News for hackers. Stuff that matters.
Re:Why the snow forecast for New York City was so bad by saalweachter on Tuesday January 27, 2015 @01:36PM (#8954791)
A single 9x% confidence interval like 2-16 inches gives the impression of a continuous bell-curve distribution, when you might have a multiple peaks based on a couple of simple factors. I wonder if it would be useful to give out a couple of probability buckets. If people heard "20% chance of 10 or more inches, 60% chance of less than 2 inches", would then know to be ready for a lot of snow, but not be surprised if it doesn't happen, or would they ignore the considerable chance of a heavy snowfall?

There are no comments on this comment yet.