ABSTRACTThe Bi‐normal density distribution function on a surface is represented by a position vector and covariance matrix. Its physical dimensions are described by the error ellipse. A generalized scalar is the radial or circular error which denotes the probability within a radius of the position. To compute the radial error probability (or probability circle) precisely, a non‐trivial numerical integration is necessary. Simpler but less accurate conventions in common use are the Drms and CEP. The error ellipse semi‐major axis is also sometimes applied to radial error. These three measures of radial error are subject to variations in probability as a function of the eccentricity of the distribution. The probability of a circle can be obtained simply and more accurately by the use of a third order polyn