search

Treeki / MeteoNook

weather prediction tool for Animal Crossing: New Horizons
GNU Affero General Public License v3.0
279 stars 59 forks source link

Incorrect shower time on overview #8

Closed Swpolo closed 4 years ago

Swpolo commented 4 years ago

The time provided on the overview for the star shower of the current day doesn't match the time provided in the day forecast.

I believe the overcast page still use the old system from 1.2 What leads me to think that is that at midnight, which has a fine weather, the time is correct. But at 1, with a cloudy sky, the time is incorrect

Example in following image for seed: 1443633326 Screenshot_20200711-010820 Screenshot_20200711-010814

Treeki commented 4 years ago

This is not a bug at all. The overview tries to give a representative sample of all the stars that appear by grouping stars that appear closely and giving an average of the time.

As the in-app FAQ says:

During a light shower, you have a small random chance of seeing five shooting stars every minute between 19:00 and 04:00. MeteoNook analyses the schedule to try and figure out when multiple groups of stars appear close to each other. The times in bold are when you can expect to see quite a few stars. On the other hand, the times that are faded out are when stars are few and far between.

This lets you see at a glance roughly when to play if you want to get more. Of course, you can always click 'see more' for exact times.

Swpolo commented 4 years ago

ah indeed, I was too focused on the prediction between midnight and 1 AM that I didn't challenged my assumption on the 2 AM prediction which was having a fine weather but still was rounding the stars time occurrence.

I'll be more careful next time, thanks for your answer

RickSeiden commented 4 years ago

Given the fact that we are supposed to track our shooting stars down to the second when entering them, averaging the times sends a very mixed signal. It would be much better to see "7:07-7:15" in the summary than "7:10".

zeroum commented 4 years ago

This is not a bug at all. The overview tries to give a representative sample of all the stars that appear by grouping stars that appear closely and giving an average of the time.

I'm sorry but this behavior is very misleading, even if it's written in the FAQ.

The app idea is about finding exact seeds, knowing the exact weather, and all its exact patterns. Averaging the times goes against its own exactness ethos.

I also support the range idea, something like 19:07—19:15 would be a better option if the times are too close together, and there are more than 3. Maybe even out a (N) to show how many of them, like 19:07—19:15(4).

gjnance commented 4 years ago

Second @zeroum 's comment. It is misleading to show a precise time when that precise time is not indicative of a precise event. I think something like "Expect star events between 7 and 7:30" for n>=1 or some such.

liangaili commented 3 years ago

This is not a bug at all. The overview tries to give a representative sample of all the stars that appear by grouping stars that appear closely and giving an average of the time.

I'm sorry but this behavior is very misleading, even if it's written in the FAQ.

The app idea is about finding exact seeds, knowing the exact weather, and all its exact patterns. Averaging the times goes against its own exactness ethos.

I also support the range idea, something like 19:07—19:15 would be a better option if the times are too close together, and there are more than 3. Maybe even out a (N) to show how many of them, like 19:07—19:15(4).

I respect the designer's idea, and I agree with zeroum's opinion that this design might be misleading for some users.