The official wiki states that for region and weather, when assigning weather to regional data, it is imperative that the chance of all weathers for that region add up to a total of EXACTLY 100. http://www.creationkit.com/Region#Weather_Tab.

This seems ridiculously impractical to me. CoT's weather mods do not obey this rule and are one of the most popular weather mods. Their mod's chance values add up to nearly 200!

I am almost inclined to believe that someone just edited the wiki without knowing why or what they were talking about, however there is no discussion tab on the regional page whatsoever to reference.

Does anyone know why this was ever deemed important, who says so, and what the consequences of a value greater than 100 might be?

I have never noticed any problems.

Open weather regions made by the devs and see if the total of their contents make 100. If they all do, there's probably a reason. And the popularity of a mod has nothing to do with how it's made; none of these aficionados knows about weather chance. The engine probably recalculate the chances so that they make 100.

EDIT: I miscounted. It seems all of the "vanilla" regions DO add up to 100...

?

?I am still not clear what happens if you go over that.

Do we know who added that line? If it was a dev or community member? Those pages tended to be pretty empty. I assume that number is actually a weight and not a chance, I.e. they add up to X then the real chance is the value over X, not value%.

Well, this is frustrating.

?

?Having them add up to more than 100 doesn't cause a crash, so now I am wondering what exactly the consequences might be, if any.

I suppose you could make a test case. Make a throwaway esp, butcher a cell so it only has two possible, highly visible weathers and then give them both a chance of 1

If the chances are in fact weights then you'll see a fifty-fifty chance for each of them. If you get no weather to speak of then you'll know it's rolling the dice on a d100 table. And if you get a CTD ... well that tells you something as well.

Hey Doc,

?

?A scientific approach is always best.

I ran a test in the SiegeTest world with two weathers assigned: SkyrimStormSnow and OvercastRain. First I tried setting them both to a chance of 1. I coc'd to the "siegestart" cell and it was snowing. Waiting several hours it started to rain. It continued to switch between the two assigned weathers after that.

?
I then tried again with both weathers set to a chance of 150. Same exact results. The engine must calculate a relative weight as was surmised. The need to meticulously tweak weather mods so all weathers add up to exactly 100 appears to be erroneous, thankfully.

EDIT: Something my limited test didn't cover was whether the weights have any effect on how long it takes weathers to change. It didn't seem to. In other words, a cumulative chance of 2 for 2 weathers seemed to take 8-ish hours to change, and a cumulative chance of 300 between the same two weathers seemed to take about the same time.

?