Thursday, April 21, 2005

It will rain...umm, maybe

The met office predicts that the rains will be normal this year. How do they know? The clouds come by only around the middle of June. Though what I'd really like to see is met observations that can't go wrong: "Met says sun will shine tomorrow" screams the Times of India.

And then, just for kicks, we could have a day-long eclipse.

3 comments:

avinash said...

:) On balance, I'd believe them.
Incidentally, in The Wisdom of Crowds, Suroweicki mentions that one of the reasons we ought to turn to a "crowd" of people for an answer rather than to an individual expert, is that experts often overestimate themselves.
"Weathermen and bridge players are the only experts that don't overestimate themselves", he says. So if the met office contains weathermen (not always sure it does), then whatever predictions they make are *probably* right.

Rahul said...

I'll check out the book because what he says sounds logical. Though I wonder whether weathermen on tv are more circumspect with their predictions than the ones at the met office where, under anonymity, almost anything goes.

The headlines that announce "Monsoons due by June 2/6/8" are the best, because a few days later the data is corrected, almost always pushing the monsoon further away. Then follows a public debate about when it should rain, about droughts, and Mid Day will do an expose on weathermen.

When it doesn't rain, it pours.

avinash said...

you know..considering how important the monsoon season is and how much rides on it, I wonder if the Met offices are under pressure (political?) to make "sound predictions".
Yeah the book is a v.interesting read (it basically argues that a crowd of people that satisfies some basic criteria - such as diversity of thought - will almost always do better than a single expert when it comes to solving diff. kinds of problems)