Case file
Clustering Illusion
- Filed under
- Not Enough Meaning
The charge
Clustering illusion is seeing meaningful streaks or clumps in data that are actually compatible with randomness.
How it operates
Human pattern detection is tuned to find structure quickly, and random processes naturally produce local clusters that feel non-random.
Logged incidents
- Incident 01
A PM sees a burst of cancellations from one city and assumes a local competitor launched there.
- Incident 02
A QA lead notices several bugs in one module over two days and concludes the architecture is uniquely broken.
- Incident 03
An investor sees three founder departures in one month and reads it as proof of sector collapse.
What to watch for
Ask: Would this pattern still look impressive if I compared it with random data or a null model?
Recommended action
Use run charts or control charts and compare observed clusters to a random baseline before inferring a cause.
Known associates
- ConfabulationConfabulation is unintentionally filling gaps in memory or explanation with details that feel true but were…
- Insensitivity to Sample SizeInsensitivity to sample size is treating small samples as if they are just as reliable as large ones.
- Neglect of ProbabilityNeglect of probability is reacting to how vivid or scary an outcome is while giving too little weight to how…
- Anecdotal FallacyAnecdotal fallacy is letting one or two vivid stories outweigh broader and better-quality evidence.
- Illusion of ValidityIllusion of validity is feeling highly confident in a judgment because the evidence forms a neat story, even…
- Masked-Man FallacyMasked-man fallacy is treating differences in what is known about something under one description versus…
Source of record