15,942 Impossible Kinematics Events in 90 Days. Low-Risk Gaps Average 104.8 Hours. High-Risk Gaps Average 17.6 Hours. The Anomaly Score Is Nearly Identical.

15,942 dark events carry the impossible_kinematics subtype over the last 90 days. The low-risk tier has gap durations 6× longer than the high-risk tier. The kinematic anomaly score is nearly the same in both.

The Setup

Dark events classify AIS gaps — periods when a vessel stops transmitting. The impossible_kinematics subtype marks gaps where the vessel reappeared at a position it could not physically reach within the elapsed time given its vessel class and expected speed envelope. The vessel went dark in one location and surfaced somewhere kinematically impossible. That is the flag.

In the 90 days ending July 19, 2026, 15,942 events across 10,840 unique vessels carried this subtype. The events distribute across four risk tiers: critical (892 events, 618 unique vessels), high (6,819 events, 3,527 vessels), medium (4,991 events, 3,899 vessels), and low (3,136 events, 2,796 vessels). An additional 104 events have no tier assigned. The distribution by volume follows an expected pattern — most events fall in high and medium. What does not follow an expected pattern is the relationship between tier and gap duration.

The Chain

Critical-tier impossible_kinematics events average 20.0 hours of gap duration. High-tier events average 17.6 hours. Medium-tier events average 45.7 hours. Low-tier events average 104.8 hours.

If risk tier reflected gap duration, longer gaps would carry higher risk: a vessel dark for 104 hours seems more alarming than one dark for 17. The data shows the inverse. Short-gap impossible kinematics are rated high or critical. Long-gap impossible kinematics are rated low.

The kinematic_anomaly_score tells a parallel story that clarifies the inversion. Across all four tiers, the score is almost flat: 0.578 (critical), 0.616 (high), 0.603 (medium), 0.621 (low). The low-tier events have the highest kinematic anomaly score. Critical-tier events have the lowest. The kinematic impossibility itself — measured by how unreachable the reappearance position is — does not vary meaningfully with risk tier.

This means the risk tier is not driven by the kinematic anomaly score. It is driven by something else. The most structurally coherent explanation is that gap duration functions as an inverse risk modifier: a 17-hour gap with an impossible reappearance is harder to dismiss by conventional means than a 104-hour gap. When a vessel is dark for four days, mundane explanations accumulate — port waiting, equipment failure, regulatory holding, severe weather. When a vessel is dark for 17 hours and reappears somewhere it could not have traveled, the mundane explanations are fewer.

The Implication

The inversion has a direct operational consequence. Triage by risk tier moves critical and high-tier events to the top of the review queue and low-tier events to the bottom. But low-tier events have a kinematic anomaly score (0.621) marginally higher than critical-tier events (0.578). They also have 3× the gap duration of high-tier events.

An operator reading "low risk" may reasonably conclude the kinematic signature is weaker. That reading is wrong. The kinematic signature is equally strong — sometimes stronger. What is weaker is the contextual evidence that a short-gap impossible movement would generate. The "low risk" label means the gap was long enough that mundane explanations remain available, not that the kinematic impossibility is less severe.

3,136 low-tier events with avg kinematic anomaly score 0.621 and avg gap of 104.8 hours represent a population that standard tier-based triage may chronically under-examine. These events have the same underlying kinematic flag as the high-risk events. They are deprioritized because of when they happened relative to the dark period, not because of what the kinematics show.

The 104 events with no tier assignment are an additional gap. Their avg kinematic anomaly score (0.606) is consistent with the rest of the population. They will not appear in any tier-based triage queue.

What to Watch

Whether the high-tier count grows faster than low-tier in subsequent 90-day windows. If short-gap impossible kinematics increase as a share of total events, it may indicate more operationally disciplined behavior — brief dark windows timed to minimize detection, at the cost of producing highly conspicuous reappearance positions.

The null-tier group also warrants attention. 104 events with no assigned tier and a kinematic anomaly score of 0.606 are currently outside any classification pathway.

Limitations

The kinematic_anomaly_score is a relative measure within the impossible_kinematics population, not an absolute probability of evasive intent. A score of 0.621 means the event is more anomalous than the population average — not that the event is confirmed evasion. The gap duration inversion hypothesis (longer gap lowers risk tier because mundane explanations are more available) is structurally consistent with the data but has not been confirmed against reviewed cases. The actual formula driving risk tier is not surfaced in the observable schema. The interpretation here is inferred from the correlation between tier, gap duration, and anomaly score across 15,942 events.

Data as of July 19–21, 2026. Sources: dark_events table, impossible_kinematics subtype. 90-day window ending July 19, 2026.