|
Sure, Pablo
|
# ? Jul 15, 2022 15:54 |
|
|
# ? May 30, 2024 13:06 |
|
also anything over a 500-year event is basically "off the charts, we dont know, there's no data, that poo poo shouldn't have happened" it's very much a sign that your distribution, as SirPablo said, is non-stationary and your old interval estimates are hosed
|
# ? Jul 15, 2022 15:58 |
|
gently caress You And Diebold posted:I hoped "a thousand weathers" would have made it clear I was joking Infinite weathers are happening simultaneously
|
# ? Jul 15, 2022 16:02 |
|
Let a thousand weathers doom?
|
# ? Jul 15, 2022 16:03 |
|
a thousand and one weatherly events
|
# ? Jul 15, 2022 16:05 |
|
Grundulum posted:Isn’t that exactly how that works? *IF* you assume that all weather stations are independent, would you not expect one in a thousand to have a 1000-year event every year? I would like to offer an apology; I took your bolding and asterisks to be an excessive extravagance that undermined the joke and the evidence tells me I was wrong.
|
# ? Jul 15, 2022 16:07 |
|
GFS predicting the day after tomorrow
|
# ? Jul 15, 2022 16:08 |
|
Seville is 106 right now after being 108 yesterday. https://www.fastcompany.com/90763832/seville-is-the-first-city-to-name-and-categorize-heatwaves-like-hurricanes Is the picture saying it was 116 on June 13th accurate? I don't remember hearing anything about that.
|
# ? Jul 15, 2022 16:09 |
|
I've gotten better at converting Fahrenheit to Celsius, but I kept looking at 40C and thinking "nah, I'm messing this up. No way it's going to be a hundred loving degrees over there." So I checked
|
# ? Jul 15, 2022 16:31 |
|
SirPablo posted:Nah just trying to help people understand that the most doomer outcome doesn't always happen. So many models are run so frequently, you can always find a doomer scenario somewhere. It's all about probability - 40C was always possible but initially at a <5% chance, now up to 40% (from post above). It's been clear for a long time you don't know numbers. The probability doesn't change. That is not how probability works. You innumerate gently caress.
|
# ? Jul 15, 2022 16:41 |
|
SirPablo posted:Infinite weathers are happening simultaneously You don't need a Weatherman to know which way the wind blows
|
# ? Jul 15, 2022 16:42 |
|
Looking at the models makes you go "welp gently caress, we're in trouble" because the models output predictions well out of historical range. Our models aren't perfect or reliable, though repeated cases of "that can't happen" right into "it happened" does make me question all the models that don't have the occasional really out there results, because now those models are incapable of accurately predicting reality. And thus are useless except for people who need "scientific" excuses to stick their heads in the sand. As the date draws closer, the probability doesn't change. Our ability to predict the outcome improves. And the outcome is "that's a lot of dead people".
|
# ? Jul 15, 2022 16:48 |
|
What? The probability changes against the model's latest inputs. The probability of the weather event occurring doesn't change but you're presuming perfect modeling for some reason?
|
# ? Jul 15, 2022 16:52 |
|
SirPablo posted:Ok maybe goons don't know numbers. So if something that only supposed to happen 1% of the time ends up happening every year or even multiple times in a year is only 1%
|
# ? Jul 15, 2022 16:52 |
|
NeonPunk posted:So if something that only supposed to happen 1% of the time ends up happening every year or even multiple times in a year is only 1% That's correct. Think about it as, the probability of getting heads is 1 in 2. But there isn't actually anything stopping an outcome of heads, heads, heads, heads, heads... except extreme improbability. Perry Mason Jar has issued a correction as of 16:57 on Jul 15, 2022 |
# ? Jul 15, 2022 16:54 |
|
Perry Mason Jar posted:That's correct. Think about as, the probability of getting heads is 1 in 2. But there isn't actually anything stopping an outcome of heads, heads, heads, heads, heads... except extreme improbability. Really? Say, wanna bet on a coin toss then? (No, you may not check to see if it's a trick coin.)
|
# ? Jul 15, 2022 16:57 |
|
Perry Mason Jar posted:What? The probability changes against the model's latest inputs. The probability of the weather event occurring doesn't change but you're presuming perfect modeling for some reason? I'm saying our modeling is hosed because it doesn't have historical data to draw on anymore, because the circumstances have changed. And those probability numbers don't - can't - account for changes we don't understand. SirPablo's stuck pining for a world that isn't coming back.
|
# ? Jul 15, 2022 17:00 |
|
On a fair coin the odds don't change so yes. That's exactly the point. If you roll heads, heads, heads, heads, the odds that your next flip lands on tails is 50%. The probability of landing on tails does not increase because of the four prior heads rolls.
|
# ? Jul 15, 2022 17:05 |
|
endlessmonotony posted:I'm saying our modeling is hosed because it doesn't have historical data to draw on anymore, because the circumstances have changed. And those probability numbers don't - can't - account for changes we don't understand. SirPablo's stuck pining for a world that isn't coming back. I'm operating under the assumption that Pablo's aware of the modeling crisis that's been happening since, er, 2019?
|
# ? Jul 15, 2022 17:06 |
|
Perry Mason Jar posted:I'm operating under the assumption that Pablo's aware of the modeling crisis that's been happening since, er, 2019? It's pretty clear they're choosing to ignore it at every turn.
|
# ? Jul 15, 2022 17:06 |
|
I thought I understood numbers but every post I read in the past page is making my brain hurt That might just be the beginning glimmers of heatstroke though
|
# ? Jul 15, 2022 17:19 |
|
endlessmonotony posted:I'm saying our modeling is hosed because it doesn't have historical data to draw on anymore, because the circumstances have changed. And those probability numbers don't - can't - account for changes we don't understand. SirPablo's stuck pining for a world that isn't coming back. These are dynamic models. That means they use current weather data and fundamental physics to predict the weather. You clearly don't understand that circumstances change and thus the probability of events occurring changes. What's the probability you get in a car accident today? We'll, if your plans right now are to sit in your house and go no where, then almost zero (I suppose a car could drive into your house). Now what if a friend calls and asks you to help them with something? Wow, the circumstances have changed and you'll be getting in your car, now the probability of a car accident has increased!
|
# ? Jul 15, 2022 17:19 |
|
SirPablo posted:These are dynamic models. That means they use current weather data and fundamental physics to predict the weather. Well, QED. Amazing. Your models aren't reality you idiot mathfucker, they predict reality, except that most of those models are now incapable of predicting reality because they were built for circumstances that no longer exist.
|
# ? Jul 15, 2022 17:23 |
|
SirPablo posted:These are dynamic models. That means they use current weather data and fundamental physics to predict the weather. That's a perfect analogy since the risk of getting into a car accident has vastly increased in the past 2 years for *mysterious* reasons. You wouldn't say that we had the same probability of getting into a car accident three years ago today right?
|
# ? Jul 15, 2022 17:25 |
endlessmonotony posted:Well, QED. Amazing. climate is not weather. model forecasts arent dependent on climate, they're dependent on how we know weather works. why are you being such an rear end in a top hat
|
|
# ? Jul 15, 2022 17:25 |
you guys are arguing over a philosophical question models aren't reality, but probabilities in reality do also change as events unfold, unless you think the universe is deterministic
|
|
# ? Jul 15, 2022 17:26 |
|
endlessmonotony posted:Well, QED. Amazing. jesus man what did the weather models do to you?
|
# ? Jul 15, 2022 17:27 |
|
Lol I almost mentioned Laplace's demon when I said "you're presuming perfect modeling for some reason" Because the demon can't exist in our universe as we know it, even in a "billiard ball" universe
|
# ? Jul 15, 2022 17:29 |
|
endlessmonotony posted:Well, QED. Amazing. Idg why you're being obtuse. There are many dynamic weather models where the underlying physics are being updated as understanding proves and the science evolves. Here, you can go run this one yourself even. https://www.mmm.ucar.edu/weather-research-and-forecasting-model
|
# ? Jul 15, 2022 17:30 |
|
new thread title: Weatherdammerung 2022: told you so or Weatherdammerung 2022: standard deviation not enough for pervert forecaster
|
# ? Jul 15, 2022 17:31 |
|
NeonPunk posted:That's a perfect analogy since the risk of getting into a car accident has vastly increased in the past 2 years for *mysterious* reasons. You wouldn't say that we had the same probability of getting into a car accident three years ago today right? Right. Circumstances change. Unless you're a time traveler and already know the future.
|
# ? Jul 15, 2022 17:33 |
|
when someone gets called obtuse its probably time we all just grab a beer and talk about how hot it is
|
# ? Jul 15, 2022 17:34 |
|
Notorious R.I.M. posted:jesus man what did the weather models do to you? Oh, no, this goes deeper into personal grudges with the math underlying this issue. Literally the hill I died on.
|
# ? Jul 15, 2022 17:42 |
|
endlessmonotony posted:Oh, no, this goes deeper into personal grudges with the math underlying this issue. You understand close to nothing about how weather modeling works though
|
# ? Jul 15, 2022 17:45 |
neutral milf hotel posted:Weatherdammerung 2022: standard deviation not enough for pervert forecaster lol.
|
|
# ? Jul 15, 2022 17:48 |
|
neutral milf hotel posted:Weatherdammerung 2022: standard deviation not enough for pervert forecaster Lol
|
# ? Jul 15, 2022 17:50 |
neutral milf hotel posted:Weatherdammerung 2022: standard deviation not enough for pervert forecaster
|
|
# ? Jul 15, 2022 17:54 |
|
Notorious R.I.M. posted:You understand close to nothing about how weather modeling works though I understand it fine, and Perry Mason Jar is proof I'm communicating my point too, but in a hilariously inefficient way. Well, hilarious to me anyway. I know, I know, I'm fundamentally also an idiot mathfucker, and this is just me ranting about the interface problem between the math and how it's used.
|
# ? Jul 15, 2022 17:57 |
|
neutral milf hotel posted:new thread title: pretty easy choice here IMO
|
# ? Jul 15, 2022 17:58 |
|
|
# ? May 30, 2024 13:06 |
|
starting to doubt these "meteorologists" i've yet to see a single meteor in any of these forcasts
|
# ? Jul 15, 2022 18:03 |