The most powerful idea

Darwinists say that Natural selection is the most powerful idea ever discovered. I am trying to understand why they say this, even tho the idea is very nearly a tautology with no direct scientific consequences. So I looked for analogies in other sciences.

The closest I found in physics was Maxwell–Boltzmann statistics, but that has direct quantitative consequences and is not politicized. The best analogy is the Invisible hand in economics.

Here are the similarities between the invisible hand and natural selection.

• Sacred words. Each phrases was invented and popularized by a great master
  who wrote a great book that became the bible of the field.
  
• Tautology. Both are true by definition, and not subject to empirical test.
  
• Metaphor. Among the more serious scholars, the phrase is treated as just
  a metaphor to guide further investigations.
  
• Emergent. Both phrases are about how individual selfishness can lead to
  system equilibrium, without the pieces understanding events in the large.
  
• Anthropomorphizing nature. Both attempt to describe a force of nature,
  but do it by pretending that there is some imaginary being applying the force.
  The word "select" normally means the voluntary choice of a conscious being. Likewise with "hand".
  
• Deity denial. While the metaphor suggests a conscious force, it also denies that a
  supernatural god is involved.

To Dawkins, the really important point is that natural selection allows an atheistic worldview:

Although atheism might have been logically tenable before Darwin, Darwin made it possible to be an intellectually fulfilled atheist.

Just for comparison, here are some powerful ideas from the 20th century, mostly in mathematics.

• Goedel's completeness and incompleteness theorems
  
• Axiomatization of set theory
  
• Infinite dimensional metric spaces
  
• Noether's theorem on symmetry and conservation laws
  
• DNA digitally encodes components for proteins
  
• Relativistic symmetry and causality
  
• Algebraic topology
  
• Noncommuting observables
  
• Connections on bundles