Are we all above average?

In Garrison Keillor's Lake Wobegon, all the women are strong, all the men are good-looking, and all the children are above average.

When presented with GA "little plane" accident statistics, a pilot of such planes often has a predictable response, namely that he's not too worried about accident stats, because he's an above average pilot when it comes to safety.

We know that pilots are both strong and good-looking, but are we all above average?

Well, perhaps we shouldn't say "all of us", as the corollary to this sentiment is that GA accidents are the domain of a tiny subset of the pilot population who are knuckle-heads, the bumbling incompetents who:

What would the accident statistics look like if all pilots were as safety-minded, skilled, and of course, good-looking, as we are?

Looking at the stats

According to the 2007 Nall Report, only 10% of fatal accidents are attributed to "Mechanical/Maintenance" causes, leaving 80% attributed to the pilot, and 10% for which the cause is unknown (implying, of course, that if the cause was known, some chunk of these could be attributed to the pilot as well).

So while we spend a lot of time practicing engine-out maneuvers, the component most likely to fail is, as they say, "the loose nut between the yoke and the seat".

This leads to a very satisfying conclusion: GA planes crash because bad pilots crash them. Of course, since I'm not a bad pilot, I would never do that, and so I'm not subject to 80% of the risk. Right?

Well, a perception that's popular, for obvious reasons, is that if you eliminate just a handful of the most common bone-headed stunts, mistakes clearly made only by bone-headed pilots, then you've eliminated the lion's share of risk.

So let's go back to the Nall report, and see if we can identify these "few bone-headed mistakes".

Let's start with fuel management, which includes both running out of gas, and also the ultra-boneheaded "plane had gas, but pilot had fuel selector to wrong tank". In 2007, there were 11 fatal accidents attributed to this cause, or 4% of all the fatals.

How about flying under the influence? This gives us another 5 fatals, or about 2%.

How about "buzzing"? Well, this gets a bit complicated, as the Nall Report throws all of the following into the same "Maneuvering" bucket: aerobatics, low pass, buzzing, pull-up, aerial application maneuver, turn to reverse direction (box-canyon-type maneuver), or engine failure after takeoff and pilot tries to return to runway. Accidents attributed to any of these causes represents about 25% of fatals. The Nall Report does break these down into "Aerobatics", "Lost Control", and "Hit Terrain/Wires/Trees", but it's difficult to say what percentage of these results stemmed from actions that were obviously reckless in nature, e.g. "buzzing".

What about continuing VFR flight into IFR conditions? The Nall Report lists 20 such fatal accidents in 2007, accounting for another 7%.


Even considering that we're throwing all maneuvering accidents in with our "boneheaded" categories, we've still only accounted for about 38% of fatal GA accidents. It seems clear that we are unable to point to a few "mistakes" that we can simply "choose to not make", and thus eliminate the bulk of our risk. The uncomfortable truth is that pilots find numerous and varied ways to kill themselves, and that there are thus no easy answers. Ultimately, we're left with little more than the notion that it just boils down to making the right decisions. Unfortunately, on this front, a pilot is only as good as he is on his worst day.


While it's obvious why pilots want to believe that "bad things happen to bad pilots (and only bad pilots)", and thus claim personal risk that's a small fraction of that demonstrated by the overall GA pilot population, it's virtually impossible to demonstrate that. We cannot even demonstrate whether experienced pilots are more or less safe than inexperienced pilots, because the statistics just aren't available.

It's entirely possible that the pilot population consists of a "bottom 2%", and that accidents occur primarily to those "two percenters", but we have no evidence to support or dispel that notion. What we do have, though, is recent anecdotal evidence of extremely experienced and accomplished pilots, such as Steve Fossett and Scott Crossfield, dying in what appear to be very typical GA accidents. If we're going to make the case that "we're better than accident pilots", then we do have to consider how we reconcile that notion with this anecdotal evidence.

The number of pilots who die attempting reckless "stunts" does not appear to be very high, the bulk of fatal GA accidents are the result of very mundane judgments by pilots. Very experienced pilots usually have one or two "close call" stories, typified by magazine features named "Never Again" and "I Learned About Flying From That". It's often clear, from these stories, that faults in judgment by a pilot who would have no reason to suspect he was being intentionally reckless or bone-headed, led him into a high-risk situation where if chance had swung the wrong way, his "Never Again" story could have been an NTSB report, instead.

In my opinion, regardless of its truth or untruth, I find the notion that there are "accident pilots", and that we can identify ourselves as being outside that group, as wrong-headed and possibly dangerous. Given the anecdotal evidence, and at least the strong possibility, that we are all subject to these lapses in judgment, I find it far safer to assume that I'm just as subject to the kind of lapses in judgment that lead to accidents due to causes such as fuel management or VFR flight into IFR conditions, requiring me to maintain constant vigilance of my own decision-making.

Harry Mantakos /