FAQ
Which of two extremely safe flights is actually safer?
Every mainline flight is astonishingly safe. The model estimates a fatal risk per flight near 1 in 13 million for the safest carriers, which is 99.9999923 percent safe, down to roughly 1 in 1 million, or 99.9999 percent, for the weakest. That whole range is safer than the drive to the airport.
This site ranks that margin, which is the only thing left to rank once everything is this safe, and it says so plainly when the gap is too small to mean anything. For anxious flyers who want the marginally safer option grounded in real records, not reassurance.
Aren't all the major airlines basically equally safe?
Largely yes. The developed world (US, Canada, Europe, Japan, South Korea, Singapore, Australia, the Gulf) flies at roughly 1 in 10 to 1 in 80 million fatal risk per flight. The score doesn't separate safe from dangerous, because there's no dangerous option here. It ranks the margin between two safe flights, and flags when that margin is statistical noise. It's for choosing the marginally better of two excellent options, not a claim that any mainline flight is unsafe.
What's the difference between an incident and an accident?
An accident involves a death, a serious injury, or major aircraft damage. An incident is below that line: a system fault, a precautionary diversion, a go-around. Incident counts are kept out of the score, because what counts as reportable varies too much between countries to compare fairly. They're shown as context and linked to source, but never scored.
What do you mean by airline "certifications"?
Three public signals. IOSA, the IATA Operational Safety Audit: registered carriers have run several times lower accident rates. The EU Air Safety List: a regulatory ban, scored as a hard penalty. FAA IASA: it grades a country's regulator, and a Category 2 state (below the ICAO oversight standard) adds a bounded penalty to carriers based there. No reputation or brand score.
Where do the weather and turbulence numbers come from?
Airport weather is NOAA METAR observations, refreshed through the day. The route turbulence outlook is computed from NOAA GFS model winds at cruise altitude (via Open-Meteo), sampled along the route. It's a model forecast, and it's labeled as one.
You can't forecast turbulence a week out. Isn't showing it misleading?
Turbulence has a useful forecast horizon of about 36 to 48 hours, so there's no outlook shown beyond 48 hours from departure. Inside that window it's a forecast like any weather forecast: worth acting on, not a promise, and it can change. Crews also adjust altitude and route around rough air in real time. Knowing it's likely lets you pick a seat and keep your belt on.
Don't AIRMETs and SIGMETs just scare anxious people?
They're shown with context: what the advisory covers, how routine it is, and what it means for a large jet whose crew already plans around it, which is usually very little. Raw, they alarm; framed, they inform. A specific you understand beats a blank you imagine.
How do you compare two flights?
Each flight gets an estimated fatal-risk-per-flight. The airline sets the base rate from validated signals (audit status, regulator oversight, recent accident record, flight volume), refined with empirical-Bayes shrinkage so a clean record only counts once enough flying backs it. Aircraft type and engine history are bounded modifiers, because the operator matters far more than the airframe. It's published as odds (1 in X million) with an uncertainty range, and when two flights land within about a factor of two, the model calls it a tie.
Do aircraft age, incident history, or weather really tell me anything about my flight?
Individually, barely. No model predicts a single flight. The airline dominates the risk, the gap between top carriers is often a coin flip, and fleet age and weather are smaller still. The records are shown because the unknown is what fuels anxiety, and a sourced record (usually reassuring) beats a blank space. Where a factor barely matters, the site says so.
Is this just cashing in on anxious people?
The core is free and stays free: the records, the ranking, the calming tools. Pro is optional, with a money-back guarantee. Running a live site costs real money every month, flight data, weather data, servers, and Pro covers that so the free version stays free. Repeatedly telling you flying is extremely safe is a poor way to monetize fear.
Are you affiliated with an airline or a regulator?
No. Independent, and not affiliated with any airline, or with the NTSB, FAA, ICAO, or NOAA. Everything is built from public records, with links back to the originals.
I'm a nervous flyer. Should I even use this?
Depends on you. For many anxious flyers, a concrete sourced record calms the fear better than a blank. For some, more information feeds the loop and refreshing makes it worse; if that's you, close the tab. If you use it: lean on the free calm tools, read the record once instead of refreshing, and remember every number points the same way. Flying is extremely safe.
A pilot said some of this is wrong. What's your answer?
Good. Experts catch what I can't, and that feedback has already improved the site. Where it's right, it gets fixed: I cut two factors from the score that couldn't be defended, removed numbers I couldn't source, and capped turbulence to its real 48-hour window. If something's wrong, tell me and I'll fix it or show the work.