Nate Silver, a statistician and political forecaster at The New York Times, first achieved fame in 2008, when his forecast for the US presidential election proved more accurate than most mainstream polls. He repeated this feat again in 2012, when he successfully predicted the election results in all 50 states.
In “The Signal and the Noise”, Silver examines a variety of different scenarios in a quest to uncover the “Art and Science of Prediction”. The aim of the book is to discover “why some predictions succeed and why some fail.”
Separating the Signal from the Noise
Silver argues that the ever increasing amount of data available to us has not made it easier for forecasters to make accurate predictions. In fact it has become more difficult to distinguish between key (The Signal) and irrelevant (The Noise) information.
Searching for a solution to this issue, Silver turns to the writings of Thomas Bayes. This 18th century English minster theorised that we learn about the universe through approximation, getting closer and closer to the truth as we gather more evidence.
This approach flies in the face of the tactics employed by many economists, political and sports commentators who rely on statistical models. Their habit of making confident predictions based on limited or irrelevant data is one of Silver’s pet peeves.
Throughout the book, Silver takes aim at a number of forecasters from various sectors. In one passage he examines how economists have fared when it comes to predicting recessions. He uncovers a number of alarming facts, including that in the 1990’s, only two of the 60 recessions around the world were predicted a year ahead of time. Similar findings are made with regards to politics, baseball and earthquakes.
The issue, he explains, is not using statistical models, but rather relying solely on machines to do the “thinking”. Without human analysis and insight, coincidences such as a sports team winning an event and an economic crisis occurring at the same time, may find themselves aligned.
Playing with Probability
In one chapter, “The Poker Bubble”, he applies the Bayesian principles to the popular card game. Silver explains how the best players use a variety of factors to make estimates regarding their opponent’s chances.
For example, statistically speaking women are more likely to play conservatively. Although this may only be true in 55 per cent of instances it still provides an insight. As the game evolves, crude initial assumptions are refined as players update their probabilistic assessments based on the behaviour of their opponents. “If you doubt the practical uses of Bayes’s theorem, you have probably never witnessed a poker game.”
This approach is similar to that employed by Silver when he predicted the results of the 2012 election. The reasoning being that because we will never have 100 per cent of the data to be able to predict 100 per cent of the future, probabilistic inferences are as close as we can get.
A Humble Approach
“The Signal and the Noise: The Art and Science of Prediction” provides a fascinating insight into the world of forecasting. The sheer range of examples Silver provides, coupled with his undoubted knowledge of the topic, make for a very entertaining read.
His refreshingly humble writing style serves to highlight his belief that human beings need to exercise more caution when it comes to forecasting. By thinking in terms of probabilities and not certitudes, it is possible to make far more accurate predictions, a key element for any business looking to thrive in a volatile economic environment.