Skitour Switzerland 2020 Patraflon

Ski touring vs the daily commute: how risky is the backcountry?

What is the risk of unassisted backcountry ski touring? I decided to clarify the concept by comparing backcountry ski trips to daily commutes.

1 percent

That is the amount of residual risk promised by snow and mountain safety guru Werner Munter. This is, of course, absurd.

I first heard this percentage during a backcountry skiing course. Thankfully, I was not alone in my disbelief. A ‘one-in-a-hundred chance’ implies that we won’t survive more than a hundred skiing trips on average! Even if we planned every tour by the book – Munter’s book to be precise. Even if we base our decision on his ‘3 × 3 Filter- and Reduction’ methodology that should make ski-touring risk acceptable.

This article concerns the risk involved in unassisted backcountry ski touring. It argues that we do not really know how much or little safety is provided by risk-reduction techniques, such as the Werner Munter method. Fortunately, this does not mean that these risk management approaches are irrelevant; merely that there is significant work yet to be done.

The Netherlands Royal Mountain Sports Association (NKBV) provides its flatland members with opportunities to develop hiking, climbing, and backcountry skiing skills. Understandably, the Dutch alpinist spends a considerable amount of time on theory and practicing indoor techniques, as compared to mountain-dwelling enthusiasts. There are risks involved in all outdoor sports endeavors. Still, where sport climbing has the benefit of certified materials and thoroughly tested procedures (here too, Munter’s legacy will forever be secured in the Munter hitch), backcountry skiing relies heavily on theoretically informed choices. The sport leans on the skier’s judgment of the reliability of the snowpack and the slope. By choosing which slope to enter – and which not – the skier can considerably increase the safety of her travel through avalanche-prone backcountry terrain. However, even perfect application of theory, planning, and execution cannot reduce the skier’s risk to zero. In the mountains, the weather is changeable and the snowpack fickle. The skier has no choice but to accept that some amount of residual risk is inherent in every trip into the backcountry. Despite her precautions and informed choices, she may encounter trouble of the type that she will be unable to recount.

What is this ‘residual risk’? Traffic provides a standard measure of risk that we all seem to accept without much question. I decided to clarify the concept by comparing backcountry ski trips to daily commutes. How many backcountry skiing trips would I have to make, in order to reach the corresponding level of residual risk that we accept on commuter trips in traffic?


To speak about risk, you need the micromort. The micromort is a unit that expresses the probability that a specific activity, repeated a million times, will result in fatality. In traffic, it can indicate the number of lethal accidents per one million commuter trips. If you want to know how much safer your transport of choice is, you simply compare the micromorts of one option with the micromorts of the alternative. An example: consider the quintessentially Dutch statistics for the car and the bike. Between 2015 and 2017, the Dutch used their car for a yearly 7 billion trips and their bikes for 4 billion trips. The annual number of car trip fatalities was around 220. For bike trips, this number was 190.

So that means that a billion car trips will result in approximately 30 fatal accidents while a billion bike trips will result in approximately 50 fatal accidents. Hence, one car trip carries 0.03 micromort and one cycling trip carries 0.05 micromort. While both are still incredibly small probabilities, the likelihood of being killed in your car or on your bike in the Netherlands might be larger than that of winning the lottery: which can be as small as 25 in a billion.

Two such micromort estimates merely serve as a way to compare the possible outcomes of two actions. What’s more, reducing your risk by using your car, instead of your bike, actually increases your risk of killing a cyclist: your ‘microkill’. The point of this macabre section is to show that we have this data for our everyday commutes, and ask whether it is possible to talk about backcountry skiing in the same manner.

In the Netherlands, we know a lot about traffic because Statistics Netherlands (CBS) conducts meticulous survey investigations on its inhabitants’ behavior in traffic and combines traffic accidents from various registries. Alpine countries (Austria, France, Italy, and Switzerland) have also been keeping track of fatal avalanche accidents since 1936. To get the micromort value though, we also need to know the number of skiing trips that ended well. Until 2016, such figures were nowhere to be found. So how was Werner Munter able to be so self-assured in the 90s? And how should we then perceive the methodology which is universally referred to in snow safety training materials?

Italy 2019 glacier crossing to Rifugio ai Caduti dell’Adamello Photo (c): Bart Videler

The Stats

Werner Munter used the data of 91 avalanche accidents, occurring between the 70s and 90s, to distill specific scenarios, which every backcountry skier is trained to avoid. For example, half of the avalanches in the data set occurred on slopes with a gradient steeper than 39 degrees. Ergo, according to Munter’s method, a backcountry skier can cut her risk in half by avoiding slopes over 39 degrees.

Let’s take a moment to apply this use of scenarios to traffic accidents involving cyclists in the Netherlands. Here, cycling is perceived as transport and not as a sport. Moreover, the attitude to safety is very relaxed due, in part, to cyclists’ well-guarded right of way. Nevertheless, we know that more than half of cycling accidents happen on a bike with a step-through frame, rather than a cross-bar frame. So, if you were to be a, rather stiff, ‘Munter of bike accidents’, which rule would you enforce? Would you halve the accident risk by imposing cross-bar frames? Of course not.

We see a higher occurrence of accidents on step-through frame bikes because there are more of them in circulation, not because they are more dangerous, per se. In other words: the number of bike rides that end happily, and the number of rides ending in serious accidents, probably both involve the same amount of step-through frames.

Cue the important role for common sense in constructing Werner Munter’s method. An experienced alpinist, Munter knew that half the backcountry skiers, at all times, were not located on slopes over 39 degrees. So, it was conspicuous that more of the lethal avalanche accidents took place there. In the same way, Munter factored in the hillside exposition where avalanches were located, as well as the forecast level avalanche danger of the day the avalanches in his data set occurred (using the ranges 1 to 4). Hence the Munter method still relies on common sense to use the statistics, but that does not necessarily make it less reliable.

We actually act similarly in traffic. Take, for example, the ban on drunk driving. A large number of lethal accidents still involves alcohol consumption: approximately 12-23 percent, let’s say one in six accidents, according to a Dutch estimate from 2015. However, common sense tells us that incidents of drunk driving are rare. Think of the billions of humdrum commuter trips taking place every day. It’s not like, at any time, one in six drivers is drunk! Therefore, we can conclude that if you avoid combining alcohol and driving you diminish the risk of having a car accident.

So sometimes, if you do not have the complete data available, you can get around with common sense. To make the very best choice, however, we still have to know the micromorts. Do we ask the police to prioritize penalizing either alcohol consumption or the use of smartphones while driving or cycling? We know that both increase risks in traffic. However, which transgression is more dangerous and should be fined more stringently? We cannot quantify any of that if we don’t know the micromorts.

The Dutch and their theoretic approach to mountain risk

It does speak to our flatland approach that so many Dutch ski tourers know the 1 percent statistic reported in Werner Munter’s book. After all, we do not ’feel’ the risk of mountain sports every day, so we need to prepare for it from home. This approach is contrasted with the Dutch approach to cycling. Every Dutch person observes the risk of cycling – or actually, the lack of risk – every day. What we observe, we can understand. Most of us, for example, consider cycling without a helmet to be safe. Yet, we will judge motor scooter drivers without a helmet to be irresponsible. We have a feeling for the risk of speed, and the external threat that others in traffic pose to the cyclist. This ‘feeling’ for risk is of course based on our own observations, but those also concur with the data: the average scooter trip in the Netherlands really is almost 5 times as deadly as the average bicycle trip, with 0.24 micromort vs 0.05 per trip.

Maybe the reason I’ve never heard a mountain guide complain about the 1 percent statistic, is that they immediately discard it. 1 percent would be 10.000 micromorts per tour. Their job is really not thát dangerous. A mountain guide once told me that an Austrian insurer places mountain guides in about the same risk category as scaffolders, but with a smaller risk of disability.

For traffic, we can compare cycling and driving because we know both the number of fatal accidents and the total number of journeys. An estimate for backcountry skiing has been available since 2016 when the Swiss avalanche research institute SLF collaborated with Sport Schweiz to conduct a large population survey to define the number of hours that the Swiss practice (mountain) sports. An estimate of the total number of ski tours was combined with the number of avalanche accidents. The result was 9 micromorts per trip.

Comparing micromorts

9 micromorts is considerably higher than the 0.03 and 0.05 for driving and cycling. Nonetheless, many backcountry skiers are convinced that the risk of the sport is acceptable, thanks to risk reduction methods like Munter’s. The Swiss avalanche researchers reporting the 9 micromorts are also optimistic. According to them, one mid-week ski trip per year is just as dangerous as a whole year of participating in Swiss traffic (including cyclists, drivers, pedestrians, etc.). Dutch commuter traffic is actually safer than this Swiss average; one year of commuting by car and bicycle in the Netherlands only buys a Dutch person 3 days in the Swiss backcountry.

For a long time, the risk of backcountry skiing couldn’t be quantified in any more detail, just as we observed for the risk of drunk driving and ‘phone’ driving. Except in skiing, the phone isn’t the culprit, but rather a savior. Skiers have progressively used their phone’s GPS trackers, telling us exactly which ski tours end well.

Which circumstances did backcountry skiers confront most often? Which slope gradients exactly, which snow conditions, on which mountains, at which altitude? These were all factors Munter still had to guess while lumping together many situations (notably those on slopes steeper than 39 degrees), until the phones were able to provide the exact specs for their specific location.

If I make my backcountry trips as safe as possible, those 9 micromorts actually don’t apply. Because the 9 micromorts are estimated for a mix of all backcountry tips. This includes trips that are tightly planned, by the books, using risk reduction methods such as the Munter method, and also the sloppier trips which might only have ended well due to good luck. How much safer than 9 micromorts the Munter method makes our trips, or how much more dangerous it is to disregard it, has not been quantified.

Austria 2018 Mittagskogel (Ötztaler Alpen) Nord face (view into Pitztal) Photo (c): Irene Fikkers

The Future

Since the data was collected for the 9 micromorts estimate in 1999, 2007, and 2013, the smartphone has provided much information. Avalanche researchers are now busying themselves with this data. The first analyses of data from tools such as gpstracks,, and are already providing a treasury of insights.

This data will help backcountry skiers plan their tours with as little residual risk as possible. Not just using the slope gradient with the avalanche bulletin and the filter method recommended by Munter, but with detailed data from ski trips in that period of the year, on that same mountain, even on the different parts of the slope on the route. That is why I pose that the biggest gains yet to be made in risk reduction still lay before us. Imagine that you could plan your tour online and that red, orange and green patches would automatically be projected over your itinerary. Like you used to try to do the old fashioned way with the graphical version of the Munter method: you get red, orange, or green when you combine the avalanche bulletin with the information on your gradient. The modern way would work the same, but at a much higher level of detail and without the use of common-sense estimations. Pure data; relevant to the exact ski tour I would like to make.

According to the first analyses, such enhanced risk reduction tools could still greatly reduce residual risk, if you stay within their newly defined green area. That means that backcountry skiing would be cut off the list of extreme sports. Currently, diving equals 5-10 micromorts per dive, while parachute jumping is 8 micromorts per jump, and delta flying 8 micromorts per flight. The possible future of backcountry tour skiing stands at 2 micromorts per trip, rendering a year in Dutch traffic comparable to 2 weeks of an annual ski holiday.

The initial research into risk of traffic and ski touring was performed while writing this (Dutch) article for De Correspondent. A rewrite of the research focusing more on ski touring was originally published (in Dutch) by the magazine Hoogtelijn of the Netherlands Royal Mountain Sports Association (NKBV). That piece was translated into English, edited and extended for an international audience with incredible help from Kim Ménage. Kim, thanks for the ‘ski-nerding’. You’re the best ski touring buddy I can imagine.


This blogpost appeared originally on Judith ter Schure’s blog.

Main image: Switzerland, 2020, Patraflon, Nicole Clerx

Source of the 1 percent (German):
Munter, W. (1997). 3 x 3 Lawinen: entscheiden in kritischen Situationen. Agentur Pohlmann & Schellhammer.
The 1 percent estimate appears on page 120: “Da die Kriterien weitgehend voneinander unabhängig sind, entsteht ein Restrisiko von 0.4 x 0.25 x 0.1 = 0.01 […] Da die Sicherheit von 99% immer noch ungenügend ist (auf 100 Touren eine Lawinenauslösung), wird sie mit der Reduktionsmethode kombiniert.” These numbers 0.4, 0.25 and 0.1 originate in the three stages that reduce risk in Munter’s method: 40 percent in the “Regeonales Netz”, 25 percent in the “Lokales Netz” and another 10 percent in het “Zonales Netz”. Munter does not report a specific source for these numbers, but they are referred to widely, for example on this German Wikipedia page about decision strategies in avalanche-prone backcountry terrain.

Source of the 9 micromorts (English):
Winkler, K., Fischer, A., & Techel, F. (2016). Avalanche risk in winter backcountry touring: Status and recent trends in Switzerland. In Proceedings of the international snow science workshop (pp. 270-276)

Source of the 2 micromorts (English):
Schmudlach, G., Winkler, K., & Köhler, J. (2018) Quantitative Risk Reduction Method (QRM) a Data-Driven Avalanche Risk Estimator.

Combined links to sources and calculations of the 0.03 (car), 0.05 (bicycle) and 0.24 (scooter) micromorts (Dutch):

Source of micromorts for diving, parachute jumping etc.. (English and trusted, i.e. collected by David Spiegelhalter):

Avatar photo

Judith ter Schure

Wat krijg je als een bergsporter besluit statisticus te worden? Micromorts, survival analyse en een hoop gemijmer over de rol van risico, pech en geluk in het leven. En zoals bij wel meer statistici: een grote liefde voor knap ontworpen experimenten, strak bijgehouden tijdsreeksen en open data met al hun beperkingen.

Add comment

To the VVSOR website