ETA accuracy benchmark · 🇮🇹 Trieste

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

15.77 h

MAE 22.13 h · 787 voyages

Broadcast ETA (RMSE)

62.30 h

crew-declared, 473 voyages (same set)

Our advantage

74.7 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (285 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


357244000	container	10/12, 04:26 AM UTC	1.78 h	0.57 h
241963000	product	10/11, 11:34 PM UTC	1.06 h	-15.58 h
636021153	container	10/11, 07:16 PM UTC	1.87 h	-12.28 h
636025297	container	10/11, 05:43 PM UTC	-0.14 h	-589.73 h
249087000	container	10/11, 05:41 PM UTC	0.22 h	-991.69 h
314001131	container	10/11, 04:05 PM UTC	2.93 h	-0.09 h
253000119	container	10/11, 03:08 PM UTC	3.75 h	-368.65 h
538006499	product	10/11, 10:30 AM UTC	3.07 h	-1.50 h
636020783	container	10/11, 03:53 AM UTC	0.20 h	-0.88 h
563221900	container	10/11, 01:20 AM UTC	2.00 h	-610.84 h
255915842	general-cargo	10/10, 11:46 PM UTC	3.57 h	-494.78 h
636021506	container	10/10, 11:16 PM UTC	1.58 h	-391.27 h
538009721	container	10/10, 09:19 PM UTC	3.21 h	-404.33 h
636020455	container	10/10, 02:11 PM UTC	0.95 h	-154.20 h
271050376	general-cargo	10/10, 09:46 AM UTC	0.05 h	-168.78 h
636022626	product	10/10, 09:17 AM UTC	-0.64 h	-122.30 h
247383800	general-cargo	10/10, 08:40 AM UTC	-0.52 h	-182.67 h
256566000	container	10/10, 08:07 AM UTC	2.08 h	1.88 h
215958000	product	10/10, 06:56 AM UTC	0.37 h	-2.94 h
271046489	general-cargo	10/10, 05:14 AM UTC	-1.47 h	-85.00 h
241456000	product	10/09, 08:31 PM UTC	0.80 h	-11.53 h
271044633	general-cargo	10/09, 08:13 PM UTC	3.97 h	-184.23 h
574005750	chemical	10/09, 04:34 PM UTC	1.24 h	1.43 h
636021517	container	10/09, 04:18 PM UTC	-0.61 h	-636.32 h
246519000	container	10/09, 03:13 PM UTC	-1.97 h	-0.73 h
255815000	general-cargo	10/09, 07:58 AM UTC	-0.90 h	-136.97 h
511101667	container	10/09, 12:07 AM UTC	3.06 h	-1419.13 h
477155200	container	10/08, 10:37 PM UTC	3.53 h	-1550.20 h
215020000	general-cargo	10/08, 08:58 PM UTC	1.82 h	-77.48 h
248046000	container	10/08, 07:44 PM UTC	2.51 h	-26.73 h
255806043	container	10/08, 05:35 PM UTC	4.58 h	-359.59 h
642122016	product	10/08, 03:45 PM UTC	0.63 h	-21.25 h
371024000	container	10/08, 02:28 PM UTC	0.79 h	-23.48 h
271000662	general-cargo	10/08, 02:02 PM UTC	1.17 h	-8.78 h
563260800	product	10/08, 12:28 PM UTC	2.64 h	-2.48 h
256263000	container	10/08, 12:19 PM UTC	-1.36 h	—
352001191	container	10/08, 12:04 PM UTC	-1.24 h	-484.08 h
636021246	container	10/08, 10:03 AM UTC	0.36 h	-3048.05 h
256263000	general-cargo	10/08, 06:54 AM UTC	1.13 h	-83.90 h
271000662	general-cargo	10/08, 05:53 AM UTC	-0.65 h	-82.88 h
247453500	general-cargo	10/08, 04:43 AM UTC	0.82 h	-192.47 h
636020244	container	10/07, 06:21 PM UTC	2.85 h	-0.85 h
271043165	general-cargo	10/07, 02:41 PM UTC	-0.29 h	-75.69 h
538009489	container	10/07, 01:50 PM UTC	0.12 h	-325.83 h
215515000	product	10/07, 12:49 PM UTC	1.44 h	-6.82 h
271043303	general-cargo	10/07, 07:23 AM UTC	0.24 h	-89.39 h
311002132	container	10/07, 07:03 AM UTC	3.84 h	-1.56 h
247241900	general-cargo	10/07, 06:46 AM UTC	1.55 h	-90.77 h
636092914	container	10/07, 06:12 AM UTC	1.91 h	-617.21 h
311013400	product	10/06, 08:12 PM UTC	0.94 h	-383.21 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.