'Once in 300 years' rain hits Thai city as floods ravage South East Asia(www.bbc.com)

24 pointsby Brajeshwara day ago3 comments

fghorowa day ago
'Once in 300 years'???
While the functional form of the statistical distributions themselves might still be valid, certainly the old parameters are no longer so.
- 171862744019 hours ago
  That is a severity scale using anecdotal past events. It isn't intended to be a sound statistical claim.
  - adammarples15 hours ago
    It's not even that, it's just the heaviest rainfall recorded in 300 years
- j16sdiz21 hours ago
  Nobody knows the future. Most of the time we just use historical data directly, or project linearly.
  These kind of headline is very misleading. Need better way to communicate these.
  - 19 hours ago
    undefined
andrewstuarta day ago
The entire thesis of once in 300 years or once in 1000 years or once in 100 years weather event really has to be let go by the media. The fact is these things are happening or more and more often and are a direct result of climate change and then not once in 300 years, they’re happening all the time
- georgefrownya day ago
  It's also a bad statistical method because if there are 300 cities/regions in the world and a storm hits a random one of them most severely each time, on average you will expect to have one city/region every year seeing a 300-year storm even in a static climate.
  Not that I think the climate isn't changing, but because if the headlines are obviously p-hacking all the time you get all climate change reporting eventually called fake news even when it isn't.
  - 171862744019 hours ago
    I think these terms state how likely some event is for some climate, which is useful for people who don't live in that climate. It isn't so much used for real statistic.
- griffzhowla day ago
  The article says it's the heaviest rainfall recorded in Hat Yai over the last 300 years. So that's the actual meaning, and interpreting it in the probabilistic sense seems to have been the initiative of the headline writer.
- 171862744019 hours ago
  This is a severity scale primed with how much likely it was in the past. We might adjust that scale in a century, but the events severity don't change and it would be useless to continuously adjust a scale, while trying to use it, that would make it meaningless.
- roughlya day ago
  I mean, we’re still working on convincing people that climate change is actually happening, so if they want to keep reporting the 100yr storms that happen every year now, that’s fine by me.
  - ragebola day ago
    They should be clearer: "storms that were once in a 100 years in the 'old' climate".
    But how do you fit in nuance and statistics into news headlines etc?
    grebc21 hours ago
    English has good words for severity which is what they’re trying to impart.
    potato373284219 hours ago
    There is a huge difference between "once per location per 100yr" and "once per 100yr".
    Every year there is at least one hurricane Katrina equivalent storm in the world. Having one in New Orleans is once in hundreds of years. Anywhere on the gulf coast is once in, IDK, a dozen.
    So you can pretty easily lie and mislead (accidentally or not the results are the same) by not being super careful about scope.
    171862744019 hours ago
    This is a well-known term, that categorizes the norm in a specific location. It would be useless to use it across locations.
    ragebol19 hours ago
    Makes all the difference indeed.
    No place nor patience for such nuance nor precise definitions in the news unfortunately.
    potato373284218 hours ago
    >No place nor patience for such nuance nor precise definitions in the news unfortunately.
    You might even say there's strong incentives in the opposite direction.
- senectus1a day ago
  yeah they should measure rainfall in swimming pools, or sydney harbours.
emacsahmeda day ago
[flagged]