Homebuilt Accidents: Fatal Factors

Pilot decision-making is a greater factor in fatal accidents.


To the owner of an Experimental Amateur-Built (E/A-B) aircraft, any accident is a tragedy. However, the FAA monitors the homebuilt accident rate based on fatal accidents only.

It’s understandable, in some ways. Unlike owners of production aircraft, builders of homebuilt aircraft are usually very familiar with disassembling and transporting a bent bird. A minor enough accident and the temptation is strong to grab the pieces and stuff them out of sight before the local constabulary shows up.

I know of at least two cases where the owner and his friends dashed the wreckage into a hangar and shut the door before outsiders arrived. One man swore up and down to a news crew that he didn’t know anything about a plane accident. He was standing with one arm in a sling and a crunched rudder tucked under the other.

But accidents resulting in death are nearly impossible to hide. They receive the highest attention from investigating agencies. This is good, as they are the cases where the builder or pilot made the worst kind of mistakes. By learning what is more likely to kill us, we have a better chance of avoiding a serious accident ourselves.

Injury Severity

We can find the fatal accidents through the “Injury Severity” entry in the NTSB records. Searchers using their web page can select only “Fatal” and “Non-Fatal,” but the downloadable database actually has four injury categories: Fatal, Serious, Minor, and None.

“Fatal” is unambiguous. But to be considered as related to a given accident, the NTSB requires that the death occur within 30 days. Injuries are rated as “Serious” if they require hospitalization for more than 48 hours, involve broken bones (other than simple fractures of fingers, toes, or nose), severe bleeding, and/ or 2nd or 3rd degree burns over more than 5% of the body surface.

The NTSB doesn’t define “Minor” injuries, but one can assume that it’s a level less than “Serious” but still not “None.”

For the analysis to come, we’ll combine the “None” and “Minor” categories.

Injury severity for E/A-B accidents 1998-2017.

What Kills Us

Pilot miscontrol—errors made in the control or guidance of the aircraft—is the major cause of homebuilt accidents regardless of injury severity. What may be surprising is that the percentage of pilot miscontrol decreases with accident severity: As seen in Figure 1, pilot miscontrol is the major cause for over 43% of the “Minor/None” injury category, but slightly less than 30% in the fatal cases.

Figure 1: Pilot miscontrol rate vs. severity of injuries

So, what else kills us? Figure 2 compares the rate of occurrence of various causes when pilot miscontrol is excluded. Builder and maintainer errors are involved less as the severity increases. Simple engine failures (engine mechanical, fuel system issues, running out of gas, carburetor icing, etc.) are less likely to result in fatalities, as well. Other than pilot miscontrol, then, what drives fatal accident statistics?

Figure 2: Non-miscontrol E/A-B aircraft accident causes vs. injury severity.
The Buccaneer struck the power lines while searching for a lost life vest on the river. Dusk and cloud cover likely made the lines difficult to see. (Photo: National Transportation Safety Board)

Pilot judgment issues. Nearly a quarter of the cases involve a deliberate decision by the pilot—mostly unnecessary low flying or continued VFR into IFR.

We can’t compute a “risk factor” associated with these practices since we don’t know how often pilots successfully buzz or scud run. We generally just find out if they die doing it.

But what the statistics say is that if you do crash during one of these activities, it’s probably going to kill you. Two-thirds of all low flying accidents are fatal, as is a stunning 83% of crashes resulting from continued VFR into IFR conditions.

Fatal Accident Rate

As mentioned earlier, about a quarter of homebuilt accidents result in fatalities. Figure 3 plots the “Fatal Accident Rate” (the percentage of accidents that produce fatalities) per year from 1998 through 2017. Overall, for that 20-year period, the average is 24.3%.

Figure 3: E/A-B aircraft fatality rate.

The overall GA rate is about 18%, so homebuilts are a bit worse. Figure 4 compares the E/A-B fatal accident rate against that of several types of common production GA aircraft.

Figure 4: E/A-B fatal accident rate vs. common production aircraft.

What drives whether an accident results in a fatality?

Two things: The first is the aircraft’s speed at impact. The energy involved in an accident depends on the weight of the aircraft multiplied by the square of its speed at impact. A Lancair IV on approach has about six times the energy of a Pietenpol. Which leads to the second factor: If an accident occurs, the occupants’ survival depends on the ability of the structure to protect the occupants from the liberated energy.

Generally, designers of homebuilts haven’t included crashworthiness when developing their aircraft. But it wasn’t a factor considered during production of most GA aircraft, either. Sure, airplanes designed to modern standards like the Cirrus and Diamond have occupant-protection features. But the average light GA aircraft is 45 years old. Safety features like airbags and deformable structures weren’t too common back then.

Looking back at the overall picture, it seems logical that fast airplanes would have a higher fatality rate. Figure 5 plots the rates for a variety of homebuilt and production-aircraft types against the aircraft’s cruise speed. (Why cruise instead of stall speed? Because its cruise speed will reflect how quickly an airplane might accelerate after a loss of control. It’s interesting to note that most lethal of accidents—continued VFR into IFR conditions, buzzing, etc.—do occur near cruise speed.) As expected, a higher fatality rate is found among airplanes that can fly faster.

Figure 5: Fatal accident rate vs. cruise speed and wing configuration.

Here’s a bit of a curiosity: Figure 5 also shows the wing position of each aircraft—high, low, or mid (typically canards). Note when a speed range has both low- and high-wing airplanes: The fatality rates of the high-wing airplanes are almost always lower.

Why is this? A high-wing airplane generally surrounds the cabin with a structural cage. As long as the occupants are prevented from bashing their heads into that structure, a high-wing airplane probably offers better protection. Speaking of protection, the figure shows three “ultralight type pushers” at the low end of the speed range with an elevated fatality rate. Sure, they fly slow. But the occupants sit forward of the main structure of the aircraft and aren’t well protected by it. Their fatal accident rate probably reflects less protection that results from this configuration.

High-wing aircraft have more protective structure around the occupants.

It probably comes to no one’s surprise that the Lancair IV has the highest fatal accident rate. But if one projects a straight line through the middle of the results (ignoring the ultralight-style pushers) to approximate a speed vs. fatality rate ratio, you’ll find the Lancair IV falls generally on that line. Given its high cruise speed, its fatality rate is about what one would expect.

Finally—note that the results for the homebuilt airplanes (blue symbols) are very similar to those of production airplanes (red symbols). E/A-B aircraft are actually doing pretty well.

Wrap Up

Ultralight-style aircraft provide the occupants a tremendous flight experience, but the degree of protection in an accident is often less.

It was pleasant to find that the fatal accident rate for most homebuilts is proportionate and roughly equal to that of production GA aircraft. Ultralight-style homebuilts don’t offer as much protection, but the lack of surrounding structure is part of the appeal.

So why is the overall GA rate 18%, vs ~25% for homebuilts? Probably because nearly a third of the single-engined GA fleet are Cessnas: High winged, strut-braced, extremely modest performance airplanes that fly slower and protect the occupants better.

Being a better pilot certainly increases your chance of survival, but your decision-making has a disproportionate effect. As a group, homebuilt owners should really cut back on the low-altitude stuff and the scud running.

I feel like a bit of a hypocrite saying that—I was once young, foolish, and thought myself immortal. But I learned my lesson (without an accident), and have been on the righteous path for the past 30 years. Go thou and do likewise.

This article originally appeared in the October 2019 issue of Kitplanes magazine.

For more great content like this, subscribe to Kitplanes!

Ron Wanttaja
Ron Wanttaja is a systems engineer, engaged in satellite orbit/constellation design and analysis, launch vehicle and onboard propulsion system trades, and operations concepts for space systems. He worked on the early design studies for the International Space Station.

Other AVwebflash Articles


  1. Ron Wanttaja posts an interesting article but there are several errors in his analysis that are worth mentioning. He states, “It’s interesting to note that most lethal of accidents—continued VFR into IFR conditions, buzzing, etc.—do occur near cruise speed.) As expected, a higher fatality rate is found among airplanes that can fly faster.” That fact is not born out by research that I participated in with the GAJSC that shows most fatal GA accidents involve a loss of control and are low speed events fmi see: https://www.gajsc.org/document-center/ Faster does not correlate to more fatal accidents, otherwise turbine aircraft would logically lead the way.

    As he notes, not all accidents get reported so his denominators are suspect (at best). Many accidents are not reported or incorrectly reported or coded. I found many Lancair accidents were incorrectly coded by the FAA and NTSB. After the introduction of the Lancair IVP almost every Lancair accident was coded as a Lancair IVP, even if it was a Lancair 320 or 360. Aircraft recognition is not their strong suit.

    As I said, the denominator is important when discussing accident rates. I encountered this several years ago when an FAA internal memo incorrectly noted a super majority of Lancair accidents had fatal results. They said, “Analysis of Lancair accidents revealed that the Lancair is over-represented in the experimental aircraft accident rate, and was not improving. For instance, Lancair aircraft only represented 3.2% of the experimental fleet, yet in FY 2008, they accounted for 6.6% of the accidents with a lethality rate approaching 60%. These percentages were worse in FY 2012 —Lancair accounted for 8.2% of the accidents with a lethality rate of 100%.” Simply not true. There were 30 accidents and incidents in the U.S. in 2008 with 11 of those fatal. Certainly not a good year but not 100%. As of 2016, when I published a white paper on the subject (found here: https://www.lancairowners.com/files/wp-content/uploads/2016-LOBO-White-Paper.pdf), our fleet of Lancairs had experienced 557 known accidents and incidents since 1988, including 116 fatal accidents involving 192 fatalities. 372 accidents and incidents involved no injuries at all. Efforts to reduce the fatal Lancair accident rate has paid off with just two fatal accidents last year but, the work continues to have an accident free fleet.

    Pilots are often their own worst enemy. LOBO commissioned a safety study that utilized the GAJSC protocols for examining the data. The white paper reports on the research and found that nearly a majority (261) of the accidents involved a failure by pilots to follow procedures. Many of the accident pilots had no training in the aircraft they were flying. If pilots would simply take training from a recognized and well respected training provider and if the training community would standardize their training syllabus across their fleets like Robinson, COPA, Cirrus and the MU2 fleet have done, then we would see a serious reduction in the accident rate.

    I appreciate Ron’s work on this subject and thank him for his thought provoking discussion.

    Best regards,

    Jeff Edwards, PhD, ATP and founder Lancair Owners and Builders Organization (LOBO)

  2. the problem might be that faster airplanes have less lift coefficient or wing area in comparison to wingload/weight,,,,so low speed is more dangerous to such airfoils and wing configurations than on a low wingloaded slow design…

    so why nobody actually gets such an idea by seeing that data i don,t get it,,,…the correlations in thise statistics are blurred completely as it seems, as Canadian UL data shows that the accident rate of registered and insured but besides that completely identical ultralights in Canada, mostly low maintenance and often homebuilt is half of those of GA airplanes🤷‍♀️the basic ULs there don,t need any checks or whatever only registration and insurance and the only difference is 25 kilos/ m2 and minimum of 10 m2 wing area.l.l…which leads to bigger designs, but besides that the wing loading of Us UL standards is basically even lower! 🤷‍♀️i leanrt and compared that for more than a year alone in many designs and sices…
    so most accidents even come from restrictive max. engine powers which lead to lack of climb out! but besides that a more heavy bird with 10 m2 than mostly minimum of 7 m2( below that an UL is technically more or less impossible even with the lightest engines, stall speed minimums give that limit…no chance for less, all others are wrongly classified experimentals with high wing loads and as far as i was able to get drawings and blueprints mostly a crap of zero calculated design features regarding G loads! nor safety) …
    so not fast airpkanes crash more often but more easily with their small wings at low speed…its simpky always stall🤷‍♀️stall under rain, under crisswings, tight banks/turns, steep climbing, brainless testosterone behaviour( aerobatics as if every shit we buy MUST endure every overload….but nope, it doesn,t! 🤷‍♀️)
    low floght is not dangerous sorry, but idioticly watwr skidding or sand skidding , zero planning…that is dangerous! and when low and slow YOU JEED A LOW WING LOAD AND HIGH LIFT SURFACE, WHICH CAN BE DESIGNED AND NOT TRIED OUT WITH SOME CRAPPY OLD RESALE FROM A GARAGE HOMEBUILT! )
    so it.s behaviour and ignorance that leads to accidents! 🤷‍♀️using something inappropriate for hazzardous application or simply wrong usage…case closed. people should study more, read manuals, get a license or some training or the knowledge from lesewhere,,,,
    ttjanks for the statistics, but i assume those show US behavior a lot…ignorance leads to wrong expectations , lack of knowledge leads to such idiocies as STOL contests, sorry..lbut adter i learnt how to calculate lift in correlation to speed amd weight STOL competitions made zero sense to me and lost all reason for doing those,l.🤷‍♀️