Unlike earlier thinkers, who had sought to improve their accuracy by getting rid of error, Laplace realized that you should try to get more error: aggregate enough flawed data, and you get a glimpse of the truth. "The genius of statistics, as Laplace defined it, was that it did not ignore errors; it quantified them," the writer Louis Menand observed. "...The right answer is, in a sense, a function of the mistakes.

Error ... is less an intellectual problem than an existential one - a crisis not in what we know, but in who we are. We hear something of that identity crisis in the questions we ask ourselves in the aftemath of error: What was I thinking? How could I have done that?

Unlike earlier thinkers, who had sought to improve their accuracy by getting rid of error, Laplace realized that you should try to get more error: aggregate enough flawed data, and you get a glimpse of the truth. "The genius of statistics, as Laplace defined it, was that it did not ignore errors; it quantified them," the writer Louis Menand observed. "...The right answer is, in a sense, a function of the mistakes.

If it is sweet to be right, then - let's not deny it - it is downright savory to point out that someone else is wrong.

The kinds of things that we can make mistakes about are essentially unlimited in number.

Regret doesn't remind us that we did badly, it reminds us that we know that we could do better.

doubt is a skill. credulity ,by contrast, appears to be something very like an instinct