How Mistake Fares Actually Work (And Why Most Are Already Gone)

In January 2019, Cathay Pacific accidentally sold first-class tickets from Vietnam to New York for $675 return. The fare lived for about eleven hours. The people who got it weren't lucky. They were watching.

That gap between lucky and watching is what makes mistake fares interesting. They look like accidents to the public and they're treated like accidents in news headlines, but inside the industry they're a known failure mode of a system that prices thirty thousand seats a minute. Understanding how mistake fares actually happen is the difference between reading about them on Twitter the next morning and being on the flight.

What a mistake fare actually is

A mistake fare (or error fare, depending on which forum you read) is a published airfare that violates one of the rules the airline's revenue management team would have applied if a human had been in the loop. The most common causes are mundane.

The first is a missing fuel surcharge. International long-haul tickets are built from a base fare plus a stack of surcharges. When a fare is refiled and the YQ/YR component drops off the filing, the total can fall by $400 to $900 instantly. The El Al $200 New York to Bangkok fare from 2019 was this exact bug. The base fare was correct. The taxes were correct. The carrier-imposed surcharge simply wasn't there.

The second is a currency conversion error. A fare filed in a foreign point of sale gets converted at the wrong rate, or the decimal moves. United's famous Denmark-to-United-States $0 first-class fares in 2015 were a currency bug at the Danish point of sale. A Lufthansa Norwegian-krone bug in 2018 sold transatlantic business class for around €530.

The third is a stale fare on a codeshare. Carrier A files a sale. Carrier B's pricing engine pulls the wrong base and combines it with the wrong rules. The result is technically a valid published fare for a few hours, until someone notices.

None of these are glitches in the colloquial sense. They are exactly what the system was told to do, given inputs nobody audited in time.

Why the difference between a mistake fare and a fire sale matters

Genuine mistake fares are filed at fares the airline does not believe is the right price. That has two consequences worth knowing about.

The first is that they get pulled fast. Most mistake fares live between 90 minutes and 14 hours. The Cathay Vietnam fare was unusually long-lived at eleven hours because it broke over a weekend. A Qatar Airways US-to-Cyprus business class bug in 2022 lived for about three hours. If your alert source is a daily email digest, you are reading an obituary, not a deal.

The second is that the airline reserves the right to cancel. In the US, DOT rules from 2011 to 2015 required carriers to honor any ticketed fare. That rule was loosened. Today, US-based carriers can void mistake fares if they refund you within a reasonable window. Non-US carriers operate under their home regulator, which is usually more permissive. Cathay honored the Vietnam fare. American Airlines, in 2018, did not honor a Hong Kong-to-US business class glitch.

A fire sale is something else entirely. It's a deliberate decision by revenue management to dump inventory into a low fare class because forecasted load factor on a specific flight is below target. TAP Portugal does this routinely on US-to-Lisbon routes shoulder-season. ANA does it on the Tokyo Haneda to JFK route in February. These are not errors. They are also not protected from cancellation, because there's nothing to cancel.

The shorthand: a mistake fare gives you a story to tell. A fire sale gives you a vacation.

What it takes to catch a real one

Three things, in order of underrated.

Polling cadence. A mistake fare priced at 02:14 UTC will be flagged by the airline's own anomaly detection by roughly 02:40 to 04:00. If your monitoring system checks once a day, your chance of seeing it is approximately zero. If it polls every 15 minutes against the right carrier, the right cabin, and the right origin, your chance is something like one in three for any given monthly mistake-fare cycle.

Booking discipline. When the alert hits, the question is not "should I go to Tokyo?" It's "do I have my passport details saved, a card that won't trigger fraud review, and 12 minutes to spare?" Mistake fares disappear because the ticketing queue rejects the ticket, not because the search result vanishes. Booked-but-unticketed is not booked. Hold the reservation, get the ticket number, then close the laptop.

Patience for the cancellation. Roughly one in four mistake fares gets cancelled. The carrier refunds you, sometimes with a goodwill voucher, occasionally with the original ticket honored after public pressure. Don't book non-refundable hotels for 72 hours.

The boring tools that actually work

Most successful mistake-fare hunters run two layers. The first is a wide net: an alert service that watches the entire market and pushes notifications by phone within a minute of detection. The second is a per-route watch on the three or four cities they actually care about, set tighter, with a lower threshold.

The mistake here is treating mistake fares as a hobby. They aren't. They're a low-frequency event with a short response window. Either you have infrastructure on it or you don't. Casual checking of Google Flights once a week will produce, on average, zero mistake fares per year.

If you'd rather have a system watching the wire for you, Flyozo does exactly that. It polls airline inventories continuously, flags error fare candidates within minutes, and pushes them by mobile notification so you find out while the fare is still bookable, not after.