Concept

Independence

The relation between two events when knowing that one happened says nothing about the chance of the other. It is what makes variances add, and it is a hypothesis that fails invisibly — no picture of a sample shows whether it holds.

Named by 22 essays across 4 fields — each of them below, with the objects they name alongside it.

A Galton board after 600 balls. 600 balls fall through 12 rows of pegs, each bouncing left or right at random, and pile up in a bell-shaped heap.

A bell curve assembled out of coin flips

Drop six hundred balls through a board of pegs, each bouncing left or right at random, and they pile up in a shape that can be predicted precisely. Nothing coordinated them.

probability · Central limit
120 needles on a lined floor. 120 needles dropped at random across evenly spaced lines; 83 of them cross a line.

Getting pi by dropping needles on the floor

Throw a needle at a lined floor enough times, count how often it crosses a line, and pi falls out. There is no circle anywhere in the experiment.

probability · Monte Carlo
When a shared birthday becomes likely. The chance that some pair in a group shares a birthday, against group size. It passes a half at 23 people, where the probability is 50.7%.

Twenty-three people

A room needs 253 people before someone probably shares a birthday with you. It needs 23 before two of them probably share one with each other. The gap between those numbers is the whole problem.

probability · Birthday problem
Nine walks, and the square root. 9 independent walks of 400 steps, each step one place left or right. The dashed curves are ±√n: the walks stay near them, spill past them, and come back — which is what a typical distance means as opposed to a limit.

A walk that always comes home, until it does not

Step left or right at random, forever, and the walk returns to where it started with certainty. On a grid it also returns. In space it does not, and about a third of walks leave and never come back.

probability · Random walk
Three doors, as areas. Staying wins 33.3% of the time and switching wins 66.7%, because the host's choice is constrained by what the host can see, so opening a door rules a region out without moving any boundary.

The door that was not opened

Three doors, one prize, a host who opens a losing door and offers a swap. Switching wins two times in three, and the reason is not about doors — it is about what the host was allowed to do.

probability · Bayes
A line, a point, and many parallels. A disc whose lines are arcs meeting the boundary at right angles, showing several lines through one point that never meet a given line.

Two worlds that both obey the rules

A statement is independent of a list of axioms when there is a structure satisfying the axioms where it holds and another where it fails. That is not a claim about what nobody has managed to prove — it is a proof that nobody can.

logic · Models
The Goodstein sequence from 4, with the ordinal beside each term. A table of the Goodstein sequence with each term's hereditary representation and the ordinal obtained by replacing the base with omega.

A sequence that explodes and still stops

Goodstein's sequence starting at 4 climbs past any number you care to name and reaches zero after about ten to the hundred and twenty million steps. The proof that it stops is a second sequence, running alongside it, that goes down.

logic · Ordinals
One set of sums, two scalings, two different limits. The exact distribution of a sum of n independent copies, scaled two ways. Divided by n it collapses onto the mean; divided by the square root of n it holds a fixed width and settles into a shape.

The average settles and the wobble does not

Two theorems are usually met a page apart and sound as though one is a sharper version of the other. They are the same sums looked at through two different magnifying glasses: divide by the number of them and everything collapses to a point, divide by its square root and a shape appears.

probability · Central limit
A lopsided distribution added to itself, and the shape that returns. On the left, the exact distribution of a sum of copies of one lopsided distribution, standardised, for several counts: the shapes converge. On the right, the bell curve convolved with itself, which is the bell curve again.

The shape that averaging leaves alone

Adding independent quantities blurs their distributions together, and rescaling restores the width. Almost every shape is changed by that operation. Exactly one is returned unaltered, and that is why sums of unrelated things keep arriving at it.

probability · Central limit
Averages of a heavy-tailed quantity, which never settle. Running averages of draws from a Cauchy distribution, which jump rather than converge, beside the cumulative distributions of averages of 1, 4 and 16 draws, which lie on top of one another.

An average that never settles

The average of many independent quantities is supposed to steady as their number grows. For one famous distribution it does not steady at all — the average of a thousand draws has exactly the same distribution as a single draw, and no amount of further averaging changes it.

probability · Central limit
Every way of choosing one thing from each of 4 pairs. A table with one row per choice function on a small family of pairs, each row giving what it takes from each pair, with the row a stated rule names picked out.

The choice nobody can write down

Given finitely many pairs, picking one thing from each is a finite list of decisions and needs no justification. Given infinitely many, the list cannot be finished — and whether one exists anyway is an axiom, independent of everything else, whose consequences include a theorem most people refuse to believe.

logic · Axiom of choice
Two unbiased estimates of one integral, and their spread. The sharply peaked integrand with the proposal density that follows it, above a strip plot of 200 estimates from each of two methods; the weighted estimates cluster 4.2 times more tightly about the same value.

Sampling where the answer lives

Monte Carlo error cannot be made to fall faster than the square root, so the only thing left to attack is the constant in front of it. Drawing points where the integrand is large, and dividing by how often they were drawn, leaves the answer alone and can shrink the noise many times over.

probability · Monte Carlo
The expected number of monochromatic sets, and where it drops below one. The logarithm of the expected number of single-coloured 4, 5, 6-point sets in a random two-colouring, plotted against the number of points, with the crossing of one marked for each.

The colouring nobody has ever seen

Count the monochromatic sets a random colouring is expected to contain. If the average is below one, some colouring has none — and the argument is finished, having produced nothing anyone can look at.

discrete · Ramsey theory
The tower of sizes, and the gap in it. A tower of infinite sizes, each the number of sub-collections of the one below, with the space between the first two marked as the one no proof decides.

The size that cannot be pinned down

There is no largest infinity, because no collection has as many members as it has sub-collections. What is not settled is whether anything sits between the first two — and that is not an open problem but a proved absence of an answer.

logic · Cardinality
The fast-growing hierarchy at its first few ordinals. A table of the fast-growing hierarchy: one row per ordinal index, one column per argument, with the cells too large to evaluate marked as such.

An ordinal as a growth rate

Index a family of functions by the ordinals, each one iterating the last, and the index becomes a measure of how fast a function grows. The point where the index leaves what arithmetic can prove is exactly where the Goodstein sequence became unprovable.

logic · Ordinals
How many of 8 people get their own hat, against the Poisson with mean 1. Paired bars for each number of people getting their own hat: the exact share of arrangements and the Poisson probability with mean one, nearly equal at every count.

How many get their own hat

The chance that nobody gets their own hat settles on 1/e. The chance that exactly one person does settles on 1/e too, exactly two on 1/(2e), exactly three on 1/(6e) — the Poisson distribution with mean 1. The reason is a set of averages that come out exactly 1 at every size, and the counts reach the limit so fast that eight hats are within six ten-thousandths of it.

probability · Inclusion exclusion
A year of birthdays with a seasonal swing of ±50%, peaking in September. Bars for the 365 days of a model calendar with a seasonal swing of ±50%, peaking in September, drawn as each day's excess or shortfall against an even year. A shared birthday among 23 people has chance 54.86% against 50.73% for the even year, the first group with an even chance is 22, and the calendar behaves like 324.4 equally likely days.

Any unevenness brings the match sooner

Real birthdays are not spread evenly across the year, and every such departure pushes the famous twenty-three down rather than up. The proof is one move on two days at a time, and what it leaves behind is a single number — the one ecologists use to count species.

probability · Birthday problem
The order type of a nonstandard model of arithmetic. The ordinary numbers as a run of dots, followed by 9 galaxies — copies of the integers — at the positions of c/2 up to 2c, ordered densely like the rationals, with c² beyond. Infinitely many more galaxies lie between those drawn.

A number larger than every number

Ask for a number bigger than 0, bigger than 1, bigger than 2, and so on for ever. Every finite piece of that request is granted by an ordinary number, so compactness grants all of it at once — in a structure that satisfies every sentence true of the whole numbers and still contains something beyond all of them. Nothing in first-order logic can say 'and nothing else'.

logic · Models
Fair bits from a biased coin. 44 flips of a coin biased 0.7 towards 1, read in 22 pairs. Mixed pairs are kept and give their first bit; matched pairs are discarded. Over a long run the output is 50.2% ones, at 0.210 output bits per flip.

Fair bits from an unfair coin

Read a biased coin's flips in pairs, keep 01 as 0 and 10 as 1, and throw away the rest: the output is exactly fair, whatever the bias, and nobody needs to know the bias. The trick wastes most of the coin, the waste can be recycled almost up to the ceiling Shannon's entropy sets — and it fails quietly the moment the flips remember each other.

computation · Pseudorandomness
An additive function that is nowhere a line. A square window with 1157 points of the graph of an additive function that sends √2 to 0, scattered across the window and meeting every part of it, beside a dashed diagonal marking the straight line y = x.

A function that adds and is nowhere a line

Every continuous function with f(x + y) = f(x) + f(y) is a straight line through the origin. Drop continuity and, given the axiom of choice, there are others — functions that add perfectly and whose graphs are scattered densely over the whole plane. A finite piece of the construction can be drawn exactly; the whole of it needs a basis of the real numbers that no one can write down.

logic · Axiom of choice
Infinitely many guessers, finitely many wrong. Three rows over the first 40 places: the hats worn, the chosen representative of their class, and a row of marks showing each guess right or wrong. The 5 wrong guesses all fall within the first 14 places, up to a marked place; every later guess is right.

Infinitely many guessers, finitely many wrong

An infinite line of people each wears a black or white hat, sees every hat in front and none of their own, and must guess their own colour. With a finite line, each guesser is right half the time whatever they agree in advance. With an infinite line and the axiom of choice, they can agree a strategy under which all but finitely many are right — and nobody can carry it out.

logic · Axiom of choice
Two children, and at least one is a boy. Four equally likely families drawn as quarters of a square: the question “is at least one a boy?” rules out only the girl–girl family, and leaves three equal quarters; the chance of two boys is 33.3%.

Two children and the sentence about one of them

A family has two children and at least one is a boy. The chance that both are boys is one in three — or one in two, or anything from one in three to certainty — and every one of those answers is right for some way the sentence could have come to be said. There is no host and no door, and the protocol is still the whole problem.

probability · Bayes

Named alongside it

The objects these essays reach for when they reach for this one.

ConvergenceExpectationNormal distributionVarianceAxiom of choiceConsistencyConvergence rateCounting argumentLimitModelNon-measurable setScaling

All concepts