Predictive Dilemmas: Accuracy vs. Certainty

What's for dinner tonight?

"50% chance it's a hamburger, 50% chance it's fried chicken.

Which is it? I know, right?

If someone says, "There is a 60% chance it will be a hamburger steak and a 40% chance it will be fried," you might think, "Well, I guess I'll just have a hamburger steak.

In order to avoid this kind of "You have to be clear about which is which! In the world of forecasting, it seems that people sometimes avoid 5-minute forecasts in order to avoid "Which is it? For example, in the U.S., private companies rarely say that there is a 50% chance of precipitation in their weather forecasts. This is because it looks indecisive in the eyes of the viewer. So, even at the expense of some accuracy, they round up to the nearest 60% or down to the nearest 40%.

There is also a story about Nate Silver, author of "Signal & Noise," who spoke to a gathering of business executives and others before the 2012 U.S. presidential election. When he predicted that Obama and Romney each had a 50% chance of winning, the people listening to his talk said, "Cut the BS and tell us the truth! They said, "Stop giving us one-sided predictions and tell us the truth!

Which do you prefer, honesty or catchiness?

References

• Signal & Noise: The "Predictive Science" of a Genius Data Analyst