Decisions Under Stress

By Guy Higgins

I recently read a very disturbing book – Jon Krakauer’s Into Thin Air.  The book recounts the story of several simultaneous expeditions climbing Mount Everest in 1996.  A little bit of background first.  Mount Everest, the tallest mountain in the world at 29,028 feet, was first climbed in 1953 by Sir Edmund Hillary and Tenzing Norgay, his Sherpa guide.  Since then, the mountain has been climbed numerous times, and starting in the late 1980’s commercial ventures have guided paying customers to the summit.

Krakauer’s story focuses on the expedition, led by Rob Hall, of which he was a part, as well as the expedition led by Scott Fischer.  He also discusses the interaction these two expeditions had with other groups climbing Everest at the same time (Taiwanese, Japanese, South African groups and one solo climber).

For the record, there is very little air in the air at 29,028 feet above sea level.  In fact, commercial and military aircraft are pressurized above 10,000 feet, military pilots are breathing pure oxygen above about 30,000 feet (even pressurized), and if you lose pressurization above 35,000 feet, the oxygen system in military aircraft will initiate “pressure breathing.”  That’s where the system forces oxygen under pressure into your lungs and you must forcibly exhale (the opposite of normal breathing where you exhale when you relax).  All that is because we need oxygen – we don’t operate well without it.

Rob Hall and Scott Fischer were leading these two expeditions up Everest.  Both Hall and Fischer were experienced Everest guides and both had summited on prior expeditions.  They both started out doing all the right things – they arranged for their clients to spend time climbing at elevations above 17,000 feet (there’s darned little air in the air at 17,000 feet).  They had guides, they had pressurized oxygen bottles, and they were climbing at the right time of year.  Yet, on the day that members of the expeditions made the summit, twelve people, including both Hall and Fischer, died.  What happened?

Having read the book, there’s not an easy answer.  At least one member of Fischer’s expedition wasn’t really ready to climb Everest – not enough climbing experience.  One member of Hall’s expedition had missed summiting it 1995, had been strongly encouraged by Hall to come back (Hall even deeply discounted his fee) and try again.

Since Fischer and Hall both died, we’ll never know exactly why they did what they did, but both men independently made the same decision – to ignore their own “turn-around-no-later-than” time and continued to help their clients struggle on toward the summit.  We’re not talking about ignoring the time by ten minutes, but by two to five hours.  That’s two to five hours longer at very high altitude – far beyond their planned oxygen supply and in extreme cold.  At the same time, a violent storm was brewing at lower elevations and rushing up the glacial valley toward the climbers.

Both Fischer and Hall remained near the summit to help their clients reach the summit and get back.  Out of oxygen and caught by a storm with gale-force winds, they both continued to try to help their clients – sounds like the right thing to do, and it was, but it was also incredibly heroic.  These men knew they were risking their lives.

Hypothesizing, Fischer and Hall may have fallen victim to over-confidence, to the “sunk cost” fallacy, to confirmatory bias and possibly other cognitive dysfunctions.  Remember, they’re at 29,000 feet with low to no supplemental oxygen, working hard and trying to think.  All three of those things are hard.

Both guides had been to the summit before.  They knew what they were doing.  They believed in themselves.  They’d never lost a client.  They knew they could get everyone up and back.  The problem is that the future is not a mirror image of the past.  This expedition was different, even before they decided to ignore their no-later-than time.  Over confidence is dangerous.

The day was clear, and the wind, as they started their final assault on the summit, was less than normal.  The information they seemed to have been considering was good.  Staying longer looked like a very low-risk gamble.  The problem was that they either didn’t access all the information (the storm below was clearly visible) or they ignored it because it didn’t fit with their decision.  The failure to use all the information, whether through oversight or cognitive bias is dangerous.

Both men, but particularly Hall, felt that it was very important to help their clients summit.  Remember, Hall urged one client to come back.  This client had been compelled to turn around the year before only 300 feet below the summit.  Hall may have felt that he could not do that to his client two years in a row.  All the climbers had invested not only thousands of dollars, but also weeks of their lives, and more importantly enormous physical discomfort and even pain to get to the “top of the world.”  The sunk cost in money, time and pain certainly played some part in the collective decisions (climbers can and often do turn around on their own) to continue to the summit.  It’s not the sunk cost, but the cost yet to be paid that counts.  In this case, that cost for twelve people was their lives.

Most of us are never going to have to make a decision while suffering from severe oxygen deficit, crushing fatigue and extreme cold.  Still, we will all be faced with decisions under stress – decisions where we can be over confident, where we can pick and choose the data that supports what we want to do.  We can consider only the sunk cost and not adequately consider the future price.  For all decision makers, but particularly for leaders – and particularly in stressful situations (such as decision making during emergencies and crises) – it’s important to understand that our natural human tendency is to be subject to cognitive biases.  We need to understand, recognize and mitigate these tendencies.

Into Thin Air disturbed me because good people doing something challenging and fulfilling behaved much the way that I might have – and they died.

4 thoughts on “Decisions Under Stress

  1. My opinion: Humans fail for 4 reasons:

    1. Lack of Attention (Situation Awareness)
    20 minute limit

    2. Fail to Perceive
    Inability to understand the situation

    3. Memory
    Short term memory (7 +/- 2 items)

    4. Logical Reasoning
    Poor judgment

    When humans are placed under stress, different people (personality types) perform differently (Yerkes-Dodson / Hans Eysenck). http://www.compsim.com/demos/d58/Yerkes-Dodson.html

    So humans as an integral part of safety critical systems is a risk. Bad things will eventually happen.

  2. Tom, there’s a lot of truth in what you say, but it’s also important to remember that, under stress, those limits might change. Studies of fighter pilots showed improved performance with increased information — but only up to a point. That point was well beyond normal information processing, but the collapse of good reasoning was then abrupt. There are many more reasons that people fail to make good decisions, and most of them fall under the heading of cognitive dysfunctions — normal human cognitive behavior includes those dysfunctions. They are not irremediable, but they are our natural way of behaving unless we understand them and act to mitigate them — recognizing that they are natural behaviors and we will revert to them given any opportunity to do so. Rob Hall had an overconfidence issue that allowed him to put himself into extremis — over confidence is one of those dysfunctions. You can’t fix it by simply thinking better or harder. It takes other remedial action. The research by Kahnemann, Tversky, Ariely et al provides insight.

  3. Not that I have studied the tests/work of Tversky & Kahneman, but their work has been compared to the KEEL model for judgment and reasoning, where negatives have more impact than positives. In our case, “impossible” overcomes “must”. Our focus is still creating machines that have 100% explainable and auditable behaviors and that are traceable to human policy makers. So when you mass produce a machine it doesn’t have the human failings. Machines (my opinion) do not need to participate in human evolution. They are just machines that don’t need to fail like humans. The hope is that human policy makers can create policies (without being under stress during their policy making role), can still account for stressful conditions, and can make better policies. The problem is that “humans” aren’t really happy about having their policies “explicit and auditable and traceable to them”. It is sometimes better to allow humans to fall back on their unauditable judgment and reasoning. (aka human evolution)????

  4. Tom, Deep Blue can win any chess match because chess is completely rule driven. There is no machine that can beat Michael Jordan at basketball because basketball is rule and reaction driven — the results of any decision are not predictable through anything less than an infinite set of possibilities. Today (and as far as I can see), a machine can perform very well in an environment that is “laws of physics” driven, but not in one that is complex adaptive systems driven — and that is where most human activity exists. Scott Fischer and Rob Hall died because they were making complex adaptive decisions under enormous physical and psychological stress. Your machine would make a different decision but it may not have turned out any differently because Fischer’s and Hall’s clients were not compelled to obey them. How does your machine respond in that situation — which may, in fact, have happened?

Leave a Reply

Your email address will not be published. Required fields are marked *