ETA accuracy benchmark · 🇬🇷 Piraeus

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

11.50 h

MAE 5.70 h · 976 voyages

Broadcast ETA (RMSE)

72.80 h

crew-declared, 351 voyages (same set)

Our advantage

84.2 %

lower error than the broadcast ETA

Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (409 excluded). Both figures are computed on the same remaining voyage set — apples to apples.

Closed voyages

error = predicted/broadcast minus actual arrival (hours)


240183900	product	10/12, 05:19 AM UTC	0.94 h	-526.32 h
239550200	chemical	10/12, 05:08 AM UTC	3.12 h	-445.95 h
240871000	chemical	10/12, 04:46 AM UTC	2.74 h	—
240897000	product	10/12, 04:27 AM UTC	2.87 h	—
240897000	product	10/12, 12:28 AM UTC	3.93 h	1943.53 h
237829800	product	10/11, 11:08 PM UTC	1.23 h	—
240205400	product	10/11, 10:03 PM UTC	3.05 h	—
241699000	product	10/11, 09:46 PM UTC	-0.08 h	-60.77 h
240285000	product	10/11, 09:32 PM UTC	-12.60 h	-245.55 h
240205400	product	10/11, 05:26 PM UTC	0.99 h	7184.33 h
240620300	product	10/11, 04:35 PM UTC	-0.42 h	284.40 h
237829800	product	10/11, 03:26 PM UTC	5.79 h	—
239869900	product	10/11, 03:01 PM UTC	1.43 h	1940.47 h
538010227	product	10/11, 12:25 PM UTC	-59.38 h	-66.93 h
237829800	product	10/11, 11:34 AM UTC	0.97 h	—
240666000	product	10/11, 10:29 AM UTC	-0.20 h	—
239869900	product	10/11, 10:26 AM UTC	-0.24 h	1945.06 h
240043800	product	10/11, 08:45 AM UTC	0.46 h	—
241349000	product	10/11, 08:09 AM UTC	1.54 h	-385.16 h
240698300	container	10/11, 07:14 AM UTC	1.08 h	-30.25 h
241654000	product	10/11, 07:02 AM UTC	2.20 h	1125.95 h
237009900	container	10/11, 06:02 AM UTC	1.54 h	-10.05 h
237112600	container	10/11, 05:58 AM UTC	4.20 h	—
237112600	container	10/11, 05:38 AM UTC	0.83 h	-94.65 h
215587000	container	10/11, 05:30 AM UTC	1.90 h	-118.51 h
240620300	product	10/11, 05:30 AM UTC	1.54 h	295.49 h
240666000	product	10/11, 05:21 AM UTC	2.37 h	—
241694000	container	10/11, 04:37 AM UTC	1.85 h	-0.63 h
538012556	container	10/11, 04:37 AM UTC	1.79 h	4.88 h
241694000	container	10/11, 04:26 AM UTC	-0.43 h	-23.44 h
240620300	product	10/11, 04:21 AM UTC	4.67 h	—
237829800	product	10/11, 04:18 AM UTC	1.82 h	—
240620300	product	10/11, 03:37 AM UTC	2.45 h	297.37 h
241654000	product	10/11, 12:12 AM UTC	1.70 h	1132.80 h
636017256	container	10/10, 10:31 PM UTC	-7.13 h	-678.53 h
256837000	product	10/10, 09:09 PM UTC	4.10 h	-431.15 h
256652000	product	10/10, 08:20 PM UTC	1.01 h	-62.84 h
241654000	product	10/10, 07:44 PM UTC	1.80 h	1137.25 h
311001767	container	10/10, 06:59 PM UTC	0.85 h	-972.49 h
211129160	chemical	10/10, 05:45 PM UTC	1.89 h	-203.75 h
413065000	general-cargo	10/10, 05:15 PM UTC	0.98 h	-552.25 h
237009900	container	10/10, 04:10 PM UTC	0.59 h	-71.67 h
240246000	container	10/10, 03:28 PM UTC	3.25 h	-0.47 h
538005825	product	10/10, 03:26 PM UTC	4.39 h	-3.45 h
240620300	product	10/10, 02:01 PM UTC	0.94 h	310.97 h
241654000	product	10/10, 01:57 PM UTC	1.29 h	1143.04 h
249427000	container	10/10, 01:52 PM UTC	1.07 h	-90.88 h
240205400	product	10/10, 01:09 PM UTC	2.29 h	7212.60 h
271052243	general-cargo	10/10, 12:40 PM UTC	0.80 h	-2.67 h
636026639	container	10/10, 12:24 PM UTC	0.42 h	-1.92 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.