktop I have addons to remove them, but on phone sometimes I fat finger hit them and its an annoyance. Can we get a setting to remove these to make anki truly pass fail on android?
The Hard and Easy buttons are neither useless or harmful when you’re using FSRS. I don’t know where you got that information, but it’s not accurate.
If you don’t want to use them, you certainly don’t have to, but there’s nothing wrong with using them.
I’ll let someone else comment about the status of AnkiDroid’s “New Reviewer” project, and whether an option for 2-button grading is already planned for that.
Regardless of your opinion on easy and hard, we can agree that accidentally hitting easy and hard would hurt you either way. These buttons are functionally useless for people who don’t use them aside from being landmines that screw up your algorithm if you hit them.
No, we can’t agree on that. An accidental grade isn’t the end of the world. But your opinion and suggestion are clear, the hyperbole doesn’t help you much.
If you need to avoid hitting them, you have options available to you now. You can make the buttons larger – so they are easier to aim at. You can arrange them in 2 rows – effectively removing those treacherous Hard and Easy buttons from your reach. Or you can turn them off entirely and associate gestures with just Again and Good.
This suggestion is already implemented! But only in the new reviewer screen. It’s currently under development so it doesn’t have gestures and whiteboard feature. If you still want to use it, follow these steps:
- Go to Settings > About.
- Tap on the Anki logo multiple times.
- Tap OK on the pop-up that appears.
- Go back to the previous screen.
Now, you’ll see a developer options category. Turn on the new reviewer from there and the option for 2 button mode.
About whether 2 button is better or not. It certainly is even though some people will not agree with you and me. Analysis show us that 2 buttons result in lower RMSE than 4 buttons use. I assume you read that somewhere before. Here’s the data for that: Pass/Fail Grading as Default - #36 by Expertium
One more important aspect is that when we look at people who claim they are consistent with how they use the 4 buttons, the comparison is different. Here, we see RMSE being almost the same. So, if you use them consistently (whatever that might mean) you can get decent results. Here is source: Pass/Fail Grading as Default - #120 by Expertium
One common argument that you’d hear is something like “four buttons → more choices → probably better” but that’s simply ignoring the other aspects, for example, the params and the intervals are probably very different for 4 button users vs 2 button users. It doesn’t help to have more choices if having more choicess mean having inferior choices.
Further analysis after that was inconclusive, as you see in the second post that you linked. Too bad I can’t edit the first post now to say it.
Okay, say it now. Or do you mean the consistent vs inconsistent thing? I don’t think that changes the equation much. I mean, I can’t try to be consistent because I have no idea what it means precisely.
I mean I can’t be confident about “consistency” right? It’s some weird magical thing.
It’s just:
- Hard = pass, not fail
- Don’t press Easy on hard cards
- Don’t press Hard on easy cards
That’s it, it’s that simple
As for statistics, I would need a much larger sample of consistent 4-button users to draw a conclusion, and collecting that much data (hundreds of collections from such users( would be extremely difficult.
Oh, did you explain to the users your standard for consistency?
Don’t say you just assumed…
Okay, then it should be safe to say 2-button vs 4-button doesn’t matter for users who know their sh*t (maybe 4-button is slightly better but probably not that much). If you however are the average user, it’s probably better for you to move towards 2 button use.
The thread isn’t locked, so I’m not sure why it isn’t editable. Discourse might just be protecting it because it’s old/has many replies. If there’s something I can do to make it editable for you, let me know. Or if you just want me to add a note to it, I can do that too. You can DM
@ moderators
.
I think editing is allowed only for the first [insert number here] days since posting. You can add a “Further research was inconclusive, see Pass/Fail Grading as Default - #120 by Expertium” note maybe
I genuinely have no clue why this suggestion is so controversial… I’m not going to address any of these semantic points.
I never hit hard and easy on purpose. I sometimes hit them on accident and its aggravating to know that I probably have done it many times without noticing. I don’t even like using AnkiDroid because of this, I prefer to use desktop client to ensure I’m not making mistakes. I am asking for a setting to remove hard and easy because they are bad buttons for me. Simple as
@sorata thanks, just saw your post that’s great.
Just turned it on to check it out and it looks GREAT! Thanks Sorata. Are there any drawbacks to using new reviewer mode I should know of before I turn this on? Thanks in advance.
–
Regarding 2 vs 4 buttons I made up my mind long ago. I’ve been using anki for almost 10 years now and fortunately received good advice to only hit again or good when I first started and always kept that simply for mental fatigue reasons so I don’t waste time thinking about what button to hit. The fact that all these years later, it might also reduce RMSE is a just a cherry on top.
As I said, whiteboard and gestures aren’t yet implemented. They’re in the pipeline. I also think navigation isn’t good because until the day we have bottom tabs, there is no way to quickly switch to the browse or stats screen.
Other than that, the new reviewer is much better than the old one. If you have complex styling and rendering issues for some elements the new reviewer would solve quite a lot of that.
If it helps you trust, the previewer in AnkiDroid is actually based on similar code. So, it’s quite solid.
Amazing, thanks appreciate it I will be using this. I don’t use gestures or anything, I was just checking to make sure there is no changes to the algorithm or anything. Thanks again.