What actually kills people in the mountains

The Austrian register draws a picture that is not the one in the photographs. ÖKAS, the national alpine-safety board, recorded 309 alpine deaths in 2024 against a ten-year mean of 284 — the deadliest recent year on record:

  • 127 of them (41%) were hikers — far ahead of ski slopes (35) and ski touring (26).
  • Leading causes across all activities: cardiovascular events, 86 (27%), falls from height, 65 (20%), slips and trips, 46 (14%).
  • Across ten years (November 2015 to October 2025), hiking produces roughly 2,600 casualties a year, around 106 of them fatal — 37% of all Austrian alpine fatalities; 81% of the dead are male.

Switzerland agrees. The bfu, the federal accident-prevention council, counts over 40,000 hiking casualties a year among residents and 56 resident deaths a year in mountain sports, most often from a fall; mountain sport is its largest fatal-sports category, 84 deaths a year across 2000–2023. SAC-CAS 2025 adds the rescue side: almost 4,000 people rescued, 46% of emergencies while hiking, 98 deaths — the lowest in ten years against a 2015–2024 mean of 119, falls the most frequent cause.

Turning back is not a climber's specialism. It is a walker's skill, and the walkers are the ones dying.

Sources: ÖKAS — Alpinunfälle in Österreich 2024; ÖKAS — Wanderunfälle: Zahlen, Daten und Fakten; bfu — Dossier Wandern & Bergsport; bfu — tödliche Sportunfälle; SAC-CAS — Bergnotfallstatistik 2025.

The descent is the dangerous half

This is the finding the article rests on. In a peer-reviewed study of mountain hiking in the Austrian Alps, Faulhaber and colleagues found that "About 75% of all falls happened during the descent." Falls are roughly 46–50% of hiking accidents, nearly 60% of them caused by slipping, and 75% of fall accidents happen on rubble or grass — only 6% on snow. A second Austrian cohort agrees: Rausch and colleagues, across 7,368 accidents from 2015 to 2021 of which 7.1% were fatal, found women had 69% of their accidents on the descent.

Most people plan a day as if the summit were the finish line; the statistics treat it as the midpoint. Every metre gained is given back on tireder legs, on rubble and wet grass, later and in worse light — the dangerous ground is the easy path down.

The question is not whether you can reach the top, but whether you can still walk down afterwards.

Sources: German Journal of Sports Medicine — Faulhaber et al., 2020; PMC — Rausch et al., 2024.

The time budget: one third down, two thirds up

The clearest official statement of descent budgeting comes from an inverted mountain. The National Park Service tells Grand Canyon hikers: "Allow 1/3 of your time to descend and 2/3 of your time to ascend", and "Plan on taking twice as long to hike up as it took to hike down."

The canyon is the mirror image of a summit day — you descend first and pay later — but the arithmetic transfers: the slow half is the second half. A day planned as a ratio rather than as one round-trip figure produces a defensible turnaround time, fixed in the valley before you are tired and invested. The Park Service's Longs Peak page states the commitment in its own words — be "prepared to turn back during sudden, drastic weather changes" — and names the seasonal pattern behind it: "In June, July and August, be aware of afternoon thunderstorms and lightning."

An honesty note our sources force on us. The slogan "summit by noon" appears in no verified official source we could check. What the guidance actually gives you is better: a pattern (convective risk builds after midday) plus a commitment (a time set before departure and honoured whatever the summit looks like from where you stand). The same standard applies across this section: the lightning crouch is retired in current NWS guidance, and "spread the group out" is no longer in it either, so we do not repeat them. A rule you cannot source is a rule you cannot defend at 3,000 m.

Sources: NPS Grand Canyon — hiking tips; NPS Rocky Mountain — Longs Peak.

Why parties push on anyway

The gap between knowing the right decision and making it has been measured. Ian McCammon's study of 598 US avalanche accidents between 1972 and 2001, involving 1,180 people, found that ordinary decision cues — familiarity with the terrain, social proof, commitment to a stated goal, and scarcity — measurably influenced victims' choices. It held for trained parties: the paper cites Atkins (2001), who found 83% of accidents involving avalanche-aware victims were decision-making errors, not gaps in knowledge.

That is the uncomfortable core of it. The failure mode is not ignorance but a competent party, on familiar ground, that has told everyone where it is going, following tracks suggesting someone else already judged it fine.

An accuracy note: the famous six-trap version — FACETS — is McCammon's later formulation, published in a magazine article for which no stable URL survives verification; we cite the 2002 conference paper.

France's ANENA keeps its own tally: 126 avalanche accidents, 190 people caught and 21 deaths in the season from 1 October 2024 to 30 September 2025, against 25.6 deaths a year over the last ten years and 29.6 a year since 1971, with swings from 9 deaths in 2021-22 to 40 the season before. Touring and off-piste dominate the count; a quiet season is not evidence that your judgement is good.

Sources: McCammon — Evidence of heuristic traps in recreational avalanche accidents, ISSW 2002 (PDF); ISSW catalogue record; ANENA — archives et données d'accidents; ANENA — bilan 2021-2022 (PDF).

The trigger that has nothing to do with the weather

This is general safety information, not medical advice. Consult a doctor before travelling to altitude or taking any medication. In an emergency, call the local emergency number.

Most turnaround guidance treats the deciding variable as external; the mortality data says otherwise. Burtscher's review states that "Sudden cardiac death (SCD) is the major cause of fatalities in males over 34 years of age during hiking or downhill skiing", that risk is greatest on the first day at altitude, that deceased hikers had a prior myocardial infarction in 17% of cases against 0.9% of controls, and that CPR was initiated in only 14% of hiker sudden cardiac deaths. The Austrian figures agree: cardiac events were the largest single cause of alpine death in 2024 (86 of 309) and 37% of hiking deaths over the decade.

So many turnaround decisions are physiological, not meteorological — and they arrive on day one, when the weather is good and nothing has gone wrong. A party that only turns back for storms has no protocol for the commonest killer it faces.

Sources: PubMed — Burtscher, 2007; ÖKAS — Alpinunfälle in Österreich 2024; ÖKAS — Wanderunfälle.

Blocked, not injured

One whole category of rescue is simply a late turnaround. In Switzerland in 2025, 38% of the people rescued were unhurt — up from 29% in 2020 — with alarms for being blocked or exhausted rising. German data shows the same structure: of over 1.4 million DAV members, 1,243 were involved in accidents or emergencies in 2022, and being "blocked" accounted for 23% of hiking incidents — down from a 34% pandemic peak, but structurally large (the DAV reports biennially and covers members only).

Nearly every one was a turnaround that had been available an hour earlier, for free, on your own legs.

Sources: SAC-CAS — Bergnotfallstatistik 2025; DAV — Bergunfallstatistik 2022; DAV — Bergunfallstatistik overview.

Who you go with, and who knows your plan

On going alone, the official guidance is unqualified. NPS Grand Canyon: "Don't hike alone." NPS Mount Rainier: "Do not travel alone. If visibility is poor, do not travel at all." The Swiss bfu advises tackling difficult tours "am besten zu zweit oder in einer Gruppe" — ideally in a pair or a group — and telling a third party, especially when setting out alone. Honesty note: no verified official statistic exists for the share of victims who were alone — this is guidance from the agencies that run the rescues, not a number.

The New Zealand Mountain Safety Council gives the mechanism: "When things go wrong and you're unable to call for help, your emergency contact can raise the alarm if you don't arrive back on time." Leave your route, your party's names and contacts, and your expected return time with someone who will act. Mount Rainier: "Always tell someone your travel plans so they can notify the park if you fail to return."

A return time given to someone at home is a second clock on the same day — and unlike the summit, it does not negotiate.

Sources: NPS Grand Canyon — hiking tips; NPS Mount Rainier — hiking safety; bfu — Ratgeber Wandern/Bergwandern; NZ Mountain Safety Council — leaving trip intentions.

What turnaround discipline looks like

The best library of decisions that went wrong is free: the American Alpine Club has published Accidents in North American Climbing annually since 1948; it "documents the year's most significant and teachable climbing accidents", each analysed "to show what went wrong", with a searchable database of more than 5,000 reports. Reading them is the cheapest way to acquire experience of a decision you have never had to make.

None of this makes the call easy. Turning back is a judgement under uncertainty, usually made while you still feel fine. The evidence only says where to put the guard rails: set the time before you leave; budget the descent as the harder half; name the cues that will push you on; and give someone at home a return time you intend to honour.

The summit is optional. The descent is not.

Sources: American Alpine Club — About the Accidents; AAC — The Prescription.

Sources