In subjects that do not allow precision, a rough answer that is right is worth more than an exact answer that is wrong.
Read wrote it in his textbook Logic: Deductive and Inductive (1898), in a chapter on the terms of ordinary language, warning that in reasoning upon some subjects it is a mistake to aim at an unattainable precision. The line is now usually credited to Keynes, often in the form roughly right rather than precisely wrong; quotation researchers first find it attached to him in an accountants' magazine in 1966, twenty years after his death.
In practice
- A council planning school places works from a population forecast with a wide range, rather than a single figure from a model whose assumptions it cannot check.
- A doctor tells a patient that recovery usually takes two to six weeks, rather than naming a date.
- Possible explanations
- One explanation is that precision without accuracy invites false confidence: a figure given to the decimal place looks measured whether or not it is. Another is decision value, since a roughly right estimate points towards the right action while an exact but mistaken one points confidently the wrong way.
- In its favour
- Consistent with forecasting research, where calibrated ranges outperform confident point estimates, and with the practice of reporting uncertainty in science and engineering.
- Against it
- A rival adage holds the opposite, that it is better to be definitely wrong than vaguely right, because a definite claim can be tested and corrected; the idea goes back to Bacon's remark that truth emerges more readily from error than from confusion. Read's maxim gives no test for when vagueness has become evasion.
- Standing
- Usually misattributed to Keynes. Compare All Models Are Wrong, which makes the same trade from the modeller's side.
- Subject
- Science & Epistemology
- Essay
- Rules for weighing claims