Dilution effect

Biological Parallel

Predicting rabbit population growth from current abundance, weather data, and predator counts—high accuracy. Adding irrelevant information (lunar phase, wind direction, atmospheric pressure) dilutes the model: prediction accuracy decreases as noise drowns signal. The dilution effect: more data isn't always better. Biological forecasting faces this constantly: adding marginally relevant variables (distant climate indices, tangentially related species) worsens predictions. The mechanism: finite data spread across more parameters increases uncertainty. Parsimony isn't elegance; it's predictive necessity.