Why Coincidences Happen: The Law of Truly Large Numbers

Quick answer: Coincidences happen because millions of events, people and possible matches create countless opportunities for unlikely outcomes. Statisticians Persi Diaconis and Frederick Mosteller called this the law of truly large numbers (1989): with a large enough sample, any outrageous thing is likely to happen. Memory adds to the effect, because we remember surprising matches and forget ordinary misses. For example, an event with a one-in-a-million chance on each independent try has about a 63.2% chance of happening at least once in a million tries, from P(at least once) = 1 − (1 − p)^n.

Facts checked November 22, 2026.

You think of an old friend and they call that afternoon. Two strangers at a party share a birthday. A woman wins the lottery twice. It is hard not to feel that something mysterious is going on. Mathematics has a calmer explanation: in a world this big, surprising things are not just possible but expected. Here is the math of coincidences, the famous examples, and a calculator that shows how “impossible” odds become likely.

The Law of Truly Large Numbers

In a 1989 paper, “Methods for Studying Coincidences,” Persi Diaconis and Frederick Mosteller summarized it in one line: “With a large enough sample, any outrageous thing is likely to happen.” They gave an example: if a one-in-a-million coincidence happens to one person in a million each day, then in a population of 250 million (the US at the time) we expect about 250 such coincidences every day, close to 100,000 a year.

They identified four main sources of coincidences:

  • Hidden causes: the events were connected after all.
  • Psychology: we remember hits and forget the many misses.
  • Multiple endpoints: many different matches would have felt surprising, and “close” matches count too.
  • The law of truly large numbers: with enough people and chances, rare events pile up.

The Key Formula

If something has probability \(p\) of happening on each try, the chance it happens at least once in \(n\) independent tries is

\[ P(\text{at least once}) = 1 – (1 – p)^n. \]

For a one-in-a-million event (\(p = 10^{-6}\)) with a million chances, \(1 – (1 – 10^{-6})^{10^6} \approx 1 – e^{-1} \approx 63.2\%\). The event becomes more likely than not once \(n \ge \ln 2 / p\), about \(0.693/p\): for one in a million, about 693,000 tries.

Two cautions. First, the expected number of occurrences, \(n \times p\), is not the same as the probability of at least one: with \(n p = 1\) you expect one occurrence on average, but the chance of at least one is only 63%. Second, the formula assumes independent tries. If events share a cause, like everyone in a city watching the same TV show, matches become much more likely than this simple model predicts.

Littlewood’s Law of Miracles

Littlewood’s law of miracles says a person should expect an apparently extraordinary event about once a month. It is named after British mathematician J. E. Littlewood, who discussed improbable coincidences in A Mathematician’s Miscellany (1953), but the familiar once-a-month version was popularized by physicist Freeman Dyson in his 2004 essay “One in a Million.” Dyson’s illustration: call a one-in-a-million event a “miracle,” and suppose you notice about one event per second for 8 hours a day. Then you experience about a million events in roughly 35 days.

\[ 8 \text{ hours} \times 3{,}600 \text{ s/hour} = 28{,}800 \text{ events a day}, \qquad \frac{1{,}000{,}000}{28{,}800} \approx 35 \text{ days}. \]

The one-event-per-second rate is an illustration, not a measurement, so the model predicts roughly one remarkable event a month on average, not exactly one.

Try It: Coincidence Calculator



Famous Coincidences, Explained

The Double Lottery Winner

In the 1980s, a New Jersey woman won the state lottery twice in four months, and a New York Times story called it a “1 in 17 trillion” long shot. Diaconis and Mosteller pointed out that this is the chance that one specific person, buying one ticket in each of two specific drawings, wins both: a “not-very-relevant question.” The relevant question is whether anyone among millions of regular players ever wins twice. Purdue statisticians Stephen Samuels and George McCabe found it “practically a sure thing” over time: better than even odds somewhere in the US within seven years, and better than 1 in 30 within a four-month window. These were estimates for lottery play at the time, not predictions for today’s games.

Shared Birthdays

With just 23 people, the chance that two share a birthday is about 50%, because there are \(\binom{23}{2} = 253\) pairs. Loosen the rule to birthdays within one day (the same date, the day before or the day after), and only 14 people give even odds; within a week, about 7. Small changes in what counts as a “match” change the numbers a lot. See the birthday paradox explained.

“I Was Just Thinking About You”

You think about many people every day, and many people call you. Over months, a call from someone you recently thought about is almost certain to happen sometimes. You remember those moments and forget the thousands of times you thought of someone and they did not call.

Are Coincidences the Same as Synchronicity?

Psychologist Carl Jung used the word synchronicity for coincidences that feel meaningfully connected without an ordinary cause; Diaconis and Mosteller’s paper discusses Jung’s examples. A coincidence describes events; synchronicity is an interpretation. Statistics does not say a coincidence cannot feel personally meaningful. It asks a narrower question: does a pattern happen more often than a clear probability model predicts? If it does, and keeps happening, that is a reason to look for a real cause.

How to Think About a Coincidence

  1. Ask “what are the chances of anything like this?” not “what are the chances of exactly this?”
  2. Count the opportunities. How many people, days and possible matches were there?
  3. Remember the misses. Surprising hits get noticed; ordinary non-events do not.
  4. Look for hidden causes. Sometimes events are connected, like two friends recommending the same viral video.
  5. Beware the look-elsewhere effect. If you search thousands of possible patterns after the fact, some will look amazing by chance. Scientists correct for this when testing many hypotheses.

The same thinking protects against misleading claims, from “hot streaks” to the lottery, and connects to regression to the mean.

Common Misconceptions

“A one-in-a-million event almost never happens.” To you on a given try, no. Somewhere in a big population, it happens constantly.

“Coincidences must have a hidden meaning.” Most have none; they are expected results of huge numbers of chances. Sometimes there is a real hidden cause, which is worth checking.

“Law of large numbers and law of truly large numbers are the same.” No. The law of large numbers says averages settle down with many trials; the law of truly large numbers says rare events become common when there are enough chances.

Watch: The Math of Coincidences

If you learn better by watching, these videos are worth your time.

Probability and Outrageous Events: David J. Hand on The Law of Truly Large Numbers (Samuel Devis). Statistician David Hand explains why extremely unlikely events happen all the time.
The Strange Math That Predicts (Almost) Anything (Veritasium). How probability tames randomness over many events.

Check Your Understanding

Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.

Answer each question and press Submit to see your result right away.

1. What does the law of truly large numbers say?
Found a bug? Let us know!
2. What is P(at least once) for probability p over n independent tries?
Found a bug? Let us know!
3. By Littlewood’s law, how often should you expect a one-in-a-million “miracle”?
Found a bug? Let us know!
4. How many people give even odds that two have birthdays within one day?
Found a bug? Let us know!
5. Why was “1 in 17 trillion” misleading for the double lottery winner?
Found a bug? Let us know!

Frequently Asked Questions

Why do coincidences happen so often?

Because there are enormous numbers of people, events and possible matches. Even very unlikely events become likely when there are enough chances, an idea called the law of truly large numbers.

What is the law of truly large numbers?

A principle described by Persi Diaconis and Frederick Mosteller in 1989: with a large enough sample, any outrageous thing is likely to happen.

What is Littlewood’s law of miracles?

The idea that a person should expect an apparently extraordinary event about once a month. It is named after J. E. Littlewood (1953) and was popularized by Freeman Dyson in 2004: about one noticed event per second for 8 hours a day gives about a million events in 35 days.

What is the difference between coincidence and synchronicity?

A coincidence is a surprising combination of events without an established causal link. Synchronicity, Carl Jung’s term, interprets such events as meaningfully connected.

What are the odds of winning the lottery twice?

For one specific person and two specific tickets, astronomically small. But Purdue statisticians found it better than even odds that someone in the US wins twice within seven years.

How do you calculate the chance of a coincidence?

Use P(at least once) = 1 − (1 − p)^n, where p is the chance per opportunity and n is the number of opportunities.

Keep Exploring

Sources and Further Reading

Related to This Article

What people say about "Why Coincidences Happen: The Law of Truly Large Numbers - Effortless Math"?

No one replied yet.

Leave a Reply