Two things broke together and it meant nothing

A parcel is late and the market is slow. An account behaves oddly on the same evening a message goes unanswered. The mind draws a line between them immediately, because two unusual things happening together must be connected.

Sometimes they are. Often the line is drawn between two entirely separate events by a brain that was built to find patterns and is not fussy about whether they exist.

Why coincidence feels impossible

Because the mind counts the hits and not the misses. Nobody notices the thousands of days on which two ordinary things happened together and meant nothing. The one day where two notable things coincided is memorable precisely because it was strange.

Strangeness is not evidence of connection. It is a measure of how often a combination occurs, and rare combinations occur constantly, because there are a great many possible combinations and only a few of them need to land.

How to break the two apart

1. Write the two events down separately, with times.
2. For each, ask which actor could have caused it on its own, using the attribution rule.
3. If the two answers are different actors, you almost certainly have a coincidence.
4. If they are the same actor, ask whether that actor could plausibly have caused both, and be strict about it.

Step three is where most of these resolve. A late parcel is a vendor or postal question. A slow site is a network question. There is no mechanism connecting them and nobody has ever proposed one that survives being stated out loud.

The version that is not coincidence

Some clusters are real and it would be dishonest to pretend otherwise. A general network slowdown really does make several things fail at once. A market with a genuine problem really can produce trouble across messaging, deposits and pages simultaneously.

The test is whether a single mechanism explains both without any hand waving. If explaining the connection needs a sentence with probably in it, you have a theory rather than a cause.

What follows from getting it right

You solve two problems instead of theorising about one. The combined theory is almost always unactionable, because it names something large and vague, and large vague culprits cannot be dealt with. Two separate problems each have a next step.

The other gain is that you stop escalating. A reader who believes two things are connected believes something big is happening, and belief that something big is happening is what produces rushed decisions about money.

Where the combined theory comes from

Usually from somebody else. A post appears connecting two unrelated observations, other people recognise one half of it from their own week, and the theory acquires supporters who have each confirmed a different piece of it.

This is how most confident claims in this subject are assembled. Nobody lied. Everybody contributed a true fragment. The connecting tissue was invented once, by one person, and then nobody went back to check it because the story already had witnesses.

The defence is boring and it works. Ask what single thing would have to be true for both halves to follow. If nobody can state it in one sentence without hedging, the theory is decoration on top of two separate facts.

• Back to the cases with no author
• The rule for assigning an actor
• Every case in one table