A clustering scheme has been applied for capturing qualitatively different surge motion patterns in the phase space. The scheme enables the identification of “high-run” incidents as soon as such motions are triggered and while their phenomenology has not yet been well developed. A “high run” is a surf-riding–like behavior, appearing in irregular following seas. The concept of finite-time coherent sets is exploited for deriving estimates of the probability of high-runs. The method is verified by identifying independently the corresponding hyperbolic Lagrangian coherent structures; then, consistency is sought between the two approaches. An important feature of the method is that it does not rely on the use of some empirical criterion for the high-run threshold, such as one based on the exceedance of an arbitrary high-speed level. Despite its computational burden, the proposed scheme offers “objective” statistical information on a ship’s high-run tendency that can be used for benchmarking simpler (approximative) probability calculation schemes.