ETA accuracy benchmark · 🇲🇦 Tanger Med

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

17.84 h

MAE 6.26 h · 811 voyages

Broadcast ETA (RMSE)

54.30 h

crew-declared, 369 voyages (same set)

Our advantage

67.1 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (354 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


248096000	product	10/12, 05:26 AM UTC	-0.22 h	-517.45 h
636093333	container	10/12, 04:04 AM UTC	-1.62 h	-297.08 h
538007617	container	10/12, 02:51 AM UTC	-1.92 h	-988.86 h
219018897	chemical	10/12, 02:03 AM UTC	-187.01 h	-153.06 h
219767000	container	10/12, 01:04 AM UTC	-5.84 h	-11.07 h
249123000	container	10/12, 01:04 AM UTC	-1.15 h	-651.07 h
229340000	container	10/12, 12:48 AM UTC	-0.41 h	2902.69 h
256370000	product	10/11, 11:46 PM UTC	3.91 h	-2543.80 h
373026000	product	10/11, 11:43 PM UTC	-3.87 h	166.28 h
241298000	container	10/11, 09:59 PM UTC	-1.54 h	-755.99 h
224473460	product	10/11, 09:37 PM UTC	0.43 h	—
212777000	container	10/11, 07:23 PM UTC	-0.03 h	-496.40 h
253000159	product	10/11, 06:48 PM UTC	2.29 h	-280.81 h
212633000	product	10/11, 05:11 PM UTC	2.16 h	-20.69 h
563044300	container	10/11, 04:53 PM UTC	-2.73 h	-1127.89 h
255917253	product	10/11, 04:43 PM UTC	5.45 h	-1001.72 h
210707000	container	10/11, 03:11 PM UTC	-0.64 h	-33.19 h
636021776	container	10/11, 02:49 PM UTC	0.03 h	-452.83 h
314001074	container	10/11, 02:41 PM UTC	4.28 h	15.30 h
215402000	product	10/11, 02:35 PM UTC	-90.83 h	-136.59 h
224473460	product	10/11, 02:02 PM UTC	1.05 h	-235.04 h
373026000	product	10/11, 01:16 PM UTC	-37.50 h	-38.53 h
636023006	product	10/11, 12:02 PM UTC	1.09 h	-1984.05 h
636014059	container	10/11, 11:11 AM UTC	3.32 h	-318.19 h
538010212	product	10/11, 11:11 AM UTC	2.18 h	-2716.18 h
219506000	container	10/11, 09:56 AM UTC	3.93 h	-1169.95 h
636018094	container	10/11, 09:05 AM UTC	-0.49 h	-296.60 h
256453000	product	10/11, 08:21 AM UTC	5.00 h	-2234.36 h
219515000	container	10/11, 08:08 AM UTC	-0.99 h	-412.14 h
229933000	product	10/11, 07:29 AM UTC	6.12 h	-18.00 h
250002434	container	10/11, 06:17 AM UTC	6.88 h	-11.29 h
229933000	product	10/11, 06:14 AM UTC	7.13 h	-16.74 h
255805657	general-cargo	10/11, 04:25 AM UTC	-0.22 h	-2353.43 h
212633000	product	10/11, 03:55 AM UTC	-0.08 h	-28.93 h
636025912	product	10/11, 03:24 AM UTC	3.17 h	-2427.40 h
311000922	container	10/11, 03:23 AM UTC	0.15 h	-2.39 h
256485000	container	10/11, 02:05 AM UTC	-0.28 h	-1936.10 h
209314000	container	10/11, 01:21 AM UTC	4.94 h	-7.36 h
255805883	container	10/10, 11:55 PM UTC	-0.99 h	-443.92 h
636022873	container	10/10, 11:01 PM UTC	-0.57 h	-1.53 h
210707000	general-cargo	10/10, 10:52 PM UTC	-0.44 h	-10.88 h
563022600	product	10/10, 09:59 PM UTC	5.71 h	-11.99 h
229365000	product	10/10, 09:45 PM UTC	-0.19 h	-173.75 h
209825000	product	10/10, 09:18 PM UTC	-1.40 h	-41.30 h
577414000	container	10/10, 09:13 PM UTC	6.11 h	-12.22 h
636018701	container	10/10, 08:17 PM UTC	-1.54 h	-2.29 h
563220200	container	10/10, 07:39 PM UTC	-1.11 h	-1.15 h
577414000	container	10/10, 07:37 PM UTC	7.69 h	-541.63 h
636026315	product	10/10, 05:54 PM UTC	0.16 h	-10.91 h
236590000	product	10/10, 05:53 PM UTC	0.73 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.