# How far out can weathermen accurately predict the weather?

This is all about probabilities. How long can the weather be predicted with 100% accuracy? That is not a 1-in-a-1,000,000 chance of being wrong? Then the answer is not even one minute.

Alternately, I can tell you right now that there’s a very, very, very good chance there will be rain somewhere in the world, on this day, 100 years from now.

The trade-off between certainty and accuracy is something weathermen struggle with every day. Every bit further into the future, the prediction certainty tends to go down just a little bit.

There are a few question people ask weathermen on a regular basis:

1) Is it going to rain today? (People want to know whether to take a coat.)

2) Is it going to rain this weekend? (People want to know whether to plan a picnic.)

3) Will it be exceptionally hot in the next few days? (People want to know whether to stock up on ice and drinks.)

4) Is the wind going to be strong in the next few days on the ocean? (People want to know whether it will be safe for small boats.)

5) When’s the first major snow going to fall at the ski resort? (The owners and tourists want to schedule their holidays.)

In none of these cases can the weatherman be sure. There’s a misconception that the weatherman has failed when a prediction goes wrong. In fact, there’s always a chance of failure. But over the last decades the accuracy of weather prediction has increased very significantly.

Notice that the common questions all have a relatively specific purpose. Often an economic purpose. I.e., if fishing boats can’t be sent out because of high winds, fishermen will lose money. So our society invests money to learn that information. But if the wind is unusually high in an uninhabited mountain range? Few people will care, little or no money will be lost, whether the wind is high or low. So our society invests almost no money in that!

If we invested 100 times as much, would our weather predictions be 100 times better? Not at all. There’s a point of diminishing returns where the cost is high, and the economic advantage is low. There’s also the “Butterfly Effect”, which means that small, random or unforeseen events — such as a butterfly flapping its wings — might eventually make the difference between it raining one day, or not.