Streetlight Effect
Looking where the light is good rather than where the answer is.
The streetlight effect is the tendency to search where observation is easy rather than where the thing you are looking for actually is. The name comes from the parable of a drunk man searching under a streetlight for keys he dropped elsewhere, because, as he explains, "this is where the light is." The joke works because the logic is locally sound and globally useless.
In practice the effect is rarely a personal failing so much as a property of Instrumentation: a system logs what someone thought to log, a metric measures what was cheap to measure, and both attract disproportionate attention afterward simply because they are the data that exists. A dashboard full of request latency says nothing about the request that never got dispatched. Investigation converges on the visible layer, and the write-up mistakes thoroughness of the search for coverage of the problem.
Because the cause is instrumentation rather than laziness, the fix is rarely "look harder" — a wider search under the same streetlight still only covers the same patch of ground. The fix is Observability: extending what can be seen at all, so that looking becomes possible in places it previously was not. This is also why the effect compounds with Survivorship Bias and Selection Bias — cases that never reach the lit area are excluded from the sample before any bias in interpreting the sample even gets a chance to operate, which is a Root Cause Analysis built entirely on the wrong half of the evidence.
See also4
Ground Truth
The measurement that arbitrates, as opposed to the artifact that merely reports.
Method95 connections
Silent Failure
A failure that produces no signal, so the absence of an error reads as success.
Method63 connections
Truncation Bias
Reading a truncated result as if it were the whole result, so every counterexample is invisible.
Method26 connections
Fermi Estimation
Reaching a defensible order-of-magnitude answer by decomposing a question into estimable factors.
Method23 connections
Linked from5
- Base Rate FallacyMethod
judging a specific case from vivid evidence while discounting how common the categories actually are.
- Premature OptimizationMethod
Spending effort making code fast before you know where it's actually slow trades clarity for speed you may not need.
- Selection BiasMethod
A sample distorted because inclusion in it was never random.
- Survivorship BiasMethod
Drawing conclusions from a sample already filtered by success.
- XY ProblemMethod
Asking for help with your attempted solution (Y) instead of your actual problem (X), so the help solves the wrong thing.