ETA accuracy benchmark · 🇺🇸 Los Angeles

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

56.38 h

MAE 7.11 h · 995 voyages

Broadcast ETA (RMSE)

40.87 h

crew-declared, 540 voyages (same set)

Gap

38.0 %

higher error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (96 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)

MMSI	Cargo	Arrival	Our error	Broadcast error
352002754	container	07/12, 07:36 AM UTC	0.76 h	-2.12 h
367646810	product	07/12, 01:55 AM UTC	-2.54 h	-2.92 h
563109900	product	07/12, 01:01 AM UTC	2.28 h	1.97 h
367646810	product	07/11, 09:56 PM UTC	0.02 h	1.06 h
538012661	chemical	07/11, 06:45 PM UTC	0.45 h	-17.75 h
354004000	container	07/11, 04:59 PM UTC	1.08 h	-781.98 h
367353080	product	07/11, 03:42 PM UTC	1.24 h	4160.30 h
369703000	product	07/11, 03:29 PM UTC	-0.66 h	-190.49 h
338881000	container	07/11, 02:36 PM UTC	-0.89 h	-762.61 h
367353080	product	07/11, 02:21 PM UTC	-47.66 h	4161.65 h
538008412	product	07/11, 01:19 PM UTC	-22.96 h	-37.82 h
636018038	general-cargo	07/11, 01:15 PM UTC	-0.41 h	-1.26 h
636093436	container	07/11, 01:12 PM UTC	-4.22 h	-8.70 h
563222300	container	07/11, 11:17 AM UTC	-0.57 h	-1.29 h
367719640	container	07/11, 09:55 AM UTC	-15.99 h	-20.93 h
370024000	container	07/11, 09:44 AM UTC	0.07 h	-82.74 h
538010889	product	07/11, 08:08 AM UTC	2.10 h	-308.15 h
305524000	container	07/11, 01:23 AM UTC	0.61 h	—
563219700	container	07/10, 09:49 PM UTC	1.51 h	-707.82 h
369040000	product	07/10, 06:39 PM UTC	0.41 h	96.35 h
352003169	container	07/10, 02:43 PM UTC	2.00 h	-0.73 h
477183900	container	07/10, 01:14 PM UTC	1.33 h	-1.23 h
431009000	container	07/10, 12:39 PM UTC	0.85 h	-1.66 h
431009000	container	07/10, 12:21 PM UTC	0.05 h	-1.37 h
255994000	container	07/10, 12:03 PM UTC	0.38 h	433.94 h
431009000	container	07/10, 11:38 AM UTC	-0.10 h	-0.63 h
368305000	container	07/10, 08:44 AM UTC	1.57 h	-199.74 h
538003274	container	07/10, 02:56 AM UTC	-14.14 h	221.06 h
431296000	general-cargo	07/09, 11:09 PM UTC	0.41 h	-1.16 h
368127000	product	07/09, 08:29 PM UTC	1.02 h	-47.50 h
636025014	container	07/09, 07:54 PM UTC	1.17 h	-0.91 h
416464000	container	07/09, 07:02 PM UTC	-0.04 h	-1.04 h
356466000	product	07/09, 06:36 PM UTC	0.34 h	-7.61 h
367719640	container	07/09, 05:30 PM UTC	-0.75 h	-29.67 h
369040000	product	07/09, 05:19 PM UTC	0.70 h	-190.32 h
229957000	product	07/09, 03:32 PM UTC	2.00 h	-1135.53 h
441502000	general-cargo	07/09, 02:50 PM UTC	-0.19 h	-48.84 h
538004619	container	07/09, 02:02 PM UTC	0.73 h	-1.04 h
477067900	container	07/09, 01:37 PM UTC	0.52 h	-600.62 h
368237190	container	07/09, 01:35 PM UTC	0.91 h	—
477637600	container	07/09, 01:21 PM UTC	3.16 h	-927.36 h
368237190	container	07/09, 12:44 PM UTC	-0.02 h	-144.74 h
636022332	container	07/09, 12:00 PM UTC	1.96 h	-1.01 h
636021470	container	07/09, 10:55 AM UTC	0.08 h	743.08 h
538008585	container	07/09, 09:25 AM UTC	1.92 h	-105.43 h
305524000	container	07/09, 08:37 AM UTC	0.57 h	-6.62 h
367719640	container	07/09, 08:07 AM UTC	-9.29 h	-41.21 h
538004934	product	07/09, 06:56 AM UTC	1.73 h	-27.94 h
538007896	product	07/09, 12:53 AM UTC	0.62 h	-11.88 h
538006131	product	07/08, 11:03 PM UTC	0.07 h	-26.05 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.