Reviews graph - units of total time studied suggestion

Hello.

My suggestion is to display total study duration on the stats board in units of hours.

Currently on the Linux client’s stats, total time studied displays amounts in what I could at best describe as calendar units; once it passes 24 hours it goes into days, (maybe weeks), months. It seems to me unnatural to count absolute durations in units greater than hours. When somebody says that they had worked somewhere ‘for a month’, they don’t mean ‘for 30.5 x 24 hours’ but ‘for ~4 x 5 x 8 hours.’
I wouldn’t necessarily say that a total review of ‘2500 hours’ is more graspable than ‘3 months’, but I’d say it’s more fitting, as well as more exact (how many days are in a month?).

Cheers and thank you!
Mark

5 Likes

makes sense. people talk about learning something for 10,000 hours, not 13 months.

4 Likes

Discussed previously –

3 Likes

The previous discussion wasn’t able to convince me that it’s a good idea as the talk was mostly about precision which I don’t think is important.

But, it now appears to me that changing the units make comparisons harder. Is it more obvious here that 29d is half of 1.9m? Or, is it more obvious that 2102h is half of 4204h?

OP also says something that I feel is correct:

It’s not something we see very often, at least not me. I see “Your flight is in 48h” not “Your flight is in 2d”. The latter can even sound bizarre.

4 Likes

Thank you, when I searched for previous threads I couldn’t find any.

1 Like

I’m not sure I follow this as an example. I think the first sounds like an odd thing to say about a flight that is 2d away – while the second sounds completely normal.

No problem! That thread popped into my mind because of the attitude and trouble-making of that particular user. I’m happy to leave that thread behind and see this suggestion getting the more reasonable and measured discussion it deserves!

1 Like

I’m not sure I follow this as an example. I think the first sounds like an odd thing to say about a flight that is 2d away – while the second sounds completely normal.

I think the stress is on ‘see’. When you talk about your flight to others, you’d say ‘my flight is in two days’, but if there’s an indicator online for a flight, it would more likely say ‘in X hours’. Nevertheless, I think it’s not quite an analogous example, since here it’s about duration and not accumulated time.

(couldn’t figure out how to cite a post, retaining the meta info about the original author)

2 Likes

(discourse shows you a quote button.)


Yes, that’s a slightly different case.

On a different note, I have found some people who think fractional months are an A11Y issue. It’s harder to understand.

In addition, Steam does this for their statistics for their games; in hours played.

Maybe it would be a nice ego boost to see that Anki Study times overtaking Steam hours played, etc.

2 Likes

I commented on this issue in this recent reddit post.

It would’ve been easier to see if Anki displayed “4.18 days (100.32 hours)” instead of me having to bring up the calculator.

2 Likes

I would like to give my arguments in favor of displaying time studied in hours:

  1. Anki already shows % of days studied in parentheses, which isn’t any different from my suggestion above.
  2. We don’t think in terms of studying for days or months, we study for at most a few hours at a time. Hours is the natural unit for studying.
  3. It’s more rewarding to see that you’ve studied a total of 10,000 hours than 416.7 days 13.9 months.

Edit: Or would it display 1.1 years? OK scratch that example, 1000 hours = 41.7 days is more realistic.

4 Likes

I would personally prefer not to have the decimal places like that.

I’d prefer something like “4 days, 4 hours, 32 minutes”. That’s what the “4.18 days” actually means, but “4.18 days” isn’t obvious at all for me, unlike the version I just proposed.

I also don’t think showing it as hours only is intuitive either. I mean: I studied 1435.94 hours this month – great but, what does that mean? To me it’s more difficult to grasp then having it converted to “X days, Y hours, Z minutes”.

Maybe it could be made more obvious that that graph is showing full days, (i.e. that 24h means 24h studied in total, accumulated)?

I kinda agree. Not the main point though.

Maybe that’s too much info.

That’s what the “Average for days studied” is for. The “1435 hours” is just a benchmark. It’s a point of reference that people often use in discussions and a motivation to keep going.

4 Likes

The arguments have swayed my opinion - limiting durations to hours does seem reasonable. I suggest you throw up a poll in a new topic, and if you can get the majority of people preferring the new method, someone can implement it.

5 Likes

I decided against pinging people and just sharing the poll here:

I opted to raise it here over opening a new thread, as it is rather related.

I’d like to suggest that when displaying stats for the entire history (>1 year), the time axis would use dates (i.e. 27th Feb.-2nd Jun. '24) instead of delta time (i.e. 3,501-3,600 days ago) as it currently does.
I think that when one inspects longingly past achievements or shortcomings, as one does, it’s of greater interest to know when the peaks and valleys happened, rather than ‘how long ago in absolute duration’ it happened.

2 Likes

I guess dae’s reasoning was a new thread would get more of us to vote; not a lot of us would be following old threads.

I was referring to my comment, not the poll.

Oh, start a new thread I’d say.

1 Like