How I Learned to Stop Worrying and Love the Bomb

Frank,

On Friday, January 17, 2014 10:38:48 PM UTC-5, Frank Whiteley wrote:

1/3333 beat the odds. 2007, 6, but one was most assuredly a suicide, so a skewed factor,
thus 5 is the acceptable number. Again, we're at 1/2000. 1/1800 just kind of looks about
right given what we know about the pilot population. The number is likely better than overall
odds of traffic, ladders, being a pedestrian, or cleaning the gutters, etc.

There's an article by Tom Knauff that comes up with very similar numbers:

http://www.eglider.org/NewsArticles/...mergencies.htm

Now, the thing is, coming up with informative statistics is difficult in this context, and it is difficult to put these statistics into perspective for the average glider pilot or person potentially interested in gliding.

As for the latter, a lot of this has to do with the fact that people are notoriously bad at judging and handling probabilities of rare events. No disrespect to anyone here. (I make my living by doing university research into people's cognitive "in"abilities, and people's mistakes in decision-making are very well established.)

Second, I would point out that while the probability of dying seems acceptable, risk (and "expected utility") has a second component: the cost in the event that something does happen. Many of us have lost a friend in a gliding accident. It was after I talked to the dad who lost his twenty-something year-old daughter (and my friend) in a gliding accident, that I understood what such fatalities mean to one's family, friends, club mates, workplace, and so on. I think we might underestimate this, and the numbers don't really do this any justice.

As for the statistics, I wish we could express the probability of a fatality per glider flight. If we assume 11,000 SSA members and 6.7 fatalities in the US annually, how many flights do people do, on average? If it's 20 flights (given that many people don't fly much), I get to 30 micromorts. The right way to do this would be to survey logged hours (ensuring random sampling and an acceptable response rate), and count accidents for sample population (i.e., sample from SSA membership, and count only SSA-member accidents). I think that the OLC database might be a good population, provided your result then applies to people enthusiastic and experienced enough to log their flights there.

I want to point out that risk expressed in terms of number of flights or flight hours is also difficult to interpret, because the relationship between flights and total risk is not linear. Highly inexperienced, not-current pilots are likely to be less safe. Very high-time pilots might also be less safe, as they would tend to fly in a broader range of conditions and over other terrain. That's why I would call the "10-years of risk equals 10 times .18 probability" perhaps a back-of-the envelope approximation, but also armchair statistics.

Gliding brings plenty of long-term health benefits as well. Many of us are 70 y/o and pretty fit, physically and mentally. Self-selection or causation? I don't know. We're not taking this into accounts so far.

A much more informative estimate might be one's individual assessment, or even our assessment of the other guy rigging his glider. Am I fit to fly today? Have I checked those control connections? Am I diligent with checklists? Do I know the weather? Am I current? And so on.

All of that said - nothing is a "safe" activity. We're all going to die of something. The question is: when that day comes, have we lived?