ETA accuracy benchmark · 🇩🇪 Hamburg

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

0.90 h

MAE 0.66 h · 987 voyages

Broadcast ETA (RMSE)

53.73 h

crew-declared, 233 voyages (same set)

Our advantage

98.3 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (407 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


255916314	container	10/12, 05:23 AM UTC	-0.11 h	2.11 h
211149890	product	10/12, 05:20 AM UTC	0.03 h	—
211518720	container	10/12, 05:17 AM UTC	-0.21 h	-995.30 h
211510990	container	10/12, 05:08 AM UTC	0.29 h	3354.85 h
311001911	container	10/12, 05:05 AM UTC	1.85 h	-1.09 h
209276000	general-cargo	10/12, 05:01 AM UTC	1.30 h	-22.03 h
229630000	container	10/12, 04:58 AM UTC	0.30 h	16.02 h
211178650	product	10/12, 04:58 AM UTC	-0.13 h	-488.97 h
218172000	container	10/12, 04:33 AM UTC	0.18 h	-48.06 h
211741210	container	10/12, 04:31 AM UTC	0.91 h	—
211513140	container	10/12, 04:21 AM UTC	1.15 h	—
563296400	container	10/12, 04:11 AM UTC	0.82 h	-2262.19 h
211888150	product	10/12, 03:57 AM UTC	-0.32 h	—
209276000	general-cargo	10/12, 03:46 AM UTC	-0.35 h	-20.78 h
211201080	product	10/12, 03:44 AM UTC	0.02 h	-81.38 h
636025672	general-cargo	10/12, 03:37 AM UTC	0.66 h	-26.62 h
211187040	container	10/12, 01:00 AM UTC	1.00 h	3527.99 h
305994000	container	10/12, 12:38 AM UTC	-0.18 h	701.36 h
211196020	product	10/12, 12:37 AM UTC	1.79 h	5303.38 h
255805899	general-cargo	10/12, 12:03 AM UTC	1.53 h	-25.06 h
211319650	chemical	10/12, 12:03 AM UTC	0.58 h	—
211196020	product	10/11, 11:31 PM UTC	0.23 h	5304.47 h
636025672	general-cargo	10/11, 11:29 PM UTC	0.48 h	-22.49 h
305654000	container	10/11, 11:21 PM UTC	0.29 h	-370.85 h
255916314	container	10/11, 10:35 PM UTC	1.99 h	-115.59 h
209276000	general-cargo	10/11, 10:10 PM UTC	0.51 h	-15.17 h
211468560	product	10/11, 09:28 PM UTC	0.56 h	—
229630000	container	10/11, 09:05 PM UTC	1.38 h	-65.09 h
305994000	container	10/11, 08:35 PM UTC	1.31 h	705.41 h
215949000	container	10/11, 08:21 PM UTC	0.48 h	-7.10 h
215949000	container	10/11, 08:07 PM UTC	-0.15 h	-2363.12 h
209276000	general-cargo	10/11, 07:23 PM UTC	-0.36 h	-12.39 h
211513140	container	10/11, 07:08 PM UTC	0.61 h	—
211789680	product	10/11, 06:56 PM UTC	-0.48 h	—
211201080	product	10/11, 06:49 PM UTC	-0.30 h	-72.46 h
211319650	chemical	10/11, 06:42 PM UTC	1.08 h	—
211283460	product	10/11, 06:40 PM UTC	-0.12 h	-30.68 h
636025672	general-cargo	10/11, 06:37 PM UTC	-0.03 h	—
211468560	product	10/11, 05:48 PM UTC	0.41 h	-91.14 h
229424000	product	10/11, 05:44 PM UTC	1.20 h	-2721.74 h
229375000	container	10/11, 05:42 PM UTC	-0.11 h	-2895.22 h
211510990	container	10/11, 05:40 PM UTC	0.65 h	3366.32 h
563339300	container	10/11, 04:56 PM UTC	0.00 h	-8.94 h
209276000	general-cargo	10/11, 04:46 PM UTC	-0.64 h	-9.78 h
211774670	product	10/11, 04:44 PM UTC	-0.32 h	—
244670800	container	10/11, 04:29 PM UTC	-0.08 h	—
211783290	product	10/11, 04:25 PM UTC	0.11 h	1169.30 h
255805899	general-cargo	10/11, 04:19 PM UTC	0.17 h	-17.32 h
236112746	product	10/11, 04:16 PM UTC	0.08 h	-29.27 h
211188880	product	10/11, 04:04 PM UTC	0.50 h	3980.16 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.