Dilution of Precision (DOP) is a dimensionless geometry factor describing how satellite line-of-sight geometry maps range-measurement uncertainty into estimated-state uncertainty under a specified model.
For a linearised observation model with design matrix G and observation covariance R, the formal least-squares covariance is P = (Gᵀ R⁻¹ G)⁻¹ when the model is full-rank and its assumptions hold. With independent, equal-variance observations, R = sigma^2 I and P = sigma^2 Q, where Q = (GᵀG)⁻¹. Conventional DOP values are square roots of selected sums of diagonal entries of Q.1
This familiar unweighted DOP convention describes geometry, not the measurement-error scale. Converting it to a formal uncertainty requires a compatible range-error standard deviation and parameterisation. Unequal weights, correlations, biases, added clock states or other estimated parameters require the corresponding covariance model; rank deficiency or poor geometry can make a solution unstable or unavailable.1
Lower DOP generally means more favourable geometry for the same state vector and observation model. It does not establish lower realised position error, accuracy in metres, integrity, service availability or a performance guarantee. Galileo documentation, for example, treats PDOP-threshold availability separately from position-error accuracy and receiver measurements.123