Quick guide: coincidences in ten minutes

A fast, scannable overview of this site's core ideas, each linked to the fuller explainer, for a reader who wants the shape of the argument before the depth.

Established Supported by convergent, high-quality evidence.

Plain-language answer

This page is the whole site compressed into five ideas. Each one links to a fuller explainer — read this straight through for the shape of the argument, or follow any link to go deeper on that one point.

Five ideas

1. A coincidence is a relationship between events, not a verdict about their cause. What is a coincidence? separates the event itself from the surprise it produced, how improbable it was under some model, what meaning someone attaches to it, and the separate question of whether a real causal connection exists. Bundling all of these into one judgment (“that was just chance” or “that must mean something”) skips past several genuinely different questions.

2. Every probability is relative to a model you have to state. “What were the odds?” is never a complete question — probability is a model shows how the same event can carry very different odds depending on which comparison class and which assumptions were used to calculate them.

3. More opportunities make a “hit” much more likely than intuition expects. Multiple opportunities and selection effects and the birthday problem both show how quickly the probability of some match rises once you count every chance for one, not just the one that got noticed. Across a large enough population, even very rare events become close to guaranteed.(Diaconis & Mosteller, 1989)

4. Minds are built to notice, remember, and narrate patterns — including ones that aren’t really there. Pattern perception, availability and memory, and agency and causal inference cover why a coincidence gets noticed, remembered vividly, and woven into a causal story almost automatically — independent of whether anything unusual actually happened.

5. There’s a method for weighing a specific claim, and it isn’t a single number. How to evaluate a coincidence claim walks through verifying the facts, defining a precise match, choosing a reference class, counting opportunities, and comparing explanations — ending in an honest statement of what’s established, plausible, unsupported, or still unresolved, rather than a single verdict dressed up as certainty.

What this site is not trying to do

This site does not try to prove every coincidence is “just” chance, debunk personal meaning, or offer legal, medical, or financial conclusions about any specific case. A chance origin and a real personal significance can both be true of the same event — see personal meaning without causal proof.

Where to go next

Key takeaways

  • A coincidence is several separable questions (event, surprise, probability, meaning, cause) bundled together — separate them before drawing a conclusion.
  • Probabilities are only meaningful relative to a stated model and reference class.
  • The best response to a striking claim is a method, not a single probability calculation.

Sources

  1. Diaconis, Mosteller (1989). Methods for Studying Coincidences. Journal of the American Statistical Association, 84(408), 853-861. https://doi.org/10.1080/01621459.1989.10478847 ↩