The archive contains 208,801 sailing records. The primary status marks 21,772 as cancelled: 10.4% of all records. This is neither a forecast for an individual trip nor a complete account of every sailing in IDO’s history.
| Year | Records | Cancelled | Share | Days with records |
|---|---|---|---|---|
| 2018 | 26,044 | 1,390 | 5.3% | 314 |
| 2019 | 29,080 | 1,017 | 3.5% | 365 |
| 2020 | 27,140 | 7,202 | 26.5% | 366 |
| 2021 | 22,200 | 1,712 | 7.7% | 322 |
| 2022 | 26,911 | 4,606 | 17.1% | 365 |
| 2023 | 20,140 | 1,832 | 9.1% | 365 |
| 2024 | 22,683 | 2,204 | 9.7% | 366 |
| 2025 | 20,015 | 780 | 3.9% | 365 |
| 2026 | 14,588 | 1,029 | 7.1% | 257 |
Why the denominator and coverage matter
The 2018 count starts on 21 February; 2026 ends on 14 September. The archive was retrieved on 14 September 2026. Of 3,128 requested calendar dates, 3,085 contain records. Successful API responses were empty on 43 dates in 2021. We do not invent cancelled or completed sailings on those days. Changing route mix and schedule publication also mean a difference between years cannot, by itself, demonstrate worsening weather.
Which status field did we count?
Each unique archive row is an observation. The denominator contains 187,024 completed, 21,772 cancelled and 5 unresolved records. Primary and web status disagree about cancellation on 850 rows; those rows remain in this table. The result is an explicitly defined count of a source field, not a verified individual cancellation probability.
Archive rows, physical sailings and passengers are different units
For example, if a departure record is revised or a new record is created for a different vessel, those two source records do not necessarily represent two completed journeys. This is a counting example, not evidence that a particular cancellation received a replacement sailing. Timetable revisions, different vessel or departure-time records and intermediate stops can change the set of records being counted. The archive does not reconstruct the full sequence of every change. The tables therefore count “archive records”; they must not be assumed to count unique physical departures, tickets or people who boarded.
The weather study is a different sample
The weather study excludes unresolved or conflicting labels, March–July 2020 and selected prolonged cancellation runs. Its lower cancellation share must not replace the all-archive figure on this page. None of those weather-study filters is applied here. This summary cannot establish causes, last-minute changes or whether a different vessel replaced a cancelled service.
Reproduction and corrections
The tables on this page explain the observation periods, annual counts and chart values. We do not offer the raw archive or CSV or JSON data downloads. Share = rows marked cancelled / all rows × 100, rounded to one decimal place for display. We do not publish raw booking or capacity records. Report a suspected error through the contact page.
Related research
Which months have more IDO cancellations? A 2023–2025 comparison · Why can IDO ferries be cancelled in calm weather?
Sources and data
İDO: official sailing and ticket information · MGM: Marmara marine forecast
Cite this page
When using this analysis, please credit Sefer İptal and link to this page. İDO supplies the sailing records; Sefer İptal prepares the explanations and calculations on this page.
Sefer İptal. IDO cancellation statistics: the 2018–2026 archive. 2026-09-17. https://seferiptal.com/en/guides/ido-cancellation-statistics-2018-2026/