In a "perfectly normal" year half of soil would be drier than normal, half would be wetter than normal, and none would be exactly normal.
The most reasonable way to interpret the headline would be to say soil is slightly wetter than normal, which is obviously not the intended message.
Though given how much glsl I write, I appreciate that
So I think this example fails, at least. The original article is trying to make the point that Europe's soil is alarmingly, abnormally dry.
[1]: http://edo.jrc.ec.europa.eu/documents/factsheets/factsheet_s...
Even if we are looking at the mean average, it's still not necessarily the case that half the soil will be on one side of the mean average; long tails tugs that mean average above and below the median.
But if what's meant here is the median, then yes, half will always be below.
Excuse me now while I go and beat myself up in the car park for being that person on HN.
The point I was trying to make though is that "normal" is not in any way equivalent to any kind of average. It always refers to a range or distribution, even if that range is very small (eg if there is only one normal value, the normal range is the margin of error). 5'8" and 5'10" are both normal heights; you would not say they are taller or shorter than normal. They're shorter/taller than average.
Yao Ming is taller than normal. Dwarfism is by definition abnormally small size. Normal can't be the average, because there wouldn't be anything to describe the range of heights between abnormally short and normal height. Also, normally short and normally tall would be the same thing. Makes no sense.
> Any distribution can define a normal range relative to that distribution.
Any particular sample will always have a corresponding probability in a normal distribution. So what you're saying is kind of right- values always fall within the range of a normal dostribution. That's not what it means to be outside the normal distribution, though. If you have another year that only partly overlaps with the expected distribution, you'll have an area that does not overlap despite being within the same range. That area is the fraction of values outside the normal range.
Log-normal is extremely common in hydrology. Turns out an anomaly in this case is defined as outside one standard deviation[1], so in a perfectly normal year you would expect ~15.9% of soil to be dryer than normal.
[1]: http://edo.jrc.ec.europa.eu/documents/factsheets/factsheet_s...
This way of thinking about language is not just logical and economical but also conforming with a friendly approach on communication. Of course, this is something unattractive to many people who celebrate their incredible and superior intelligence for the purpose of grandstanding. And I fear this attitude is relatively more normal on HN.