Conclusion Summary and Analysis

Nate Silver encourages forecasters and analysts to follow the Bayesian theorem and to think probabilistically. The process is not rigid and allows for the forecaster to reassess his forecast when confronted with new data. A fundamental requirement of Bayesian predictions is that the forecaster must have a valid starting point, a prior belief built on past experience and historical data. "Prediction is difficult for us for the same reason that it is so important: it is where the objective and subjective intersect." (453) A successful forecaster must learn to distinguish what is important in his data - the signal - and what is basically meaningless data - the noise. Above all, good predictions can only be made if we aren't under the delusion that they are better than they are.