Single-Seat Accident Records (Was BD-5B)

With the current discussion in another thread regarding the accident record
of the BD-5, I thought it might be interesting to do a little more in-depth
look. I decided to examine the accidents between 1 January 1990 and 13
November 2003 and do a breakdown as to relative accident/fatality rate,
pilot experience, and cause of accidents.

To give a little depth to the situation, I also ran the same analysis for
two other single-seat homebuilts. I selected the RV-3 as an airplane with
a similar "mission" as the BD-5. And I use the Fly Baby because, well, I
ALREADY had the downloaded NTSB files on it. :-)

DATA SOURCE:

I downloaded the accident reports from the NTSB web pages for these three
aircraft. In all cases, I used several variants of the names to try to
avoid missing any reports.

I also ran a simple search for all amateur-built accidents and
amateur-built accidents with fatalities over the same period.

I do not have any data indicating how many of each type of aircraft existed
in each year. Therefore, I used the number of each type registered in
January 2003 for the basis for some of the percentages. I used the FAA/EAA
practice of only counting those airplanes which are listed as having
airworthiness certificates, even though the listings are somewhat
inaccurate. Thus, the following fleet-size figures are used:

Fly Baby: 265
BD-5: 81
RV-3: 170

Using the January 2003 fleet size numbers is only for comparison; remember,
it does not reflect the number of a given aircraft in the fleet at the time
of any particular accident.

When running percentages based on this fleet size, I did *not* add back the
crashed airplanes into the total fleet size. This is different from my
earlier, BD-5-only report, so the percentages will be slightly different.

ANALYSIS METHOD:

Using the NTSB reports, I extracted key features and transferred them to a
database. The database included items like the type of aircraft, date of
accidents, number of fatalities and injuries, the ratings held by the
pilot, whether the aircraft was on its first flight, a "test flight", or
whether the pilot had newly purchased the plane, and 36 separate categories
for the accident cause.

I would enter the NTSB primary cause as a "1" in each category, and any
related events as "2"s. However, when developing the cause statistics, I
ignored the primary/secondary indications and just counted whether the
categorized item was a factor.

I did vary from the NTSB cause determination in one kind of case: When an
accident occurred after a loss of power. Often, the NTSB assigns pilot
error as a cause (either primary or contributing) if the pilot in not able
to safely land the aircraft after an engine failure. In my analysis, I
considered only the even that *caused* the emergency; if the pilot
misjudged a deadstick landing approach, I did not flag that as an accident
cause.

Keep in mind that the statistical base is NOT high for any of these three
aircraft. We're talking 15-20 accidents over a 13-year period, here. But
things are easier to compare when presented in percentages.

Finally, the data presented may appear better if you switch your newsreader
to a fixed-spacing font rather than variable spacing.

OK, on to the results!

OVERALL ACCIDENTS:

Here are the total number of accidents I extracted from the NTSB online
database. Again, this is the period from 1 Jan 1990 to 13 November 2003.

Type Accidents fatal
All Homebuilts 2881 837
Fly Baby 15 5
BD-5 22 9
RV-3 14 5

OVERALL ACCIDENT RATE:

Let's look at the overall accident rate for the homebuilt fleet and the
individual types. The "Accidents" is the total number of accidents vs. the
number of that type on the Jan 2003 register, and the "Fatals" column is
the number of accidents in which the pilot was killed.

Accidents Fatals
All Homebuilts 11.1% 3.2%
Fly Baby 5.7% 1.9%
BD-5 27.2% 11.1%
RV-3 8.2% 2.9%

FATAL ACCIDENTS TO TOTAL ACCIDENTS

These ratios show the relative number of fatal accidents vs. the total
accidents for that type.

Type Fatal Rate
All Homebuilts 29.1%
Fly Baby 33.3%
BD-5 40.9%
RV-3 35.7%

The number of accidents involving fatalities is one indicator of the
relative crash-worthiness of a design. However, these factors do not
reflect the type of accident...a structural failure at altitude is
generally not survivable, no matter the design of the aircraft.

PILOT QUALIFICATIONS

The NTSB reports, especially the older ones, did not always list the
ratings held by the accident pilot. Less than half the reports contained
the pilot rating information. For the purposes of this analysis, I have
presumed that the pilots whose qualifications were not listed had Private
licenses.

Private Comm ATP
Fly Baby 80.0% 13.3% 0.0% (One unlicensed pilot)
BD-5 63.6% 18.2% 18.2%
RV-3 71.4% 21.4% 7.1%

Again, keep in mind the small sample size here. The 7.1% for ATPs involved
in RV-3 crashes represents *one* individual.

AIRCRAFT TEST STATUS

Occasionally, the NTSB reports indicate that the airplane was on its first
flight, or was newly-purchased. Or, they may indicate that the airplane
was on a "test flight." It is unknown whether other accidents actually
occurred during these phases; the NTSB investigator may not have been aware
of it or didn't consider that it was a significant contributor to the cause
of the accident.

"First flight": Aircraft specifically identified as never having flown
before.

"Test Phase": Aircraft identified as still in its FAA-specified test
phase, or the NTSB report indicates the purpose was a test flight.

"New Pilot": NTSB report says the owner had recently purchased the
aircraft.

First Flight Test Period New Pilot
Fly Baby 0.0% 0.0% 13.3%
BD-5 9.1% 22.7% 9.1%
RV-3 0.0% 0.0% 21.4%

For the Fly Baby, a single accident was 6.7% of the total, the BD-4 was
about 4.5%, and the RV-3 was about 7.1%.

ACCIDENT CAUSES

As mentioned earlier, I include both the primary and any secondary causes
in this list, with the exception of any pilot error occurring after a loss
of power.

The accident categories I use a

"LOP - Non-Pilot": A loss of power that is not attributable to the pilot's
actions. This could include blockage of a fuel line, magneto failure, etc.
This category is also used for unexplained losses of power.
"LOP - Pilot": A loss of power attributable to pilot actions...fuel
exhaustion, carb ice, etc.
"Pilot Judgement/ mishandling": The classic "Pilot Error" category. These
factors include both cases of bad decision-making as well as those of
losing control of the aircraft due to windy conditions. Keep in mind that
this really isn't the only "pilot error" category..."LOP - Pilot," "Density
altitude,", "Maneuvering at Low Altitude," etc. are cataloged individually
and are NOT reflected here.
"VFR to IFR": Basically, scud running.
"Mechanical Failure - Airframe": Generally, structural failure of some
sort, though not necessarily of critical structure. Includes door, canopy,
panel, etc. failures.
"Mechanical Failure - other": Generally failures of non-engine or fuel
system accessories. Does include propeller and spinner failures.
"Maneuvering Low Alt": Unnecessary low flying. Does not include cases of
high-altitude flying where the terrain outclimbs the airplane.
"Density Altitude": Hot/High conditions
"Builder error:" Workmanship problems, deviation from instructions, etc.
"Mechanical Failure - Maintenance": Failures due to mistakes while
performing maintenance...leaving bolts out, etc.
"Midair": Collision with another aircraft.
"Inadequate Preflight": Failure to detect flawed conditions or plan the
flight properly.
"Inexperience": Lack of experience, either in total time or in the
accident airplane.
"Fire": Pre-crash fires.
"Undetermined": No surviving witnesses and the wreckage is too damaged to
show any potential mechanical flaws.

Again, keep in mind that both primary and secondary causes are included
here, so the columns won't add up to 100%.

(FB is Fly Baby)

FB BD-5 RV-3
LOP - Non-Pilot 20.0% 40.9% 14.3%
LOP - Pilot 13.3% 0.0% 14.3%
Pilot Judgment/ mishandling 13.3% 36.4% 28.6%
VFR to IFR 6.7% 0.0% 7.1%
Mechanical Failure - Airframe 13.3% 0.0% 21.4%
Mechanical Failure - other 0.0% 0.0% 14.3%
Maneuvering Low Alt 0.0% 0.0% 7.1%
Density Altitude 0.0% 4.5% 0.0%
Builder error 13.3% 4.5% 0.0%
Mechanical Failure - Maint. 20.0% 9.1% 7.1%
Midair 0.0% 0.0% 7.1%
Inadequate Preflight 6.7% 0.0% 0.0%
Inexperience 6.7% 18.2% 7.1%
Fire 0.0% 0.0% 7.1%
Undetermined 6.7% 0.0% 0.0%

Percentage one accident
represents: 6.7% 4.5% 7.1%

The BD-5 has a loss of power accident rate twice that of the Fly Baby and
almost three times of the RV-3, reflecting the design's long problems with
finding a reliable powerplant. The Pilot Judgment/Mishandling categories
are fairly close for the RV-3 and BD-5, but the BD-5 has trigear and the
RV-3 is a taildragger. The Fly Baby, though, is half their rates, which
probably reflects the BD-5/RV-3 high-performance status. The BD-5's
"Inexperience" results are interesting, too.

Anyway, that's what the statistics look like.

Ron Wanttaja

Great job Ron, thanks.
Could you tell me where to find the information on the RV-3 midair?

Jerry

Ron Wanttaja wrote:
With the current discussion in another thread regarding the accident record
of the BD-5, I thought it might be interesting to do a little more in-depth
look. I decided to examine the accidents between 1 January 1990 and 13
November 2003 and do a breakdown as to relative accident/fatality rate,
pilot experience, and cause of accidents.

To give a little depth to the situation, I also ran the same analysis for
two other single-seat homebuilts. I selected the RV-3 as an airplane with
a similar "mission" as the BD-5. And I use the Fly Baby because, well, I
ALREADY had the downloaded NTSB files on it. :-)


snip

On Sat, 15 Nov 2003 22:03:25 GMT, Jerry Springer
wrote:

Great job Ron, thanks.
Could you tell me where to find the information on the RV-3 midair?

Go to:

http://www.ntsb.gov/ntsb/query.asp

and paste:

ATL02LA012A

into the NTSB accident number field, near the bottom of the form.

I've pasted the full narrative below.

Accident occurred as the airplanes were touching down...no injuries!
Neither had seen the other, nor heard him on the radio.

Ron Wanttaja
-----------------------------------------
NTSB Narrative report:
ATL02LA012A
On November 10, 2001, an Extra-Flugzeugbau GMBH EA300/L, N2XA, registered
to Aero Sport Inc., operating as a 14 CFR Part 91 demonstration flight, and
a James D. Smith RV-3, N93HS, registered to a private owner, operating as a
14 CFR Part 91 personal flight, experienced a mid-air collision during
landing flare touchdown at the St. Augustine Airport (SGJ), St. Augustine,
Florida. Visual meteorological conditions prevailed and no flight plan was
filed for either airplane. N2XA sustained minor damage and N93HS sustained
substantial damage. The airline transport-rated pilot and private
pilot-rated passenger on N2XA reported no injuries. The commercial pilot on
N93HS reported no injuries. N2XA departed from SGJ about 30 minutes before
the accident. N93HS departed from Haller Airpark (7FL4), Green Cove
Springs, Florida, about 30 minutes before the accident.

The pilot of N2XA stated he departed SGJ about 30 minutes before the
accident to conduct a demonstration flight to a prospective buyer. They
flew to a training area located about 20 miles southwest of SGJ and he
demonstrated the airplane. Upon completion of the maneuvers, they returned
to SGJ. They attempted radio contact with SGJ UNICOM was which uneventful,
however they overheard other airplanes indicating that they were landing to
runway 02. They continued toward the airport making radio calls in the
blind concerning their position and made a 45-degree entry for a 800 feet
left downwind to runway 02 behind a Cessna airplane. The Cessna landed and
reported clearing the active runway while they were turning from base to
final approach. The left wing of his airplane was down for a required left
crosswind landing. He started to flare the airplane just past the numbers
and felt a "bump" similar to a hard landing and observed what he thought
was a wheel pan go past the canopy of his airplane. At first he thought he
would make a go-around but elected to land straight ahead. Upon completion
of the landing, he turned his airplane 90-degrees on the runway and
observed an RV-3 on the runway. He continued turning his airplane and back
taxied down the runway to the RV-3, stopped, and shut his airplane down. At
no time during his entry into the traffic pattern or during his approach
did he or his passenger hear any radio communication from the RV-3 pilot.

The pilot of N93HS stated he departed 7FL4 about 30 minutes before the
accident. He approached SGJ from the northwest and attempted to contact SGJ
UNICOM which was uneventful. He overheard other airplane pilot's on the
UNICOM radio (A Cessna and a Piper pilot indicated they were landing to
runway 02 and a Falcon jet pilot indicated he was landing to runway 31.) He
decided to over-fly the airport at 1,500 feet to verify the wind condition
and speed. He over flew the airport and entered a left downwind at 900
feet. While on final approach for runway 02, he observed the Falcon jet
back-taxing down runway 02. He initiated a go-around, reentered left
traffic at 900 feet behind a Cessna. The Cessna landed and cleared the
active runway. He reported his position on UNICOM on short final. As the
wheels of his airplane were touching down on the runway, he immediately
felt an impact coming from above. He looked and observed the Extra on the
runway. He stopped his airplane and exited unassisted. At no time while he
was in the traffic pattern did he see or hear the Extra pilot on the UNICOM
frequency.

Examination of N93HS by the FAA revealed that both wheels of N2XA struck
the top of N93HS left wing and left elevator. In addition, N2XA right wing
struck the top of N93HS vertical stabilizer and canopy. (For additional
information see FAA Aviation Safety Inspector Statement an attachment to
this report.)

Review of 14 CFR Part 91.113 Right-of way rules states in paragraph (g)
Landing...."When two or more aircraft are approaching an airport for the
purpose of landing, the aircraft at the lower altitude has the
right-of-way."

In article , Ron Wanttaja
writes:


Anyway, that's what the statistics look like.

Ron Wanttaja


Great work and information...Thanks Ron.


Bob Reed
www.kisbuild.r-a-reed-assoc.com (KIS Builders Site)
KIS Cruiser in progress...Slow but steady progress....

"Ladies and Gentlemen, take my advice,
pull down your pants and Slide on the Ice!"
(M.A.S.H. Sidney Freedman)

On Sat, 15 Nov 2003 20:44:32 GMT, Ron Wanttaja
wrote:

With the current discussion in another thread regarding the accident record
of the BD-5,

it would be interesting to see the correlation between survivability
and stall speed.
I suspect that there is more of a correlation between those two
factors than which design was involved.

Stealth Pilot

On Sat, 15 Nov 2003 22:03:25 GMT, Jerry Springer
wrote:

Great job Ron, thanks.
Could you tell me where to find the information on the RV-3 midair?

Jerry
++++++++++++++++++++++++++++++++++++++

I'm surprised it's just ONE.

I could report several near misses a week.
Even on short finals of 500 feet, too many
doods take off directly in front of me.
The answer is ALWAYS the same.
Sorry, I just didn't SEE you.

The last really frightening near miss was with a T-6
at 4500 feet. I was westbound. The T-6, going east.
Yes, he could have been climbing through 4500, but
when I made a 180 and overtook him......
Yep, still sitting at 4500 feet and totally oblivious to me
now safely off his left wing. He and his passenger were
having a grand time just yucking it up from what could
be observed as total indifference to air traffic and the
VFR altitude scheme.

At 5500 feet or more, the sky has been relatively
event free over the last three years and 500 hours.
So, that's where I can be found on flights even as
short as 40 miles.... and I try to stay in the hangar for
less than 50 miles. A 50 mile flight is usually around
20 minutes on the tach.

I suspect a WHITE RV3 can be equated as just another
invisible motorcycle...
even though strobes flash anytime the engine is running.
Perhaps, a set of H.I.D. headlights flashing high beams
are the ticket. However, they won't help when the other
guy has his head buried in the cockpit... or up his arse.


Barnyard BOb -- super swivel head

On Sat, 15 Nov 2003 20:44:32 GMT, Ron Wanttaja wrote:

OVERALL ACCIDENT RATE:

Let's look at the overall accident rate for the homebuilt fleet and the
individual types. The "Accidents" is the total number of accidents vs. the
number of that type on the Jan 2003 register, and the "Fatals" column is
the number of accidents in which the pilot was killed.

Accidents Fatals
All Homebuilts 11.1% 3.2%
Fly Baby 5.7% 1.9%
BD-5 27.2% 11.1%
RV-3 8.2% 2.9%

An email discussion with a fellow netter pointed out that these figures are
liable to misinterpretation.

The rates shown above are over the 14-year period of the data collected,
and are not *ANNUAL* rates! Divide by 14 to get the average annual
accident and fatality rate. Or, heck, just read it off this table:

Acc. Fatals
1990 1.33% 0.45%
1991 1.17% 0.38%
1992 1.36% 0.44%
1993 1.32% 0.36%
1994 1.24% 0.34%
1995 1.29% 0.43%
1996 1.14% 0.32%
1997 0.99% 0.34%
1998 1.19% 0.34%
1999 1.09% 0.30%
2000 1.14% 0.26%

Ron Wanttaja

On Sun, 16 Nov 2003 18:28:58 +0000, Ron Wanttaja wrote:
...
The rates shown above are over the 14-year period of the data collected,
and are not *ANNUAL* rates! Divide by 14 to get the average annual
accident and fatality rate. Or, heck, just read it off this table:
...

Thanks, Ron!

You used the number of registered aircrafts as of January 2003, right? How
many aircrafts do you think have been removed from the register during
these 14 years? If this is a substantial number, they would push the
accident rate further down.

- Holger

On Sun, 16 Nov 2003 11:16:39 -0800, "Holger Stephan"
wrote:

On Sun, 16 Nov 2003 18:28:58 +0000, Ron Wanttaja wrote:
...
The rates shown above are over the 14-year period of the data collected,
and are not *ANNUAL* rates! Divide by 14 to get the average annual
accident and fatality rate. Or, heck, just read it off this table:

Thanks, Ron!

You used the number of registered aircrafts as of January 2003, right? How
many aircrafts do you think have been removed from the register during
these 14 years? If this is a substantial number, they would push the
accident rate further down.

That was so for the 14-year rate in the first posting, but for the
annual-rate table, I used the number of homebuilts that the EAA said was on
the registry for that year. So the per-year data should be fairly
accurate.

I got that data several years ago, when I was researching my homebuilt
registration-statistics articles for KITPLANES. Here's the breakdown of
the per-year registration, with the net gain over the previous year.
"Total" is the total number of homebuilts on the registry that year, "Net"
is the net percentage increase in the number of homebuilts over the
previous year.

Year Total Net
1990 13432 6.1%
1991 14227 5.9%
1992 14916 4.8%
1993 15068 1.0%
1994 15995 6.2%
1995 16876 5.5%
1996 17837 5.7%
1997 18732 5.0%
1998 19643 4.9%
1999 20494 4.3%
2000 21087 2.9%

While the figures through 2000 were declining, they came up in subsequent
years:

2001 22186 5.2%
2002 24496 10.4%
2003 25656 4.7%

5.2% was the average annual net increase between 1990 and 2003. I took a
cut at adding the accident aircraft back into the totals to get an
approximation of the number of new homebuilts registered that year (the net
value above stems from the new-registered homebuilts minus those removed
from the registry in a given year). However, the FAA doesn't automatically
remove crashed planes from the registry, so the approximation wouldn't
work.

Ron Wanttaja

On Sun, 16 Nov 2003 21:00:00 +0000, Ron Wanttaja wrote:
...
That was so for the 14-year rate in the first posting, but for the
annual-rate table, I used the number of homebuilts that the EAA said was
on the registry for that year. So the per-year data should be fairly
accurate.

Ah - then it is good, thanks Ron. Of course, for us as individuals, the
probability of driving a vehicle that will kill us may be more important
than the probability of this happening during a certain year within the 14
year period.

- Holger