<![CDATA[Automatic Identification System (AIS) data provide continuous operational records that can be transformed into computational indicators for evaluating maritime service performance. This study aims to develop an AIS-based probabilistic model for assessing Ro-Ro ferry service-speed reliability through sailing-speed compliance. Speed over Ground (SOG) observations were used to determine whether vessels operated within a predefined reliable speed range during sailing. The reliable speed range was defined as 7–9 knots based on the performance indicator required by the local transportation authority responsible for route licensing, while sailing observations were identified using a 3–15 knot SOG filter. From 32,358 raw AIS messages collected between 1 January and 31 March 2026, 11,166 sailing observations were retained. Observation-based reliability was 35.63%, while duration-based reliability was 36.63%, indicating that 324.76 of 886.64 estimated sailing hours were spent within the reliable speed range. Vessel-level analysis showed substantial variation, with two vessels recording duration-based reliability above 50% and two others below 10%. The Weibull distribution provided the best parametric fit, with an Akaike Information Criterion (AIC) of 44,051.48, the lowest among the fitted candidate distributions. The cumulative distribution function indicated a 35.88% probability of operating within the 7–9 knot range. Monte Carlo bootstrap resampling with 10,000 iterations produced 95% confidence intervals of 34.75–36.53% and 35.38–37.88% for observation-based and duration-based reliability, respectively. These findings indicate that AIS data can support a reproducible screening approach for evaluating aggregate and vessel-level ferry service-speed reliability.]]>
