ETA accuracy benchmark · 🇺🇸 Corpus Christi

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

11.07 h

MAE 23.64 h · 392 voyages

Broadcast ETA (RMSE)

61.89 h

crew-declared, 268 voyages (same set)

Our advantage

82.1 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (81 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


367457000	product	08/26, 10:50 PM UTC	-0.68 h	-55.85 h
636022580	product	08/26, 09:07 PM UTC	-0.99 h	-1141.12 h
538004174	product	08/26, 09:05 PM UTC	4.46 h	-1532.09 h
355037000	product	08/26, 06:30 PM UTC	0.68 h	-4.20 h
538011662	product	08/26, 04:56 PM UTC	4.82 h	-1.95 h
477607200	product	08/26, 08:55 AM UTC	1.12 h	-1148.92 h
305663000	container	08/26, 05:56 AM UTC	-0.38 h	-88.94 h
338203000	product	08/26, 01:09 AM UTC	1.31 h	-91.15 h
368388590	product	08/25, 11:57 PM UTC	0.50 h	—
367633000	product	08/25, 05:29 PM UTC	0.45 h	-5.49 h
367633000	product	08/25, 05:12 PM UTC	0.73 h	-5.21 h
338203000	product	08/25, 10:10 AM UTC	-0.70 h	—
241613000	product	08/25, 06:54 AM UTC	5.80 h	-1531.40 h
338203000	product	08/25, 04:29 AM UTC	0.15 h	-70.50 h
338203000	product	08/25, 02:50 AM UTC	-0.88 h	-68.84 h
366495000	product	08/25, 01:23 AM UTC	-0.85 h	-682.64 h
368388590	product	08/24, 10:35 PM UTC	0.65 h	—
368388590	product	08/24, 10:03 PM UTC	-0.21 h	—
368388590	product	08/24, 05:32 PM UTC	1.02 h	-24.88 h
368388590	product	08/24, 04:19 PM UTC	0.80 h	—
368388590	product	08/24, 02:25 PM UTC	-0.07 h	—
368388590	product	08/24, 01:56 PM UTC	0.20 h	—
368388590	product	08/24, 10:37 AM UTC	-0.27 h	—
538012219	product	08/24, 08:42 AM UTC	0.58 h	-3.71 h
257607000	product	08/23, 10:41 PM UTC	-0.08 h	-100.69 h
367455770	chemical	08/23, 09:07 PM UTC	-0.59 h	-108.12 h
368388590	product	08/23, 08:03 PM UTC	0.89 h	—
368388590	product	08/23, 07:58 PM UTC	0.60 h	-3.32 h
367643270	lpg	08/23, 07:34 PM UTC	-0.57 h	2.42 h
636014705	container	08/23, 07:25 PM UTC	-0.57 h	-2.43 h
241854000	product	08/23, 06:20 PM UTC	7.25 h	-316.34 h
311001246	container	08/23, 06:10 PM UTC	-0.60 h	-82.18 h
419001244	product	08/23, 05:30 PM UTC	8.52 h	-1082.51 h
636015884	container	08/23, 05:27 PM UTC	1.90 h	-19.31 h
636017950	product	08/23, 01:57 PM UTC	-3.85 h	-72.96 h
636021119	product	08/23, 01:01 PM UTC	1.63 h	-71.03 h
368004000	product	08/23, 05:14 AM UTC	0.01 h	-557.24 h
538006168	product	08/22, 10:26 PM UTC	-0.23 h	-80.94 h
636025912	product	08/22, 02:47 PM UTC	-90.47 h	-92.80 h
249910000	product	08/22, 02:47 PM UTC	-540.37 h	-497.79 h
477043500	product	08/22, 02:38 PM UTC	-2322.24 h	-1801.64 h
538011635	product	08/19, 05:37 AM UTC	-0.51 h	-49.63 h
367643270	lpg	08/19, 12:16 AM UTC	2.14 h	-739.28 h
538010650	product	08/18, 10:14 PM UTC	2.82 h	-4.25 h
241785000	product	08/18, 10:07 PM UTC	1.02 h	-26.62 h
241785000	product	08/18, 09:38 PM UTC	-0.05 h	-26.14 h
305315000	container	08/18, 07:05 PM UTC	-1.89 h	-128.08 h
249986000	product	08/18, 08:54 AM UTC	0.27 h	-1797.91 h
352004782	container	08/18, 04:16 AM UTC	1.83 h	—
538006434	container	08/18, 02:15 AM UTC	2.25 h	-3.26 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.