The Signal and The Noise

The Signal and the Noise: Why So Many Predictions Fail—But Some Don't by Nate Silver

My rating: 5 of 5 stars


*** Possible Spoilers ***

'Prediction is difficult' is the message of this book. Whether it's gambling or politics, determining the future isn't easy. Part of the problem is the data you have to work with. Is it relevant - signal - or irrelevant - noise. As the internet expands, and data is measured in petabytes - or beyond - determining its relevancy is becoming increasingly difficult. Unfortunately the author fails to provide any concrete advice for solving the problem but at least he delineates it nicely. The thing to rememeber is correlation does not imply causation although it MIGHT or it MAY signify some overlooked factor related to both. The author also suggests outliers are noise as opposed to signal and this is true in most instances; however, they should not be disregarded. There is always a reason for an outlier. It may, and frequently is, a mistake in measurement. But sometimes it indicates some unknown factor of considerable importance. As a former colleague once remarked to me, 'pay attention to the outliers. That's where the research papers live.'



View all my reviews