Quotes from David J. Hand

Unfortunately, we generally find it difficult to assess very small probabilities. We typically overestimate them (thinking the events more likely than they are) and underestimate very high probabilities.

Events don't actually occur as often as people predict they will. And this in turn is related to hindsight bias (the tendency to see past events as being more predictable than they were at the time), which I'll discuss shortly.

Chaos: when the present determines the future, but the approximate present does not approximately determine the future.

in real statistical analyses the computer takes over the tedium of arithmetic juggling.

Raw data, like raw potatoes, usually require cleaning before use. Ronald A. Thisted

It makes sense to predict likely future performance from past performance. Indeed, we often don't have much else to go on. Unfortunately, however, the past can be an uncertain guide to the future.

It's the very essence of science that its conclusions can change, that is, that its truths are not absolute. The intrinsic good sense of this is contained within the remark reportedly made by the eminent economist John Maynard Keynes, responding to the criticism that he had changed his position on monetary policy during the 1930s Depression: "When the facts change, I change my mind. What do you do, sir?

A dice maker agreed with Scarne: "Once in a blue moon you might find a single store die that classes as a perfect, but the chance of finding two of them in a box of 60 and the chance that they would both be sold to the same purchaser as a pair is so small you can forget about it. It never happened.

it is easy to lie with statistics, but easier to lie without them'.

When a particular measure is used as an indicator of the performance of a system, people may choose to target that measure, improving its value but at the cost of other aspects of the system.

This recognition – that real statistics is about exploring the unknown, not about tedious arithmetic manipulation – is central to an appreciation of the modern discipline.

Admittedly, some people experience synaesthesia, in which they do associate a particular colour or sensation with particular numbers.

When you do change the circumstances under which you collect the data—when you intervene—the data are said to be "experimental." Experimental data are particularly important, because they can give you information about the counterfactuals mentioned in chapter

If you change the way data are used, it is perhaps hardly surprising that behavior in collecting those data changes—a feedback phenomenon of the kind we examine in detail in chapter

So what is going on? For the men and women separately, the crew had a higher survival rate than the third-class passengers. But overall the crew had a lower survival rate than the third-class passengers. This is not a trick—the numbers are what they are.

All this means, first, is that it's necessary to be very clear about what question you are asking, and, second, that whether data are dark or not will depend on that question. Trite though it may sound, the data you need to collect, the analysis you will undertake, and the answer you will get depend on what you want to know.

The evolutionary usefulness of the ability to recognize patterns and then infer the causal relationships they represent is demonstrated by the fact that precisely the same development of "superstitions" occurs in animals.

Smyth's predictions of the date of the Second Coming, as with everyone else's predictions of this event, have proved inaccurate.

A grasp of the cause of the colors of the rainbow doesn't detract from its wonder.