ETA accuracy benchmark · 🇨🇴 Cartagena
How Port Flow's predicted ETA compares to each vessel's own broadcast ETA at Cartagena, measured on closed voyages over the last 30 days. We publish the number — good or bad. Lower error is better.
Port Flow predicted ETA (RMSE)
113.48 h
MAE 25.35 h · 159 voyages
Broadcast ETA (RMSE)
Building…
only 2 comparable voyages so far (need 20)
Gap
—
benchmark builds as more voyages close
Closed voyages
error = predicted/broadcast minus actual arrival (hours)| 730001308 | container | 10/07, 02:24 PM UTC | 0.86 h | — |
| 245980000 | container | 10/07, 02:14 PM UTC | 0.17 h | — |
| 677026300 | general-cargo | 10/07, 02:09 PM UTC | 4.12 h | — |
| 538009817 | container | 10/07, 12:06 PM UTC | -0.70 h | -1045.10 h |
| 636023410 | container | 10/07, 11:49 AM UTC | -0.87 h | — |
| 538002733 | container | 10/07, 11:05 AM UTC | -0.00 h | -245.10 h |
| 219830000 | product | 10/07, 10:19 AM UTC | 0.30 h | -621.32 h |
| 636014459 | container | 10/07, 10:07 AM UTC | 0.31 h | — |
| 636092798 | container | 10/07, 06:05 AM UTC | 1.31 h | — |
| 248924000 | container | 10/07, 01:36 AM UTC | 0.02 h | -282.10 h |
| 636022175 | container | 10/07, 12:08 AM UTC | -0.23 h | -283.12 h |
| 538005418 | product | 10/06, 09:50 PM UTC | 2.02 h | — |
| 248649000 | container | 10/06, 04:18 PM UTC | 0.34 h | — |
| 256703000 | general-cargo | 10/06, 02:54 PM UTC | 1.50 h | — |
| 232006297 | product | 10/06, 01:13 PM UTC | -0.90 h | — |
| 636023371 | container | 10/06, 11:54 AM UTC | -0.00 h | -484.41 h |
| 636015945 | product | 10/06, 10:57 AM UTC | 4.89 h | — |
| 636093082 | container | 10/06, 05:31 AM UTC | 0.14 h | — |
| 538010081 | container | 10/06, 04:56 AM UTC | 0.34 h | — |
| 256612000 | container | 10/06, 02:39 AM UTC | 0.32 h | — |
| 229914000 | container | 10/05, 10:09 PM UTC | -0.36 h | — |
| 565843000 | product | 10/05, 10:05 PM UTC | 2.53 h | -1045.09 h |
| 305904000 | container | 10/05, 07:54 PM UTC | 1.22 h | — |
| 255917226 | dry-bulk | 10/05, 06:12 PM UTC | 3.72 h | — |
| 730152039 | container | 10/05, 04:52 PM UTC | -0.36 h | — |
| 563281200 | general-cargo | 10/05, 03:38 PM UTC | 1.63 h | — |
| 304954000 | container | 10/05, 01:54 PM UTC | 1.81 h | — |
| 212814000 | general-cargo | 10/05, 09:28 AM UTC | -0.31 h | — |
| 218854000 | general-cargo | 10/05, 07:03 AM UTC | -7.06 h | -450.06 h |
| 538009616 | product | 10/05, 06:45 AM UTC | 0.26 h | — |
| 636026414 | general-cargo | 10/05, 05:36 AM UTC | -0.11 h | -226.61 h |
| 636021427 | container | 10/05, 01:24 AM UTC | 0.70 h | -294.17 h |
| 219785000 | container | 10/05, 12:36 AM UTC | 0.06 h | -356.11 h |
| 636013522 | container | 10/05, 12:12 AM UTC | -0.05 h | -345.71 h |
| 538012476 | product | 10/04, 07:20 PM UTC | -0.99 h | — |
| 372981000 | container | 10/04, 06:39 PM UTC | 1.13 h | -356.66 h |
| 564050000 | container | 10/04, 03:07 PM UTC | 0.49 h | — |
| 215785000 | product | 10/04, 12:55 PM UTC | 2.21 h | — |
| 730151726 | product | 10/04, 12:34 PM UTC | -1.21 h | — |
| 256948000 | container | 10/04, 09:14 AM UTC | -0.13 h | -917.24 h |
| 538008703 | product | 10/04, 08:39 AM UTC | 3.90 h | — |
| 636024483 | container | 10/04, 08:28 AM UTC | 0.17 h | — |
| 563139800 | container | 10/04, 07:28 AM UTC | 2.13 h | — |
| 636018194 | container | 10/04, 05:24 AM UTC | 0.41 h | — |
| 564050000 | container | 10/03, 08:49 PM UTC | 0.24 h | — |
| 636093155 | container | 10/03, 08:32 PM UTC | 1.45 h | — |
| 255806154 | product | 10/03, 02:41 PM UTC | 1.29 h | — |
| 247341500 | container | 10/03, 02:22 PM UTC | 3.02 h | — |
| 229573000 | container | 10/03, 11:44 AM UTC | -0.07 h | -162.74 h |
| 636021923 | product | 10/03, 10:44 AM UTC | 2.60 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.