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)
10.59 h
MAE 21.43 h · 412 voyages
Broadcast ETA (RMSE)
64.63 h
crew-declared, 248 voyages (same set)
Our advantage
83.6 %
lower error than the broadcast ETA
Broadcast ETAs more than 7 days from the actual arrival are excluded as AIS sentinel / placeholder values (134 excluded). Both figures are computed on the same remaining voyage set — apples to apples.
Closed voyages
error = predicted/broadcast minus actual arrival (hours)| 477798800 | product | 10/12, 02:29 AM UTC | -25.75 h | -1899.49 h |
| 249347000 | product | 10/12, 12:19 AM UTC | 1.21 h | 1943.67 h |
| 244457000 | container | 10/11, 08:04 PM UTC | -0.74 h | -15.07 h |
| 636025095 | product | 10/11, 07:10 AM UTC | -0.06 h | -4.18 h |
| 352514000 | product | 10/10, 11:53 PM UTC | 0.78 h | -82.89 h |
| 538009651 | product | 10/10, 11:23 PM UTC | -37.90 h | -41.39 h |
| 538008272 | product | 10/10, 09:07 PM UTC | 0.14 h | -97.13 h |
| 374880000 | container | 10/10, 03:10 PM UTC | -2.85 h | -2.68 h |
| 538010377 | container | 10/10, 07:03 AM UTC | 0.39 h | -5.66 h |
| 538009852 | product | 10/09, 10:46 PM UTC | 6.11 h | — |
| 235076282 | product | 10/09, 08:49 PM UTC | 0.33 h | -58.83 h |
| 538007389 | product | 10/09, 08:24 PM UTC | 0.67 h | -300.41 h |
| 367561320 | product | 10/09, 03:12 PM UTC | -0.33 h | 5620.79 h |
| 240565000 | product | 10/09, 08:26 AM UTC | 0.73 h | -544.44 h |
| 367561320 | product | 10/09, 02:26 AM UTC | -2.45 h | 5633.56 h |
| 244903000 | container | 10/08, 10:00 PM UTC | -1.05 h | -19.00 h |
| 538006490 | product | 10/08, 09:09 PM UTC | 0.47 h | -1730.16 h |
| 311017100 | product | 10/08, 08:41 PM UTC | -0.01 h | -25.70 h |
| 255806147 | product | 10/08, 04:56 PM UTC | 5.51 h | — |
| 538007042 | product | 10/08, 02:15 PM UTC | 1.27 h | -41.75 h |
| 636022580 | product | 10/08, 11:22 AM UTC | -0.73 h | -1024.38 h |
| 215744000 | product | 10/08, 04:22 AM UTC | 2.46 h | -3.37 h |
| 311069200 | product | 10/07, 11:18 PM UTC | 3.65 h | -19.25 h |
| 352004479 | product | 10/07, 08:24 PM UTC | -1.15 h | -108.41 h |
| 241826000 | product | 10/07, 02:25 PM UTC | -0.23 h | -719.43 h |
| 311069200 | product | 10/07, 12:49 PM UTC | -0.89 h | -8.80 h |
| 305222000 | container | 10/07, 04:47 AM UTC | 1.07 h | -189.79 h |
| 636019141 | product | 10/07, 02:14 AM UTC | -0.13 h | -22.25 h |
| 538003422 | product | 10/07, 01:50 AM UTC | 3.62 h | -48.83 h |
| 352003844 | container | 10/06, 08:15 PM UTC | 0.28 h | -2.76 h |
| 241731000 | product | 10/06, 07:28 PM UTC | -0.50 h | -75.48 h |
| 241731000 | product | 10/06, 05:39 PM UTC | -25.89 h | -73.66 h |
| 241866000 | product | 10/06, 02:13 PM UTC | -48.22 h | -133.23 h |
| 538005826 | product | 10/06, 12:51 AM UTC | 2.91 h | -7.86 h |
| 229446000 | product | 10/05, 08:38 PM UTC | 26.84 h | -16.64 h |
| 241412000 | product | 10/05, 07:20 PM UTC | 3.40 h | -157.34 h |
| 372477000 | product | 10/05, 04:43 PM UTC | 7.77 h | -8.72 h |
| 255915984 | product | 10/05, 12:58 PM UTC | -2.66 h | -253.00 h |
| 636025427 | product | 10/05, 12:57 PM UTC | -11.86 h | 2075.04 h |
| 477205800 | container | 10/05, 08:04 AM UTC | -9.20 h | -12.08 h |
| 255804290 | product | 10/03, 08:25 PM UTC | -0.61 h | -3.42 h |
| 367381000 | product | 10/03, 05:50 PM UTC | -1.59 h | -541.85 h |
| 235076282 | product | 10/03, 04:32 PM UTC | -0.60 h | -1482.54 h |
| 249882000 | product | 10/03, 03:25 PM UTC | 4.53 h | -64.43 h |
| 311001662 | product | 10/03, 12:06 PM UTC | 0.66 h | -2950.11 h |
| 352005882 | product | 10/03, 10:18 AM UTC | 0.44 h | -818.31 h |
| 636023388 | product | 10/02, 08:54 PM UTC | 0.49 h | -41.92 h |
| 356075000 | product | 10/02, 04:42 PM UTC | 1.11 h | -148.72 h |
| 241866000 | product | 10/02, 11:05 AM UTC | -0.50 h | -34.10 h |
| 249997000 | product | 10/02, 02:30 AM UTC | 3.08 h | -3.50 h |
See it on a live desk
Predicted ETAs, congestion and sanctions screening across 51 ports — try it in one click, no signup.
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.