fn regularized_gamma_p(a: f64, x: f64) -> f64
Regularized lower incomplete gamma function P(a, x) = gamma(a, x) / Gamma(a). Uses series expansion for x < a+1, continued fraction otherwise.