Notes from the long tail—data with character. R, Python, and a healthy disregard for normal distributions.

fig. 1 · a rigged galton board X ∼ Skew(γ(t)) · γ ∈ [−1, +1]
mean μ̂ median mode skewness γ̂₁