ETA accuracy benchmark · 🇩🇪 Bremerhaven

How Port Flow's predicted ETA compares to each vessel's own broadcast ETA at Bremerhaven, 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)

6.82 h

MAE 1.29 h · 984 voyages

Broadcast ETA (RMSE)

53.99 h

crew-declared, 390 voyages (same set)

Our advantage

87.4 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (292 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


211523910	product	10/12, 05:38 AM UTC	-0.37 h	-24.65 h
636024193	container	10/12, 05:18 AM UTC	0.87 h	-2820.80 h
440252000	container	10/12, 01:54 AM UTC	0.38 h	—
440252000	container	10/12, 01:12 AM UTC	0.02 h	-3.71 h
352929000	container	10/12, 12:03 AM UTC	0.25 h	23.94 h
219863000	container	10/11, 11:49 PM UTC	2.63 h	-2289.32 h
231860000	container	10/11, 11:47 PM UTC	0.88 h	-442.81 h
352929000	container	10/11, 11:35 PM UTC	0.95 h	24.40 h
211189800	product	10/11, 10:58 PM UTC	0.08 h	—
311995000	container	10/11, 10:28 PM UTC	0.55 h	—
211189800	chemical	10/11, 10:16 PM UTC	-0.20 h	-136.28 h
255915667	container	10/11, 10:11 PM UTC	1.44 h	-138.20 h
311995000	container	10/11, 09:44 PM UTC	1.14 h	-10.74 h
248131000	container	10/11, 07:56 PM UTC	0.17 h	-4.45 h
248131000	container	10/11, 06:54 PM UTC	-0.14 h	-3.42 h
244110566	container	10/11, 06:51 PM UTC	-0.21 h	-2323.86 h
255805594	general-cargo	10/11, 06:40 PM UTC	1.77 h	-687.68 h
248131000	container	10/11, 06:26 PM UTC	0.04 h	-109.44 h
431190000	container	10/11, 05:19 PM UTC	0.76 h	1422.69 h
211428750	product	10/11, 05:11 PM UTC	0.40 h	—
211428750	product	10/11, 04:55 PM UTC	0.47 h	2089.17 h
211428750	product	10/11, 04:07 PM UTC	0.57 h	—
211287870	product	10/11, 04:01 PM UTC	0.03 h	-79.63 h
563277500	container	10/11, 02:43 PM UTC	1.56 h	-814.22 h
211428750	product	10/11, 12:39 PM UTC	-0.54 h	—
219678000	container	10/11, 12:15 PM UTC	1.97 h	-498.26 h
211480540	product	10/11, 11:19 AM UTC	0.12 h	—
211428750	product	10/11, 11:09 AM UTC	0.33 h	—
211812980	product	10/11, 10:51 AM UTC	-0.08 h	—
211812980	product	10/11, 10:33 AM UTC	0.14 h	-1030.55 h
211480540	product	10/11, 08:46 AM UTC	1.82 h	—
636093117	container	10/11, 08:31 AM UTC	0.39 h	-1658.53 h
211287870	product	10/11, 08:04 AM UTC	-0.02 h	—
211812980	product	10/11, 07:58 AM UTC	3.33 h	-1027.98 h
211287870	product	10/11, 07:13 AM UTC	0.23 h	—
211812980	product	10/11, 07:05 AM UTC	0.04 h	—
211428750	product	10/11, 05:35 AM UTC	-0.06 h	2100.51 h
305654000	container	10/11, 05:07 AM UTC	0.38 h	-160.12 h
211428750	product	10/11, 05:05 AM UTC	0.12 h	—
311001547	container	10/11, 04:58 AM UTC	0.42 h	-4.48 h
255806514	container	10/11, 04:42 AM UTC	2.02 h	—
563308400	container	10/11, 04:24 AM UTC	0.87 h	-2024.40 h
311001547	container	10/11, 04:12 AM UTC	0.82 h	-744.21 h
255805907	container	10/11, 12:31 AM UTC	0.74 h	—
477249500	container	10/10, 10:42 PM UTC	0.68 h	0.30 h
211287870	product	10/10, 09:43 PM UTC	0.39 h	-61.33 h
255916314	container	10/10, 09:21 PM UTC	1.54 h	-403.36 h
255806514	container	10/10, 07:31 PM UTC	1.51 h	-215.52 h
219543000	container	10/10, 06:45 PM UTC	1.18 h	-130.26 h
218627000	general-cargo	10/10, 06:08 PM UTC	0.50 h	12.86 h

See it on a live desk

Predicted ETAs, congestion and sanctions screening across 51 ports — try it in one click, no signup.

Open dashboard See pricing →

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.