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 7 days. We publish the number — good or bad. Lower error is better.
Port Flow predicted ETA (RMSE)
3.91 h
MAE 2.91 h · 35 voyages
Broadcast ETA (RMSE)
Building…
only 18 comparable voyages so far (need 20)
Gap
—
benchmark builds as more voyages close
Closed voyages
error = predicted/broadcast minus actual arrival (hours)| 367390890 | container | 10/11, 11:04 PM UTC | -0.29 h | 4412.93 h |
| 373569000 | container | 10/11, 10:50 PM UTC | 6.07 h | -85.84 h |
| 256647000 | product | 10/11, 02:30 PM UTC | 0.62 h | -422.50 h |
| 538009747 | product | 10/11, 06:11 AM UTC | 3.44 h | -665.26 h |
| 538009739 | product | 10/11, 04:22 AM UTC | 4.83 h | -31.38 h |
| 215890000 | product | 10/10, 08:28 PM UTC | -0.45 h | -15.46 h |
| 367391920 | container | 10/10, 06:57 PM UTC | -0.73 h | — |
| 367391920 | container | 10/10, 02:54 PM UTC | 8.31 h | -62.91 h |
| 367390890 | container | 10/10, 02:51 PM UTC | -0.72 h | 4445.14 h |
| 367134000 | product | 10/10, 09:26 AM UTC | 2.06 h | -202.44 h |
| 538005808 | product | 10/10, 07:08 AM UTC | 7.90 h | -34.15 h |
| 367391920 | container | 10/10, 04:42 AM UTC | 4.85 h | -52.71 h |
| 249342000 | product | 10/10, 12:23 AM UTC | 0.82 h | -22.40 h |
| 367390890 | container | 10/09, 04:18 PM UTC | -0.18 h | 4467.70 h |
| 367391920 | container | 10/09, 06:48 AM UTC | -0.38 h | -30.81 h |
| 367455580 | product | 10/09, 03:51 AM UTC | 5.45 h | -1710.85 h |
| 228464600 | product | 10/09, 12:25 AM UTC | -1.94 h | -19.41 h |
| 215754000 | product | 10/08, 04:31 AM UTC | -0.64 h | -17.53 h |
| 367391920 | container | 10/08, 03:17 AM UTC | 0.38 h | 1520.70 h |
| 419001940 | product | 10/08, 01:07 AM UTC | 6.35 h | -81.62 h |
| 311000953 | product | 10/07, 07:22 PM UTC | 5.01 h | 0.63 h |
| 538005271 | product | 10/07, 05:32 PM UTC | 1.31 h | -10.55 h |
| 366908640 | product | 10/07, 01:08 PM UTC | 0.21 h | — |
| 241981000 | product | 10/07, 08:04 AM UTC | 0.23 h | -1141.58 h |
| 538005536 | product | 10/07, 06:49 AM UTC | 7.81 h | -2388.83 h |
| 419001562 | product | 10/06, 08:07 PM UTC | 7.47 h | 67.87 h |
| 636018011 | product | 10/06, 08:00 PM UTC | -2.08 h | -33.02 h |
| 367710960 | chemical | 10/06, 03:16 PM UTC | 0.46 h | -436.27 h |
| 367710960 | chemical | 10/06, 02:26 PM UTC | 4.31 h | -435.44 h |
| 367710960 | chemical | 10/06, 02:02 PM UTC | 1.99 h | — |
| 367710960 | chemical | 10/06, 01:24 PM UTC | 3.24 h | -914.40 h |
| 538010087 | container | 10/06, 07:16 AM UTC | 5.96 h | -97.28 h |
| 241987000 | product | 10/05, 07:04 PM UTC | -0.74 h | -19.06 h |
| 636024983 | product | 10/05, 07:02 PM UTC | 1.72 h | -187.03 h |
| 259190000 | product | 10/05, 10:42 AM UTC | -2.82 h | -55.71 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.