ETA accuracy benchmark · 🇫🇷 Saint-Nazaire / Montoir

How Port Flow's predicted ETA compares to each vessel's own broadcast ETA at Saint-Nazaire / Montoir, 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)

1.94 h

MAE 17.02 h · 165 voyages

Broadcast ETA (RMSE)

60.57 h

crew-declared, 93 voyages (same set)

Our advantage

96.8 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (63 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


210283000	general-cargo	10/11, 08:35 PM UTC	0.25 h	-415.59 h
538005410	product	10/11, 03:14 PM UTC	0.59 h	0.27 h
249209000	chemical	10/11, 03:02 PM UTC	0.58 h	-105.04 h
255806464	container	10/11, 10:15 AM UTC	0.04 h	-200.25 h
226340000	product	10/10, 05:42 PM UTC	-1.43 h	-157.71 h
257089140	product	10/10, 04:05 PM UTC	0.17 h	-64.11 h
477345800	lng	10/10, 02:25 PM UTC	-0.27 h	-1.43 h
538012432	product	10/10, 10:51 AM UTC	-0.02 h	-2.86 h
305030000	container	10/09, 11:23 PM UTC	0.51 h	-6.89 h
224876000	container	10/09, 09:29 PM UTC	0.32 h	1.51 h
227681520	product	10/09, 05:55 PM UTC	0.22 h	-25.92 h
538008844	product	10/09, 03:40 PM UTC	0.99 h	—
231875000	product	10/08, 09:12 PM UTC	1.66 h	-13.21 h
538006120	product	10/07, 10:41 PM UTC	-13.02 h	-765.70 h
215345000	product	10/07, 02:55 PM UTC	-0.70 h	-30.93 h
305546000	container	10/07, 08:04 AM UTC	0.34 h	13.92 h
311000383	container	10/07, 07:03 AM UTC	0.66 h	-3.07 h
228020700	container	10/07, 06:34 AM UTC	0.34 h	-169.58 h
228084000	container	10/07, 04:43 AM UTC	-0.33 h	-818.22 h
538005406	product	10/07, 12:20 AM UTC	1.68 h	-146.33 h
230723000	container	10/06, 07:41 PM UTC	0.37 h	-1.70 h
305793000	product	10/06, 07:26 PM UTC	3.42 h	—
224876000	container	10/06, 07:21 PM UTC	0.78 h	2.64 h
305793000	product	10/06, 07:14 PM UTC	0.15 h	-218.24 h
257089140	product	10/06, 01:11 PM UTC	1.99 h	-25.20 h
636014199	container	10/06, 10:31 AM UTC	0.96 h	-4.52 h
227681520	product	10/06, 05:56 AM UTC	0.60 h	—
305229000	container	10/05, 06:46 PM UTC	0.26 h	-0.02 h
538008770	container	10/05, 05:39 PM UTC	1.27 h	-61.66 h
220264000	product	10/05, 12:56 PM UTC	-0.47 h	-257.94 h
255915615	container	10/05, 12:47 PM UTC	0.39 h	-2.30 h
477327900	lng	10/05, 11:06 AM UTC	1.13 h	-13.09 h
304070000	product	10/05, 06:00 AM UTC	2.30 h	-239.01 h
305864000	container	10/05, 05:52 AM UTC	1.06 h	-9.37 h
255806191	container	10/05, 03:11 AM UTC	-0.04 h	-121.19 h
255915770	container	10/04, 11:40 PM UTC	0.69 h	-806.68 h
236446000	product	10/04, 09:42 PM UTC	-0.23 h	-2.71 h
256784000	chemical	10/04, 09:32 PM UTC	-8.98 h	-247.54 h
210281000	container	10/04, 01:41 PM UTC	0.06 h	-329.69 h
226344000	product	10/04, 12:01 PM UTC	2.97 h	-799.28 h
636021960	product	10/04, 11:17 AM UTC	1.13 h	-66.80 h
636017059	product	10/03, 09:10 PM UTC	0.16 h	-121.17 h
210299000	product	10/03, 06:39 AM UTC	1.19 h	-44.66 h
224876000	container	10/02, 09:47 PM UTC	0.40 h	-34.80 h
219200000	chemical	10/02, 09:00 PM UTC	-2.32 h	-1997.01 h
305696000	container	10/02, 02:51 PM UTC	0.11 h	-303.86 h
227681520	product	10/02, 12:09 PM UTC	-0.11 h	-26.17 h
248156000	container	10/02, 06:14 AM UTC	-0.27 h	-493.24 h
226340000	product	10/01, 07:28 PM UTC	0.37 h	-161.98 h
636021960	product	10/01, 06:30 PM UTC	1.84 h	-2.00 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.