What we tested

Across 4,541 export port-calls with sufficient snapshot coverage, we examined:

• Whether ERDs change inside the final 72 hours
• How often CY cutoffs change late
• Where risk shows up when ERDs do hold

Ports in scope include USCHS, USEWR, USHOU, USLAX, USLGB, with East vs West Coast comparisons where applicable.

What the updated data shows

ERDs do freeze more often - but not completely

On the East Coast:

• 12.4% of port-calls still experience a late-stage ERD change
• 80.1% of port-calls have a fully stable receiving window (no late ERD or CY change)

On the West Coast (for comparison):

• 31.9% of port-calls see a late-stage ERD change
• Only 55.6% have a fully stable window

Freeze policies materially reduce early volatility - but do not eliminate execution risk.

Figure 1 - Receiving Window Integrity by Port

Plain English

ERDs are much more stable on the East Coast than elsewhere - but 1 in 8 still move after teams think they’re locked.

What this feels like

You’re safer than before.
You’re not safe enough to relax.

How big are late ERD moves when they happen?

Across 1,315 late-stage ERD change events:

• 63% move 1-3 days
• 17% move 7+ days
• Median drift: 1 day
• 75th percentile: 4 days
• 90th percentile: 11 days

So while many late changes are small, the tail is long and operationally brutal.

Figure 2 - Late-Stage ERD Drift Distribution

Plain English

Most late changes are annoying.
Enough are catastrophic that exporters have to plan as if they will happen.

What this feels like

You can survive a one-day slip.
A four-day slip means everything you staged now costs money.
An eleven-day slip means explaining misses you didn’t cause.

Where the risk goes when ERDs hold

This is the most important update - and the pattern is clear.

Even when ERDs hold, CY cutoffs often don’t.

Example: Charleston (USCHS)

• ERD stability: 89.1%
• CY stability: 94.9%
• Fully stable window: 84.1%

Charleston looks “reliable” by ERD metrics - and mostly is - but late-stage CY compression still shows up.

Los Angeles / Long Beach contrast

• USLAX stable windows: 75.2%
• USLGB stable windows: 65.7%
• CY late-stage changes dominate instability at USLGB (29.6%)

Figure 3 - Receiving Window Integrity by Port
Stable vs ERD-only breaks vs CY-only breaks vs both

Plain English

Freeze policies don’t remove uncertainty.
They push it downstream.

What this feels like

You stop worrying about when you can start.
You start worrying about how fast you have to finish.

The shift in failure mode (now quantified)

Using stability thresholds (≥80% = “high”):

• All ports cluster toward high ERD stability.

• CY stability diverges
• USCHS / USEWR / USHOU: high-high quadrant
• USLGB: high ERD, low CY stability

This is not random noise.

It is system behavior under constraint.

Figure 4 - ERD vs CY Stability Shift Matrix
Dot size = volume; color = CY late-stage change rate

Plain English

When one layer becomes rigid, another layer absorbs the shock.

What this feels like

Everything looks fine - until suddenly it’s all urgent.

East vs West: what freeze policies actually do

Policy delta

Note: Stable window percentages vary by scope. Earlier figures reference East Coast ports only; the table below reflects the full multi-port comparison set.

* West Coast drift inflated by extreme tail events - this is why distribution matters more than averages.

Figure 5 - Freeze Policy Effect: East vs West

Plain English

Policy improves one metric.
The experience is shaped by the weakest remaining link.

What this feels like

The dashboard says “better.”
Your week still breaks.

Why averages still mislead exporters

• Median ERD drift looks manageable
• 90th-percentile drift breaks plans
• CY volatility explains why “good ERDs” still lead to misses

Late-stage changes aren’t just changes.

They are decision traps.

Reliability isn’t one promise holding

Reliability is not:

• “ERDs froze”
• “Policy worked”
• “The average improved”

Reliability is:

Can exporters commit resources without being punished for trusting the system?

The data says: sometimes - not consistently.

Figure 6 - Receiving Window Failure Timeline
Decision lock-ins vs late-stage risk zones

Plain English

Exporters don’t fail schedules.
Schedules fail exporters - late, quietly, and expensively - even when everyone follows the rules.

What this feels like

You did everything right.
The system moved after you committed.

Bottom line

• ERDs mostly freeze on the East Coast
• Late changes are rarer - but still consequential
• CY cutoffs inherit volatility when ERDs stabilize
• Reliability must be measured where decisions become irreversible

That’s what the cargo receiving window lens reveals - and once you see it, the pattern is hard to unsee.