Skip to main content

Pattern Recognition

What pattern are you seeing that others aren't — and how do you know it's real?

Every good decision, investment, and innovation starts with someone recognising a pattern before the crowd does. AI excels at pattern-matching within defined datasets. Humans excel at cross-domain pattern transfer — seeing that a supply chain problem rhymes with a biological immune response.

Without Pattern RecognitionWith Pattern Recognition
React to eventsAnticipate events
See isolated factsSee connected systems
Surprised by outcomesExpected the shape
"Nobody saw that coming""The signs were there"

Pattern Types

TypeWhat It IsExample
StructuralSame shape across different domainsFeedback loops in biology, business, and software
TemporalSequence that repeats over timeMarket cycles, technology adoption curves
AnomalyWhat doesn't fit the expected patternThe data point everyone ignores is often the signal
CausalA reliably produces BIncentive changes behaviour — every time
Second-orderThe pattern behind the patternCheap credit → asset inflation → inequality → populism

Cross-Domain Mapping

The most valuable pattern recognition connects fields that don't usually talk to each other:

Pattern In...Also Appears In...The Insight
Biology (immune system)CybersecurityDetect-and-respond beats static defence
Physics (entropy)Business (culture decay)Without energy input, systems degrade
Music (tension/resolution)StorytellingUnresolved tension creates engagement
Evolution (mutation + selection)Innovation (experiment + kill)Volume of attempts matters more than quality of planning
Ecology (niche filling)Market positioningEvery gap gets filled — the question is by whom

Ask: "Where have I seen this shape before?" across completely different contexts.

Base Rate Discipline

What usually happens is more predictive than what you think will happen:

TrapWhat HappensThe Fix
Narrative fallacyCompelling story overrides probabilityCheck: what percentage of similar situations had this outcome?
Recency biasRecent events feel more likely than they areLook at 10+ year datasets, not last quarter
Selection biasYou only see the survivorsAsk: for every success story, how many failures had the same pattern?
Confirmation biasYou see patterns that confirm your theoryActively search for evidence against your pattern

The most dangerous pattern is the one you're sure of. That's where overconfidence lives.

Feedback Loop Detection

The most important pattern in any system:

Loop TypeHow to Spot ItWhat It Does
ReinforcingGrowth accelerates, or decline acceleratesCompound effect — good or bad
BalancingSystem resists change, returns to equilibriumStability, but also stagnation
DelayedEffect appears much later than causeMost dangerous — by the time you see it, it's too late

When you find a reinforcing loop, ask: is this compounding in my favour or against me?

Practice Protocol

  1. Study history — Patterns repeat because human nature doesn't change
  2. Play prediction games — Poker, markets, sports — then review your accuracy
  3. Keep a decision journal — Track what you predicted vs what happened
  4. Map across domains — "Where have I seen this shape before?"
  5. Seek disconfirmation — The pattern you're most sure of needs the most testing

The Shadow

Seeing patterns that aren't there. Conspiracy thinking. Over-fitting — forcing complex explanations onto simple events. Confirmation bias dressed as insight. Apophenia: the human tendency to see meaning in randomness.

By Archetype

ArchetypePattern Style
PhilosopherSees deep structural patterns across domains
RealistTests patterns against evidence before acting
EngineerBuilds systems that exploit patterns at scale

Context