Adding Epicycles

In scientific modeling, adding epicycles refers to augmenting a model’s structure to accommodate unexplained data. This practice, which has its roots in ancient geocentric theories of astronomy, often compromises both the model’s simplicity and its predictive accuracy.

Historical Origin

  • Ptolemaic Model: Geocentric model using circles and epicycles to account for planetary movements.
  • Copernican Revolution: Introduced heliocentric model, negating the need for epicycles.

Core Principles

  • Complexity Creep: Progressive addition of epicycles to accommodate new data.
  • Parsimony: The simpler explanation is preferred when all other factors are equal.
  • Falsifiability: Increasing epicycles compromises a model’s capacity to be proven wrong.
  • Predictive Failure: Models with excessive epicycles often fail to make accurate, reliable predictions.

Methodological Implications

  • Anomalies: Data inconsistencies that necessitate additional epicycles.
  • Methodological Criticism: Addition of epicycles renders a model less falsifiable and thus unscientific.
  • Scientific Inertia: Resistance to modifying or abandoning complex models due to intellectual or financial investment.
  • Operational Complexity: Real-world applications become impractical due to the model’s convoluted nature.

Mathematical Analogy

  • Fourier Series: Represents periodic functions as sums of sines and cosines, similar to adding terms for better data fit. \[ f(x) = a_0 + \sum (a_n \cos(nx) + b_n \sin(nx)) \]


  • Reactivity: Adjustments to models are typically reactive, not predictive.
  • Inflation of Parameters: Additional complexity risks overfitting the model to the data.
  • Loss of Explanatory Power: Additional epicycles diminish the model’s ability to make clear predictions.
  • Data Overfitting: The model begins to “memorize” rather than “understand” data, compromising its utility.


  • Derived from Greek “epi-” (upon) and “kuklos” (circle).

Contemporary Relevance

  • Economics: Efficient Market Hypothesis critiqued for including behavioral epicycles.
  • Medicine: Overcomplicated models hinder predictive accuracy in disease understanding.