Weather forecasting

Have a new general feature to suggest, or think one should be tweaked? Share your ideas here.
Post Reply
User avatar
nobody
Posts: 501
Joined: Wed Jun 17, 2020 7:43 am
Contact:

Weather forecasting

Post by nobody »

I updated the wiki to add some info about weather (it's very basic, but probably sufficient) and that had me thinking about why the real world weather locations aren't revealed. It would be anachronistic for characters to just know what the weather will be like as if they all had modern forecast data in their heads is ultimately where I landed. However, with this being problematic for farmers, I was wondering if it wouldn't be nice to add a weather forecasting service to the Twilight Eye offices (as they seem to be the best source of divination-like magic) that just pulls weather forecast data for Shadgard farmlands (and eventually Not-Shadgard farmlands). Characters could go pay for a weather forecast (for however far out is deemed appropriate) for some amount of riln, and then plant crops based on that. You could even provide a 7-10 day forecast in a rather vague fashion. For example, pull the normal forecast data, convert it to in-game temperature classifications (freezing/cold/chilly/mild/warm/hot/sweltering) and then aggregate it for maximum vagueness: "over the next several days, it is expected to reach down to <lowest temperature classification in forecast> temperatures for <number of times that shows up in the lows> days and up to <highest temperature classification in forecast> temperatures for <number of times that shows up in the highs> days (e.g. over the next several days, it is expected to reach down to cold temperatures for 3 days and up to mild temperatures for 7 days). If you want less certainty, add noise to the data (but weather data is inaccurate enough that probably isn't necessary) and/or randomize the number of days you pull into your aggregation (if 10 days are available, roll up a number between 5 and 10 and then use the first however many days of the weather data for each day's weather forecast so then not even the length of the weather data would be predictable and it would be difficult to tell how the weather data is changing even if you paid to get a forecast every day).

And last, if you do opt to include the weather forecast, for optimization I'd suggest having it be a side run process like market searches are now, only run it at most once per day, and run it when the first person pays for a weather forecast that day and save just the highly aggregated result. So it needn't be run every day if there isn't anyone paying for it, and everyone who pays for a forecast on the same day will get the same information.
Post Reply