ETA accuracy benchmark · 🇨🇳 Shenzhen

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

46.49 h

MAE 77.60 h · 628 voyages

Broadcast ETA (RMSE)

51.54 h

crew-declared, 222 voyages (same set)

Our advantage

9.8 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (355 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


563224700	general-cargo	10/12, 04:25 AM UTC	1.53 h	-404.42 h
477642100	container	10/12, 03:56 AM UTC	0.45 h	-457.45 h
413601030	container	10/12, 03:49 AM UTC	0.48 h	-2.32 h
412472910	container	10/12, 03:04 AM UTC	5.63 h	-23.07 h
256381000	container	10/12, 02:42 AM UTC	3.98 h	-114.71 h
413496010	container	10/12, 01:59 AM UTC	-4.62 h	-155.99 h
413601030	container	10/12, 01:52 AM UTC	-3.52 h	—
413270770	container	10/12, 01:43 AM UTC	7.05 h	—
413613980	container	10/12, 12:18 AM UTC	-0.90 h	-17.14 h
413054000	general-cargo	10/11, 10:55 PM UTC	10.34 h	-1200.92 h
538010654	container	10/11, 10:26 PM UTC	5.42 h	-296.45 h
566642000	container	10/11, 08:24 PM UTC	-9.97 h	—
412472910	container	10/11, 07:44 PM UTC	1.12 h	-15.75 h
636017346	container	10/11, 06:55 PM UTC	1.38 h	-870.93 h
413240560	container	10/11, 05:54 PM UTC	3.47 h	-99.24 h
563062600	container	10/11, 05:43 PM UTC	4.55 h	-329.70 h
477213400	container	10/11, 05:08 PM UTC	0.57 h	-2284.15 h
636022037	container	10/11, 05:04 PM UTC	-2.47 h	-397.07 h
255803450	container	10/11, 04:10 PM UTC	4.03 h	-3.17 h
538006520	container	10/11, 04:03 PM UTC	4.57 h	-2.05 h
538008845	container	10/11, 02:49 PM UTC	2.81 h	-6.82 h
636023534	container	10/11, 10:48 AM UTC	3.83 h	-2.81 h
413519910	container	10/11, 07:55 AM UTC	-1.93 h	0.07 h
413229070	container	10/11, 07:49 AM UTC	-142.57 h	-340.49 h
413443460	product	10/11, 07:02 AM UTC	2.52 h	-268.05 h
413900824	container	10/11, 06:43 AM UTC	-18.25 h	-2949.84 h
636021005	container	10/11, 05:18 AM UTC	4.36 h	1.69 h
413872002	container	10/11, 04:57 AM UTC	-8.76 h	-338.42 h
413973504	container	10/11, 03:59 AM UTC	-189.28 h	-2516.49 h
354690000	container	10/11, 03:44 AM UTC	-2.51 h	-836.25 h
636017882	container	10/11, 03:43 AM UTC	6.68 h	-2063.73 h
538011528	container	10/11, 03:07 AM UTC	-69.33 h	-75.10 h
538012625	container	10/11, 02:44 AM UTC	-1.80 h	-219.24 h
232018900	container	10/11, 01:40 AM UTC	0.68 h	-1130.68 h
563250800	container	10/11, 01:29 AM UTC	1.93 h	-2933.49 h
566828000	container	10/11, 12:13 AM UTC	-4.51 h	4.78 h
413901418	container	10/11, 12:04 AM UTC	-1924.75 h	-1320.58 h
412472910	container	10/10, 11:00 PM UTC	-3.52 h	-3032.00 h
477655800	general-cargo	10/10, 10:17 PM UTC	3.81 h	-2741.79 h
563235600	general-cargo	10/10, 09:29 PM UTC	-1.41 h	-780.50 h
413294260	container	10/10, 07:49 PM UTC	-9.52 h	-387.82 h
256441000	container	10/10, 07:29 PM UTC	6.17 h	-355.49 h
636026030	general-cargo	10/10, 05:48 PM UTC	1.06 h	-2393.31 h
563190100	container	10/10, 04:30 PM UTC	8.64 h	-289.51 h
249174000	container	10/10, 04:15 PM UTC	1.33 h	-27.25 h
210601000	container	10/10, 04:12 PM UTC	0.45 h	-797.21 h
567584000	general-cargo	10/10, 03:53 PM UTC	-10.02 h	-347.89 h
636023047	container	10/10, 02:06 PM UTC	2.04 h	—
353241000	general-cargo	10/10, 01:36 PM UTC	2.82 h	-23.61 h
412207040	container	10/10, 11:06 AM UTC	-0.40 h	-249.10 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.