Moderate is not mellow: why people still get caught on Level 2 days

There’s a particular kind of dangerous day that doesn’t look dangerous.

Not the obvious days. Not the big storm days when the mountains are shouting. I mean the days that look “reasonable” on paper. The bulletin says Moderate. Maybe “Moderate, trending up,” or “Moderate with pockets of Considerable.” You drink your tea, scan the problems, nod at the travel advice, and tell yourself the grown up sentence: we’ll select terrain carefully.

Then you go outside and do what humans do.

This is a deep dive into why avalanche incidents still happen on Moderate days, even among people who “know what they’re doing.” It’s not a morality story. It’s not “education fixes everything.” It’s a look at the cognitive and social machinery that quietly steers decisions once you’re in high consequence terrain.

1) The boring truth that explains a lot: exposure

One reason Moderate produces a lot of incidents is boring, statistical, and easy to forget: Moderate is common. When something is common, people are out more often. When people are out more often, the accident count rises even if the day is “less hazardous” than Considerable or High. In other words, raw accident counts are partly a story about how many people are playing the game that day.

This comes through clearly in the classic ISSW analysis by Greene and colleagues, “Fatal avalanche accidents and forecasted danger levels”, which compares distributions of danger ratings and fatal accidents across multiple regions.

2) Moderate is not “half as bad as Considerable”

A huge piece of the Moderate problem is that many users interpret the danger scale in a way that does not match how hazard behaves.

In Morgan et al. (2023), an online survey of 3195 recreationists found that about 65% perceive the North American danger scale as linear. That means many people intuitively treat the jump from Low to Moderate as similar to the jump from Moderate to Considerable, and similar again up the scale. But the scientific understanding and the consequences do not behave that way.

Here’s a data point that lands hard. In Switzerland, Winkler et al. (2021) linked avalanche accidents with millions of GPS movement points in avalanche terrain and found that risk increases about fourfold from one danger level to the next, including from Low (1) to Moderate (2). The Swiss SLF summarized the same finding in a plain-language writeup: “Avalanche risk recalculated”.

Put those together and you get a very normal failure mode: your brain reads Moderate as “manageable,” your behavior becomes only slightly more cautious, but the hazard can be meaningfully higher than Low.

3) Why Moderate feels permission-y

There’s also a psychological thing going on with the number “2.” Moderate often reads like a green light with footnotes. Not a stop sign. Not even a solid yellow. More like “Proceed if you’re smart.” That framing invites justification. It encourages the mindset of: we’re capable, so we can manage this.

And if you’re honest, you can feel the shift. On a Considerable day, many groups start with a defensive posture. On a Moderate day, many groups start with an offensive posture: let’s find something that goes. That “let’s make it work” posture is where a lot of cognitive biases slip in.

4) The classic traps: familiarity, social proof, commitment, scarcity

Ian McCammon’s work is still the backbone here because he looked at accident patterns and asked the right question: if people have avalanche education, why are they still dying?

In McCammon (ISSW 2002), and in the longer writeup “Heuristic Traps in Recreational Avalanche Accidents”, he describes four cues that show up again and again in recreational accidents: familiarity, social proof, commitment, and scarcity.

These aren’t “stupid people problems.” They’re human shortcuts.

Familiarity sounds like: “This slope has gone fine before.” “This zone is basically my backyard.” Familiarity creates comfort, and comfort is a sedative. McCammon even discusses how ideas like “skier stabilization” can become part of that comfort story, even when the evidence is weak or context dependent.

Social proof sounds like: “Other tracks are on it.” “People are skiing similar lines today.” Social proof is powerful because it feels like evidence, but it’s often just noise. You don’t see near-misses. You only see the tracks that survived.

Commitment is the big one on Moderate days: “We planned this all week.” “We drove all night.” “We came here for this line.” Moderate doesn’t tell you “no.” It tells you “maybe.” And “maybe” is easy to bend into “yes.”

Scarcity sounds like: “This is the window.” “This is the storm we’ve been waiting for.” Scarcity makes people treat a ski objective like it’s a disappearing resource. It creates urgency, and urgency eats caution.

To be clear, McCammon is careful about what accident data can and cannot prove: it strongly shows patterns and correlations, not neat causation. But the patterns match what many of us recognize in ourselves.

5) The uncomfortable part: knowledge does not automatically change behavior

One of the most useful critiques in this whole space is Johnson (2020), which reviews how the avalanche community adopted the “heuristics / heuristic traps” paradigm, what it helped us see, and where it can become a lazy explanation if we stop thinking.

The takeaway isn’t “people died because they were biased.” The better takeaway is: knowledge and behavior are not the same system. You can teach people snow science and route selection. And still, when desire, group dynamics, and uncertainty combine, people will sometimes override what they know.

That doesn’t mean education is useless. It means education can’t be only “more information.” It also has to be about decision environments, habits, and social skills.

6) Trend effects: yesterday changes how today feels

Here’s a bias that matters a lot on Moderate days, especially after a period of elevated danger.

Terum et al. (2023) ran experiments showing that people’s perception of current hazard is influenced by the recent trend in danger ratings, even when that trend is not actually relevant to the immediate decision. If danger has been trending down and today is “only Moderate,” people can feel relief. The number changed, so the emotional tone changes. You can feel the group exhale.

But the snowpack doesn’t always “exhale” on the same schedule as our emotions.

7) Overconfidence and “black swan” thinking

Bonini et al. (2015) looked at overconfidence and avalanche exposure using the “black swan” framing. For an individual skier, an avalanche accident is often a low probability event on any single outing, but with severe consequences when it happens. Overconfidence can push people to underestimate probability and increase willingness to go.

This is one of the hardest things to talk about without sounding insulting, because nobody wants to be told “you’re overconfident.” But the point isn’t that confident skiers are bad. The point is that confidence changes exposure. It changes what you consider “within tolerance.” And on a Moderate day, where the bulletin doesn’t force restraint, exposure is where the story gets written.

8) Rules and frameworks help, but Moderate can still slip through

A lot of people turn to decision frameworks to keep themselves honest. That’s not a bad instinct. But it has limits, and Moderate days are where those limits show up.

Landrø et al. (2020) reviewed and classified avalanche decision-making frameworks and the factors they include. Their work highlights an important reality: different frameworks emphasize different signals, and expert reasoning often combines analytical and experiential cues in ways that aren’t always captured by simple tools.

The other catch is that Moderate days often involve localized, terrain-specific problems. A day can be “Moderate overall,” but still have the kind of isolated instability that ruins your life if you happen to find it. That’s exactly why “I followed the framework” can still end up as a story that begins with “I don’t understand what happened.”

9) Forecast communication: what forecasters think they said vs what people hear

There’s also the simple communication gap: people don’t always absorb avalanche forecast information the way we assume they do. Research on how users process, retain, and apply forecasts is growing fast (including lots of work in ISSW proceedings). One broader way into that literature is the human factors scoping review by Hetland and colleagues: Hetland et al. (2024) scoping review preprint.

A very practical angle on interpretation is discussed in Ebert et al. (2025), which looks at differences between end-user and forecaster interpretations of the European avalanche danger scale and ties into related findings like Morgan’s “linear scale” result.

10) A field-honest model of what happens on “Moderate but high consequence” days

If you want a simple way to describe the Moderate accident pathway without turning it into a lecture, it often looks like this:

Moderate feels like permission, especially after higher danger. People interpret the scale too linearly, so behavior shifts only slightly. Heuristic cues stack up (scarcity, commitment, social proof), and group momentum forms. Exposure increases. Then a localized or low likelihood, high consequence problem bites.

That chain doesn’t require ignorance. It only requires being human.

What actually helps (without pretending we can “fix” humans)

This isn’t a “tips list,” but a few directions that are genuinely aligned with the research:

Build friction into the plan before the day starts. Moderate days are where you need pre-commitments most, because the day won’t force restraint. Decide what you will not do before you see the perfect slope.

Talk about goal pressure out loud. Commitment and scarcity traps become weaker when someone says the quiet thing in the open: “I know we drove all night and I want this badly.”

Treat Moderate as a real jump from Low. If risk can jump strongly from 1 to 2, your behavior should not shift by 5 percent. It should shift materially.

Watch for trend effects. If danger is trending down and the group is exhaling, name it. Trend effects are measurable. That should tell us we’re not immune.

Don’t outsource safety to confidence. Overconfidence isn’t a character flaw. It’s a cognitive trait that changes exposure. The fix isn’t self-shaming. The fix is designing habits that don’t rely on feeling confident.

Closing thought

Moderate days are the days where the mountains don’t scream “no,” so humans have to be the ones to say it.

And that’s why this topic matters. Because the hard part of avalanche safety isn’t memorizing terms or reciting red flags. It’s the moment when the line looks perfect, the group is humming, the bulletin feels permissive, and you realize the real decision is not technical.

It’s personal.

---

References and links

• Bonini, N. et al. (2015). Overconfident people are more exposed to “black swan” events: a case study of avalanche risk. PDF
• Ebert, P. A. et al. (2025). End user and forecaster interpretations of the European avalanche danger scale. Full text
• Greene, E. et al. (2006, ISSW). Fatal avalanche accidents and forecasted danger levels. PDF
• Hetland, A. et al. (2024). A scoping review of human factors in avalanche decision making (preprint). PDF
• Johnson, J. (2020). Rethinking the heuristic traps paradigm in avalanche education: Past, present and future. Full text
• Landrø, M. et al. (2020). Avalanche decision-making frameworks: Classification and description of underlying factors. Article
• McCammon, I. (2002, ISSW). Evidence of heuristic traps in recreational avalanche accidents. PDF
• McCammon, I. (2003/2004). Heuristic Traps in Recreational Avalanche Accidents. PDF
• Morgan, A. et al. (2023). A user perspective on the avalanche danger scale (Natural Hazards and Earth System Sciences). Article
• Terum, J. A. et al. (2023). Trend effects on perceived avalanche hazard (Risk Analysis). Article
• Winkler, K. et al. (2021). On the correlation between the forecast avalanche danger and avalanche risk taken by backcountry skiers in Switzerland (Cold Regions Science and Technology). Article
• Swiss SLF explainer (news). Avalanche risk recalculated. Link