Our brains are conditioned by our evolutionary past and we often jump to conclusions. This was once a pretty good survival tactic. But we are also endowed with the capacity to learn. So we screw up -- but we also have some good ideas about how to screw up less. This is a crude way to sum up Nate Silver's message in The Signal and the Noise: why so many predictions fail -- but some don't.
It's a great read with many good example and stories (chess, politics, climate, earthquakes, poker, anti-terrorism, etc.). If you are one who uses markers to flag the good passages, your marker will soon run dry.
Silver wants us to be Bayesians. (I will never grasp how one cannot be.) You must start with good prior beliefs (specified as odds) and you must update these systematically. To do this, you have to be open, flexible -- and smart enough to known which of the many signals you encounter are worthy (signals and not noise). This is all solid advice, easy to grasp but less easy to follow through. Silver's book is really a good pep talk that we can do it.
"Distinguishing the signal from the noise requires both scientific knowledge and self-knowledge: the serenity to accept the things we cannot predict, the courage to predict the things we can, and the wisdom to know the difference." (p. 453).
But are you now ready to play poker? I suggest that readers read the last chapter first. Then read the book, then read the last chapter again. Then decide.