Safe-to-Gate Monthly

May 2026

May 2026 held the median window near-flat around 136h across the whole horizon. The movement that mattered was still paired, still mid-horizon, still at Garden City.

Published June 22, 2026 · Framework TradeLanes Safe-to-Gate Monthly v1.1 · Data baseline 2026-07-13 (100%-coverage mart, voyage-keyed) · Status Operator-facing

Q2 archive edition. This is one of four monthly summaries - March, April, May, and June 2026 - published together as the Safe-to-Gate Q2 archive. Each links to that month's weekly editions.

Data-baseline note (2026-07-13). Computed on the 100%-coverage mart (production U.S. export observations, up from 44.5% pre-migration) and keyed on distinct vessel-voyages - normalized vessel name plus terminal - rather than booking/schedule-version rows, so a voyage published under two alliance carriers, or revised across many schedule versions, is counted once. May 2026: 701 distinct vessel-voyages, of which 655 carry an Earliest Return Date - the operational cohort where paired movement can occur; 42.1% of that cohort is co-moving. Schedules carrying a CY Cut but no ERD are a coverage gain but are excluded from the co-movement analysis, because a voyage with no ERD cannot show paired ERD-and-CY-Cut movement. Raw event counts reflect the full cohort.

This month's Safe-to-Gate weekly editions: Week of 3-9 May · Week of 10-16 May · Week of 17-23 May · Week of 24-30 May · Week of 31 May-6 Jun

1. Opening

Layer 1 · Headline · Opening Frame

May 2026 produced 4,579 CRW revision events across 701 distinct vessel-voyages. 52.8% of those events arrived beyond T-7. The mass of schedule movement sat inside the planning horizon, not at the late-stage boundary the system is designed to catch.

Revisions per voyage eased to roughly 6.5, against April's 7.2 and below March's 9.5. The quarter peaked in March and settled into a quieter, broader May: more voyages, fewer revisions each.

A methodology note covering the carrier-terminal data pipeline artifact first identified in Safe-to-Gate Monthly Dec 2025 (MAEU at Norfolk's primary container terminal) appears in Section 4. The artifact remains under engineering review and is more pronounced in May's residual data than in any prior month.

Two things were true in May. The aggregate story (the median window held near 136h at every snapshot from T-4w to final - essentially flat) beside the voyage story (the top balanced-activity voyages at Garden City each recorded 24 or more paired revisions). The contradiction is the story.

2. The Mechanism

Layer 2 · Pattern · Terminal anchor

The mechanism is paired ERD and CY Cut activity - Sustained Churn. 39.4% of May 2026's voyages with movement showed both fields revising together (co-moving), the largest class in the dual-signal breakdown. 24.1% showed CY Cut activity alone, 18.3% ERD alone, 15.8% held flat, and only 2.4% showed the two fields moving in opposite directions - the quarter's lowest divergence share. The mode is paired and sustained, not divergent.

USSAV Savannah carried the month's heaviest revision load and its top balanced-activity voyages. The terminal anchor held even as the network broadened.

The three-element read

Contradiction. The network median barely moved - near 136h at every horizon - and 701 voyages still produced 4,579 revision events. Both readings are true because the aggregate hides where the mechanism operates.

Operator implication. The plan committed at T-4w faced 52.8% of the month's revision activity beyond T-7, much of it in the T-14 to T-7 window where rail commits and drayage dispatch are being finalized. A near-flat median does not mean a settled plan; it means the movement netted out in aggregate while individual voyages kept moving.

Failure mode. Clustered revisions arrive mid-horizon. The window reshapes. Stability doesn't come. The commit deadline arrives. The exporter commits against the latest shape - which is overtaken by the next revision before the truck reaches the gate.

3. Bridge to Evidence

Layer 2 → Layer 3 · From claim to patterns

If mid-horizon Sustained Churn is the mechanism, three patterns must appear in the data:

1. Revision activity concentrated inside the planning horizon, not at its endpoints.
2. A non-trivial share of voyages with sustained, paired instability rather than directional drift.
3. That share concentrating at specific terminals, not distributed across the network.

Here's what May 2026 showed.

4. Quantified Proof

Layer 3 · Three findings that anchor the mechanism

52.8% of May 2026 revision events arrived beyond T-7. The mass of schedule movement sat between T-14 and T-7 (27.5%) and beyond T-14 (25.3%). 26.6% of events landed inside T-3. This locates the mechanism outside the 72-hour boundary and inside the planning horizon. (Timing shares are measured on ERD-anchored events; schedules carrying a CY Cut but no ERD have no T-minus anchor and are excluded from this distribution.)

Paired ERD and CY Cut activity is the dominant signature. 39.4% of voyages with movement showed both fields revising together; only 2.4% showed the two fields moving in opposite directions - the quarter's lowest divergence share. The dominant mode remained paired, not divergent.

Volume sits at 4,579 events across 701 distinct vessel-voyages, spread across 28 carriers and 12 U.S. ports. The network broadened in May - more voyages at more ports - while revisions per voyage held near April's level. The activity is broad-based, not isolated to one carrier or alliance.

5. The Illustrative Case

Layer 3 · The mechanism in one voyage

HMM VICTORY 059W at USSAV Garden City (carrier HDMU, departure May 30, 2026). 28 CRW revisions. Total absolute magnitude: 1,392 hours. Paired ERD and CY Cut activity: 14 ERD revisions and 14 CY Cut revisions across the same voyage - a perfectly balanced pair. The terminal moved the whole window as a single unit through 28 revisions.

This voyage carried roughly 50 hours of CRW motion per revision on average - the single-voyage signature of May 2026's mid-horizon Sustained Churn pattern. The exporter who committed drayage at T-4w against the published window committed against a plan the schedule had not finished drawing.

The window didn't break. It kept moving.

6. Concentration and Repeat Offenders

Part A - Concentration

Where Sustained Churn operated at scale in May 2026

USSAV Garden City - the mechanism in detail. Savannah carried 151 distinct vessel-voyages and 1,767 revision events - the month's heaviest load. The top balanced-activity voyages there show the cross-carrier paired signature at a single terminal:

Multiple carriers represented in the May 2026 USSAV concentration cluster. The signature is cross-carrier paired activity at one terminal - many carriers, one place.

USLAX - throughput gateway, not the churn center. Counted as distinct vessel-voyages rather than booking rows, Los Angeles carried 58 - far below the booking-level view and well short of Savannah's 151. Its 710 revision events run second to Garden City's load, but the sustained paired churn concentrates at Savannah, not at the largest throughput gateway. (Per the 2026-07-13 re-key, a sailing published under two alliance carriers is counted once; earlier booking-level counts overstated the West Coast vessel tallies.)

USORF - deferred pending methodology review. Several MAEU vessels at Norfolk appear in May 2026's residual top-activity list by CY Cut activity, with the artifact more pronounced this month (individual voyages recording 30 to 52 CY Cut revisions with little or no paired ERD movement). The CY Cut signature follows the same pipeline-artifact pattern first identified in Safe-to-Gate Monthly Dec 2025 - alternation between adjacent candidate timestamps independent of underlying terminal activity. These voyages are documented in the methodology note (Section 4) and are not used to support a Norfolk concentration finding in this edition.

Part B - Repeat Offenders

Prior-month qualifiers only · Temporal boundary rule

Every May 2026 vessel that qualified as a Repeat Offender under the temporal boundary rule is a MAEU vessel at Norfolk or Newark tied to the suppressed CY Cut artifact (Section 4) - for example PL GERMANY 620E (prior-month qualifier 2026-04) and MAERSK IDAHO 616E (prior-month qualifier 2026-03). Because these voyages reflect the pipeline artifact rather than operator-experienced movement, this edition makes no operational repeat-offender finding for May. The genuine concentration sits at USSAV Garden City, where the carriers rotate week to week rather than repeat.

7. Decision and Cost

Part A - The Decision Shift

Layer 4 · Operator behavior change

The commit horizon is wrong.

Exporters committing drayage, rail, and appointment decisions at T-4w to T-2w faced a schedule that took its final shape between T-14 and T-7 in May 2026. 52.8% of revision events arrived beyond T-7. A near-flat median window across the horizon does not mean the plan was settled - it means the per-voyage movement netted out in aggregate while individual schedules kept moving.

The decision shift: move the commit window inward where operational constraints allow. Defer the decision to the zone where the schedule has the highest information content - roughly T-7 to T-3 - even at the cost of tighter planning horizons.

Where the commit window can't move - and for many rail-dependent exports it structurally cannot - the alternative is to commit at T-4w but rebuild contingency at T-7. Treat T-7 as a mandatory replanning checkpoint, not an observational one.

Part B - Operational Cost Chain

Five steps from plan to cost - traced against HMM VICTORY 059W

1. Plan. At T-4w, the exporter commits drayage against the published window for HMM VICTORY 059W at USSAV Garden City. Rail slot booked. Driver scheduled. Trucking appointment reserved.
2. Churn. Through May 2026's mid-horizon, the terminal moved ERD and CY Cut forward together across 28 revisions. Total absolute motion: 1,392 hours.
3. Pressure. Smaller paired corrections arrived alongside the larger moves. Every check against the terminal returned a different gate window.
4. Commit. The commit deadline arrived. The driver was dispatched against the latest published window.
5. Cost. Driver wait time, container dwell, appointment rebooking - receipts of a plan made against a schedule that hadn't finished moving.

The cost cascade doesn't fail at step 5. It fails at step 4 - when the commit has to be made against a schedule still in motion. Steps 1 through 3 produce the failure. Step 5 receipts it.

Don't move step 5. Move step 4. Commit inward - where the schedule has resolved.

8. System Statement + Forward Hook

Layer 4 → Close

May 2026 produced 4,579 CRW revision events across 701 distinct vessel-voyages. 52.8% arrived beyond T-7, and the median window held near-flat around 136h across the entire horizon. The pattern is consistent with the Sustained Churn framing introduced in Safe-to-Gate Monthly Dec 2025 - paired mid-horizon revision activity that does not collapse into a stable plan, and that a near-flat aggregate hides rather than resolves.

The operator's problem isn't too many vessels at once. It is no stable point to act.

May extends the Q1-into-Q2 record: December through May, the window's instability has lived mid-horizon, the mode has stayed paired, and the terminal anchor has kept returning to USSAV Garden City. The volume rose into March and eased across April and May; the mechanism held the whole way. The next question is whether the Garden City concentration holds into June, and whether the mid-horizon mass shifts earlier or later in the window.

TRADELANES · SAFE-TO-GATE MONTHLY · MAY 2026

tradelanes.co

Methodology

Sustained Churn cohort definition. Voyages with paired ERD and CY Cut activity (co-moving) form the umbrella cohort under the v0.4 voyage-level pattern taxonomy. Sub-modes are named in the framework appendix. Classification thresholds are not published.

Carrier-terminal suppression. A pipeline-level artifact in CY Cut revisions for Maersk vessels at Norfolk's primary container terminal - first identified in Safe-to-Gate Monthly Dec 2025 - remains active and is more pronounced in May 2026's residual data. MAEU/USORF voyages appearing in residual top-activity lists with CY Cut activity that does not reflect underlying terminal behavior are excluded from concentration findings, repeat-offender findings, and Norfolk band assignments in this edition. A fuller treatment will appear once engineering review resolves the underlying data.

Cohort window. May 2026 U.S. departures with observable schedule revisions. 4,579 events across 701 voyages. The full cohort definition and classification thresholds are maintained internally and held consistent across publications.