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 30 days. We publish the number — good or bad. Lower error is better.

Port Flow predicted ETA (RMSE)

4.65 h

MAE 2.37 h · 445 voyages

Broadcast ETA (RMSE)

38.33 h

crew-declared, 217 voyages (same set)

Our advantage

87.9 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (200 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


352004278	product	10/12, 03:24 AM UTC	0.61 h	-922.41 h
368237190	container	10/12, 01:56 AM UTC	0.51 h	-299.94 h
366791000	container	10/12, 01:03 AM UTC	1.18 h	-163.06 h
538004202	container	10/12, 12:22 AM UTC	0.15 h	—
369006000	container	10/11, 11:56 PM UTC	0.40 h	-1035.94 h
636091473	container	10/11, 11:23 PM UTC	0.51 h	-756.90 h
636014128	product	10/11, 09:25 PM UTC	0.49 h	-1.43 h
477848700	container	10/11, 09:18 PM UTC	0.31 h	-1.30 h
369225000	product	10/11, 04:41 PM UTC	0.37 h	-276.69 h
636020979	container	10/11, 03:32 PM UTC	0.31 h	-4.55 h
367740790	product	10/11, 03:08 PM UTC	0.51 h	-281.15 h
636021471	container	10/11, 02:31 PM UTC	0.72 h	329.48 h
563199600	container	10/11, 01:54 PM UTC	-0.51 h	-1233.91 h
563281400	general-cargo	10/11, 01:00 PM UTC	-1.23 h	-817.01 h
368485000	container	10/11, 11:08 AM UTC	1.86 h	-38.14 h
636015659	container	10/11, 09:35 AM UTC	-0.67 h	-1419.09 h
538011471	product	10/11, 05:12 AM UTC	0.92 h	-23.21 h
563153600	product	10/11, 02:30 AM UTC	0.45 h	-15.01 h
212539000	container	10/10, 09:55 PM UTC	3.19 h	-194.93 h
369567000	product	10/10, 09:20 PM UTC	-0.10 h	-0.35 h
477056400	general-cargo	10/10, 08:13 PM UTC	1.66 h	-643.22 h
367719640	container	10/10, 05:49 PM UTC	-0.01 h	-24.99 h
367688000	product	10/10, 05:24 PM UTC	0.19 h	-1870.16 h
367353080	product	10/10, 03:27 PM UTC	2.01 h	1976.54 h
373048000	container	10/10, 02:38 PM UTC	0.66 h	-0.63 h
636025512	container	10/10, 02:31 PM UTC	0.85 h	-1.52 h
431153000	container	10/10, 02:03 PM UTC	0.67 h	-3145.06 h
636025512	container	10/10, 01:41 PM UTC	1.01 h	-472.70 h
563052600	container	10/10, 01:28 PM UTC	1.37 h	-556.47 h
372148000	container	10/10, 01:04 PM UTC	0.84 h	-2.07 h
563153600	product	10/10, 12:11 PM UTC	0.73 h	-1.19 h
354891000	general-cargo	10/10, 09:59 AM UTC	1.37 h	-1.00 h
338899000	product	10/10, 07:54 AM UTC	3.45 h	-183.90 h
636021962	product	10/10, 03:26 AM UTC	0.83 h	-20.44 h
311056900	container	10/10, 12:41 AM UTC	1.91 h	-731.19 h
369567000	product	10/10, 12:17 AM UTC	-0.14 h	-0.29 h
351928000	container	10/09, 10:58 PM UTC	0.58 h	-1.48 h
441701000	container	10/09, 10:30 PM UTC	-0.03 h	-886.51 h
259951000	container	10/09, 10:01 PM UTC	2.79 h	-2.02 h
338899000	product	10/09, 07:44 PM UTC	-0.19 h	-171.74 h
368237190	container	10/09, 06:05 PM UTC	1.59 h	-244.09 h
367719640	container	10/09, 05:47 PM UTC	-0.53 h	-4.95 h
636021548	container	10/09, 04:59 PM UTC	1.04 h	-1.99 h
416465000	container	10/09, 03:47 PM UTC	-0.26 h	-0.78 h
477182300	container	10/09, 02:01 PM UTC	0.48 h	-510.53 h
636018038	general-cargo	10/09, 01:25 PM UTC	0.12 h	-1735.42 h
563295900	container	10/09, 12:03 PM UTC	0.50 h	-1.05 h
566410000	general-cargo	10/09, 11:59 AM UTC	-0.79 h	-25.00 h
366562000	container	10/09, 09:19 AM UTC	0.60 h	-198.32 h
636024125	container	10/09, 08:34 AM UTC	0.73 h	-10.57 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.