Tuesday, December 6, 2016

The Art of Prediction

Superforecasting: The Art and Science of Prediction – Philip E. Tetlock and Dan Gardner (Broadway Books)

If the 2016 presidential election proved anything it proved that the game of prediction is hard…very hard.

Liberal polling and predicting guru Nate Silver and his 538.com website often hailed as the go to source had Hillary Clinton’s chance of winning the presidency pegged at 71% to Donald Tump’s 28%. Mic.com stated that “forecasters predict solid Clinton victory. Going against the whitewater rush of predictions going in Hilary Clinton’s direction was the professor who utilized a primary turnout model to factor his presidential pick, Donald Trump. Of course the professor faced an equal onslaught of criticism from those who ready to anoint Clinton.

So what was the root cause of this MASSIVE failure on the part of forecasters? The answer is a simple one. In Superforecasting: The Art and Science of Prediction, authors Philip E. Tetlock and Dan Gardner offer prescient insight with their straight forward and handy Ten Commandments of Superforecasting, which turns out to be 11 commandments long. The simple answer comes in the form of number 2, “unpack problems to expose assumptions, catch mistakes and correct for biases.”

BINGO! It was the assumptions and biases that tripped up all of these so-called experts when it came to the presidential election. I found Tetlock and Gardner’s  Ten Commandments to be the most useful part of the book, which I think focused much more on the forecasters than it did on forecasting.

No comments: