Einsteins biggest mistake?

What, if any, was Einstein’s biggest mistake, the one most affecting our physics today? There is a perhaps apocryphal story, recounted by George Gamow, that he counted his cosmological constant as his biggest blunder. We now know his hypothesized cosmological constant to be correct. His lifelong rejection of quantum mechanics, an interesting side-story in the evolution of 20th-century physics, is a candidate. None of these introduced difficulties in how our physics is done today. It can be argued that his biggest actual mistake, one that affects many subfields of physics and chemistry and bewilders students today, occurred in his naming of his A and B coefficients.

💡 Research Summary

The paper revisits the long‑standing question of “What was Einstein’s biggest mistake?” and evaluates three candidates: the cosmological constant, his rejection of quantum mechanics, and the naming of the Einstein A and B coefficients. The first two are examined historically. The cosmological constant (Λ) was indeed called a “blunder” by Einstein, a story popularised by George Gamow. Modern observations—type‑Ia supernovae, cosmic microwave background anisotropies, and large‑scale structure—have confirmed that Λ is non‑zero, turning the anecdote into a retrospective mischaracterisation rather than a genuine error that hampers current physics. Einstein’s philosophical opposition to quantum mechanics, while influential in the early development of the theory, did not introduce lasting methodological problems; quantum mechanics survived his objections and now underpins virtually all of modern physics.

The core argument of the article is that the real, pervasive mistake lies in the terminology surrounding the Einstein A and B coefficients introduced in Einstein’s 1917 paper on radiative processes. In the original German text, the word “Wahrscheinlichkeit” was used to describe dW, the number of photons emitted in a time interval dt. Modern readers have interpreted this as a probability, leading to the widespread label “transition probability” for the coefficient A. However, A has dimensions of s⁻¹ (it is a rate), whereas a true probability is dimensionless and confined to the interval