The methodology
behind the verdict.

Door-to-door, not gate-to-gate

The most common flaw in flight-versus-train comparisons is cherry-picking the start and end point. Published flight times typically run from departure gate to arrival gate. That number is useless if you are standing in your kitchen deciding how to travel.

Our door-to-door time includes every leg of the journey on both sides:

• Access time — the journey from a central point in the origin city to the airport or station, using public transport at peak times.
• Check-in and security buffer — for flights, the recommended arrival window (typically 90 minutes for short-haul Europe); for trains, a realistic platform arrival time (typically 10–15 minutes).
• The journey itself — scheduled flight or train duration, including any required transfers with their connection buffer.
• Exit time — deplaning, passport control if applicable, baggage reclaim for hold luggage, and then the journey from the terminal or arrival station to a central point in the destination city.

For train journeys with a required transfer, we add a realistic connection buffer rather than assuming the printed minimum. For direct overnight trains we count the departure evening and arrival morning but note that the journey replaces a hotel night, which can materially change the value calculus.

Realistic, not idealised

A comparison built on best-case assumptions tells you nothing useful. We research each route under normal conditions:

• Weekday, not weekend — trains run closer to timetable on weekdays; airports are busier in a predictable way.
• Average weather — no fog delays, no summer thunderstorm holds, but also no exceptional tail-wind. We note routes where weather-related disruption is structurally elevated.
• One hold bag per passenger — this affects check-in requirements for flights and baggage reclaim time on arrival.
• Off-peak fare bands — we do not research the cheapest possible advance fare, nor the walk-up peak price. We use the mid-range fare band a traveller booking a few weeks out would realistically encounter.

Honest when train loses

We are not a rail advocacy site. Some routes should be flown; we say so clearly. The verdict categories are Train, Flight, Depends, and Avoid (the last for routes where neither option is compelling for most travellers, or where the route is genuinely impractical).

A "Depends" verdict is not a cop-out. It means the answer legitimately changes based on a factor we cannot resolve for every reader — usually whether you are travelling to the city centre or an outer suburb, or whether cost or time is the primary constraint.

AI-assisted drafting with editorial fact-checking

RailOrFlight uses an AI writing assistant to help produce first-draft copy for each route. This speeds up the process of structuring information and surfacing the relevant comparisons. However, all factual claims — journey times, fare bands, transfer requirements, frequency data, disruption statistics — are verified by a human editor against primary sources: operator timetables, official journey planners, and published statistics from relevant infrastructure bodies.

We do not publish AI-generated numbers without verification. If a figure cannot be verified against a named source, it is removed.

Quarterly refresh cadence

Rail and aviation schedules change. New high-speed lines open, airlines drop routes, strike patterns shift, and fares drift. Every route on RailOrFlight carries a last-reviewed date in the format Month Year. We aim to revisit each route at least once per quarter and always when a significant schedule change affects it.

If you notice that a detail is out of date, please let us know. Reader corrections are taken seriously and credited in the revision history where relevant.

What we measure

Door-to-door minutes

Total elapsed time from city-centre origin to city-centre destination, as described above. Both train and flight times are quoted on this basis.

Scheduled flight block

Airborne minutes from published airline schedules, excluding airport access, security, and ground time. We source this from ProFlightSearch.com, our sister flight-data project.

CO₂ comparison

Carbon intensity is estimated using IEA emission factors for the relevant national electricity grid (for trains) and ICAO methodology for short-haul aviation (per-seat, economy class, including radiative forcing). We express the difference as a multiplier — "the flight emits approximately X times more CO₂ per passenger" — rather than quoting absolute grams, which vary by season and aircraft type.

Fare bands

Approximate cost comparison in a low / mid / high band, based on prices observed booking 2–4 weeks in advance. We do not track live fares; treat these as indicative, not current.

Frequency

Approximate number of direct departures per day in each direction, to convey scheduling flexibility.

Transfer fragility

A qualitative score reflecting how much a missed connection would disrupt the journey — higher for routes where a single transfer failure means a multi-hour delay with no easy alternative.

Disruption risk

A route-level score capturing structural risk from industrial action, infrastructure constraints, or weather on the specific corridor.

What we do not measure

• Real-time prices — fares change by the minute; we cannot and do not reflect current ticket prices. Use the operator, ProFlightSearch.com, or a comparison site for live flight fares.
• Real-time availability — we do not query live inventory. A route we describe as "high frequency" may be sold out for your specific date.
• Personal preferences — train comfort versus aisle seat, lounge access, loyalty points, fear of flying, motion sickness on winding mountain lines. These are real factors; they are yours to weigh.
• Connecting itineraries beyond two legs — if your journey requires three or more legs each way, the complexity is beyond our current scope.