The mathematics you actually use

School mathematics is arranged around techniques, and most people leave with a set of procedures they no longer remember and a conviction that they are bad at the subject. Meanwhile the mathematics that would genuinely help in ordinary life is a fairly small collection of ideas, several of which are barely taught at all. This is an attempt at that useful core.

Percentages, and the direction problem

A percentage is just a fraction with the denominator fixed at a hundred, which makes different fractions comparable. The trouble is almost never the arithmetic. It is that percentages are relative to something, and the something changes.

A price cut by 20 per cent and then raised by 20 per cent does not return to where it started. Take 100, remove a fifth to reach 80, then add a fifth of 80, which is 16, and you land at 96. The two twenty per cents were percentages of different numbers. This asymmetry is behind a great deal of misleading presentation, and noticing it is most of the defence.

Related is the distinction between percentage points and per cent. If an interest rate moves from 2 per cent to 3 per cent, that is a rise of one percentage point and an increase of fifty per cent. Both descriptions are accurate; they invite completely different reactions, and which one gets used is rarely accidental.

Big numbers need anchoring

Human intuition handles numbers up to a few hundred reasonably and then degrades badly. A million, a billion and a trillion feel like a sequence of similar-sized large things. They are not remotely.

The standard illustration is time. A million seconds is about eleven and a half days. A billion seconds is nearly thirty-two years. A trillion seconds is roughly thirty-one thousand seven hundred years — comfortably longer than recorded history. Nothing in ordinary experience prepares you for a gap that size, which is exactly why figures at this scale are so easy to present misleadingly.

The habit worth building is converting to something concrete: per person, per day, or as a proportion of something known. A national figure divided by the population is almost always more informative than the total.

Averages hide as much as they reveal

The word average usually means the mean — add everything up, divide by the count — and the mean is fragile. A single extreme value drags it a long way, which is why average income in a region can rise while most people there are no better off.

The median, the middle value when everything is sorted, ignores how extreme the extremes are and is far more representative for skewed data. Income, house prices and wealth are all strongly skewed, so median is nearly always the more honest measure, and mean is nearly always the more flattering one.

The mode, the most common value, is the third and is genuinely useful for things that come in discrete categories. When a report cites an average without saying which, that is worth noticing.

Compounding is the one that surprises everyone

Growth applied to a total that already includes previous growth accelerates in a way that intuition systematically underestimates, because we extrapolate in straight lines.

A useful shortcut is the rule of 72: divide 72 by a percentage growth rate to get the approximate doubling time. At 6 per cent, something doubles in about twelve years. At 3 per cent it takes twenty-four. Halving the rate does not slow things a little; it doubles the wait.

This applies to savings, debt, populations and anything else growing proportionally to its size. It is also why small differences in rate matter enormously over long periods, and why the same mechanism makes compound interest a powerful ally and high-interest debt a serious problem.

Probability, and the base rate

The most consequential probabilistic mistake in everyday reasoning is ignoring how common something is to begin with.

Consider a test for a condition affecting one person in a thousand. The test is 99 per cent accurate. You test positive. The intuitive reading is that you almost certainly have it. The correct reading is that you probably do not. Out of ten thousand people, ten have the condition and the test catches essentially all of them. Of the nine thousand nine hundred and ninety who do not, one per cent — about a hundred people — test positive anyway. So roughly a hundred and ten positives, of whom ten are genuine. Under ten per cent.

Nothing here is advanced. It requires only remembering that a rare condition means most positives come from the much larger healthy group. The same structure underlies a great deal of confused reasoning about screening, security alerts and rare events generally.

Orders of magnitude beat precision

The most transferable skill in this whole area is estimating roughly and checking whether an answer is plausible. Not calculating exactly — approximating well enough to catch nonsense.

Physicists call these Fermi problems, after Enrico Fermi, who was known for producing surprisingly good estimates from almost no data by breaking a question into parts he could each guess within a factor of a few. Errors in opposite directions tend to cancel, and the result is often right to within an order of magnitude.

Applied to daily life this is a defence mechanism. When a claim arrives — a cost, a statistic, a saving — a quick rough check catches the ones that cannot possibly be right. That single habit is worth more than most of the techniques that occupied years of the curriculum, and it needs no notation at all.