ETA accuracy benchmark · 🇺🇸 Sabine Pass
How Port Flow's predicted ETA compares to each vessel's own broadcast ETA at Sabine Pass, measured on closed voyages over the last 90 days. We publish the number — good or bad. Lower error is better.
Port Flow predicted ETA (RMSE)
15.80 h
MAE 6.28 h · 393 voyages
Broadcast ETA (RMSE)
50.64 h
crew-declared, 128 voyages (same set)
Our advantage
68.8 %
lower error than the broadcast ETA
Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (88 excluded). Both figures are computed on the same remaining voyage set — apples to apples.
Closed voyages
error = predicted/broadcast minus actual arrival (hours)| MMSI | Cargo | Arrival | Our error | Broadcast error |
|---|---|---|---|---|
| 310800000 | product | 07/12, 11:38 AM UTC | -0.74 h | -6.63 h |
| 538010480 | container | 07/11, 10:23 PM UTC | 4.49 h | -41.37 h |
| 367561320 | product | 07/11, 05:25 PM UTC | -0.39 h | -981.43 h |
| 367654230 | container | 07/11, 05:01 PM UTC | -0.47 h | — |
| 563121700 | product | 07/11, 01:32 PM UTC | 0.11 h | -8.53 h |
| 256647000 | product | 07/11, 02:03 AM UTC | 0.39 h | -391.05 h |
| 366943760 | lng | 07/10, 02:35 PM UTC | 0.25 h | 1766.40 h |
| 636025023 | product | 07/10, 09:55 AM UTC | 4.74 h | -1058.93 h |
| 372101000 | product | 07/10, 07:58 AM UTC | 4.07 h | -38.97 h |
| 355658000 | product | 07/10, 07:43 AM UTC | 0.27 h | -2.70 h |
| 367400840 | container | 07/10, 05:55 AM UTC | -0.63 h | -191.93 h |
| 241593000 | product | 07/10, 02:52 AM UTC | 4.17 h | -13.88 h |
| 366273000 | product | 07/09, 10:56 PM UTC | 7.12 h | -227.94 h |
| 366943760 | lng | 07/09, 10:22 PM UTC | -0.55 h | 1782.62 h |
| 366943760 | lng | 07/09, 09:04 PM UTC | 3.25 h | 1783.92 h |
| 367390890 | container | 07/09, 06:11 PM UTC | -0.40 h | 6673.80 h |
| 366943760 | lng | 07/09, 05:54 PM UTC | 2.09 h | 1787.09 h |
| 366943760 | lng | 07/09, 04:30 PM UTC | -4.61 h | 1788.49 h |
| 367390890 | container | 07/09, 04:02 PM UTC | -87.32 h | 6675.95 h |
| 311001308 | product | 07/09, 03:03 PM UTC | 2.26 h | — |
| 367561320 | product | 07/09, 02:43 PM UTC | 1.05 h | — |
| 228464700 | product | 07/09, 02:00 PM UTC | 2.14 h | -687.01 h |
| 367561320 | product | 07/09, 12:55 PM UTC | 6.41 h | — |
| 367561320 | product | 07/09, 12:43 PM UTC | 5.05 h | -928.72 h |
| 367561320 | product | 07/09, 12:07 PM UTC | 6.59 h | — |
| 367561320 | product | 07/09, 12:01 PM UTC | 5.90 h | — |
| 367654230 | container | 07/09, 08:21 AM UTC | 2.41 h | -5.87 h |
| 215503000 | product | 07/09, 08:00 AM UTC | -0.95 h | -17.01 h |
| 538008568 | container | 07/09, 07:20 AM UTC | 8.87 h | -30.33 h |
| 367400840 | container | 07/09, 05:27 AM UTC | 2.46 h | -167.46 h |
| 366943760 | lng | 07/08, 11:01 PM UTC | 0.46 h | 1805.97 h |
| 367654230 | container | 07/08, 08:22 PM UTC | -0.55 h | 0.13 h |
| 366943760 | lng | 07/08, 07:53 PM UTC | 3.79 h | — |
| 259082000 | product | 07/08, 01:30 PM UTC | -1.81 h | -2.51 h |
| 367584550 | product | 07/08, 12:10 PM UTC | 5.38 h | — |
| 366943760 | lng | 07/08, 11:24 AM UTC | 1.46 h | 1817.59 h |
| 367584550 | product | 07/08, 10:43 AM UTC | 0.78 h | 949.91 h |
| 255915876 | product | 07/08, 04:52 AM UTC | 7.05 h | -544.88 h |
| 367400840 | container | 07/07, 06:42 PM UTC | 0.75 h | -337.71 h |
| 366943760 | lng | 07/07, 02:55 PM UTC | 0.44 h | — |
| 366943760 | lng | 07/07, 12:18 PM UTC | 3.52 h | — |
| 367561320 | product | 07/07, 11:55 AM UTC | 5.52 h | — |
| 366943760 | lng | 07/07, 11:33 AM UTC | 0.45 h | 1841.44 h |
| 241415000 | product | 07/07, 08:36 AM UTC | 0.70 h | -21.60 h |
| 366943760 | lng | 07/07, 03:05 AM UTC | 0.28 h | 1849.92 h |
| 311001308 | product | 07/07, 02:24 AM UTC | 3.19 h | -70.41 h |
| 352004338 | product | 07/06, 07:45 AM UTC | 3.48 h | -23.76 h |
| 366943760 | lng | 07/06, 06:12 AM UTC | -1.56 h | 1870.80 h |
| 538009405 | product | 07/06, 12:45 AM UTC | 0.37 h | -420.75 h |
| 232027158 | product | 07/05, 08:36 PM UTC | 8.10 h | -49.10 h |
See it on a live desk
Predicted ETAs, congestion and sanctions screening across 51 ports — try it in one click, no signup.
How it's measured
- Predicted ETA: distance-to-berth over speed-made-good, refreshed on every AIS position, with a seasonal correction from rolling medians.
- Broadcast ETA: the crew-entered ETA in the vessel's AIS message — the de-facto industry default.
- Both are compared to the actual arrival timestamp on the same closed voyages. We report RMSE (penalises large misses) and MAE (typical error).
- Broadcast ETAs more than 7 days off are dropped as AIS sentinel / placeholder values, and our model is scored on that same set — so the comparison is honest, not cherry-picked.
- Coverage is public AIS only; ports with weak coverage show fewer closed voyages. We'd rather show a small honest sample than a big misleading one.