Source Precedence for Conflicting Timestamps
Different sources routinely report different times for the same event. Two carriers on the same vessel will give different ETDs; a terminal and a carrier will confirm the same discharge minutes or hours apart; satellite (AIS) position data may imply a departure before any carrier reports it. This is normal - each source is built on different information and updates on its own rhythm.
This page explains which reading Dockflow lets win when they disagree, so a value on your dashboard can be traced back to a rule rather than looking arbitrary.
The precedence rules
When multiple readings describe the same event, Dockflow resolves them in this order:
| Priority | Rule | Effect |
|---|---|---|
| 1 | Actual beats estimate | If any source confirms an event actually happened, that confirmation wins over every estimate, regardless of source. An estimate never overrides a confirmed actual. |
| 2 | More recent beats older | Among estimates (when no actual exists yet), a more recent reading carries more weight than an older one - the latest information supersedes stale predictions. |
| 3 | Agreement raises confidence | The more independent sources that agree on a time, the higher the confidence in it. Isolated single-source readings carry less weight than a cluster that concurs. |
These rules apply both to a single tradeflow (where several sources report on your one booking) and to the per-vessel schedule (where readings are pooled across every shipment on the vessel).
Per-tradeflow vs per-vessel
The precedence rules are the same, but what they operate on differs depending on how you track:
| Per Tradeflow (default) | Per Vessel Schedule | |
|---|---|---|
| Readings considered | Only those on your booking | All readings for the vessel, across all shipments |
| Result | Exactly what the sources on your tradeflow reported, resolved by the rules above | A single consensus time per port call, combined across sources |
| Trade-off | Preserves carrier-specific detail | One clean schedule, but individual carrier estimates are not preserved |
If you need to see precisely what a specific carrier reported, per-tradeflow tracking is the right fit. If you want one consolidated vessel timeline and are comfortable with a consensus, per-vessel is the right fit. The full mechanism - cross-shipment pooling, voyage clustering, confidence scoring - is documented in Vessel Schedule Intelligence.
Guards against bad readings
Two safeguards stop a wrong reading from silently winning:
- Voyage separation. Vessels revisit the same port on later voyages. Readings at the same port more than 5 days apart are treated as separate voyages, so an old call cannot contaminate a new one.
- Speed plausibility. Port-to-port times that would require an implausible vessel speed are flagged for review rather than accepted blindly.
Vessel identification
Precedence only works if two readings are known to be about the same vessel. Dockflow matches vessels by a hierarchy of identifiers, most reliable first:
- MMSI (most reliable)
- IMO
- Vessel name (fallback only - names are reused and mistyped)
This prevents two different vessels with a similar name from being merged, and keeps readings for one vessel from being split across duplicate profiles.
What to do when a timestamp still looks wrong
- Check whether the value is an actual or an estimate - an estimate can still move.
- In the per-vessel view, remember you are seeing a consensus, not a single carrier's figure; switch to per-tradeflow to see the raw carrier reading.
- If a confirmed actual is genuinely wrong (contradicted by a later reliable source), that is a case worth reporting - see Anomaly classification, which flags several of these situations automatically.