Accident Rates for Common GA Aircraft

If you are a pilot who is a risk taker (Like that guy who flew two feet from the ground in a Piper) and dies because you hit something, or you are the pilot who likes to push the fuel envelope because you don't want to make an extra fuel stop and dies, or if you are the pilot who likes to defer maintenance as long as you can or if you are the pilot who doesn't have a healthy respect for mother nature. The airplane model has NOTHING to do with how safe the airplane is.
 
ya but....that's a fast slippery aircraft....with no chute to compensate for lack of pilot skill. o_O

btw....I bet the V-tail pilots have a better accident rate vs. the A36 types. :D
Looking it over the one that stands out for me is the Bonanza fatality rate. Wowzers. That's an airplane that usually isn't flown by low timers, and yeah, probably many do more IMC than some of the others, but holy cow man.

That fatality rate is pretty high.
......
But... that Bonanza rate bugs me a bit.
.....

What the heck are you Bo drivers doing?! ;)
 
Looking it over the one that stands out for me is the Bonanza fatality rate. Wowzers. That's an airplane that usually isn't flown by low timers, and yeah, probably many do more IMC than some of the others, but holy cow man.

That fatality rate is pretty high.

Some of that I can see where it comes from -- there's a pretty tight correlation with landing speed and fatality rate on your chart.

Hit things fast, you die. But everyone always says they want no-flap landings in everything because they "feel like they have more control in gusts" and what-not. (Another topic, we've covered that one before.)

But... that Bonanza rate bugs me a bit.

Especially considering they're another group that has upped their game on training and proficiency quite a bit through type organizations since my early days of flying.

And this is all 2000's era data.

What the heck are you Bo drivers doing?! ;)

I believe that those numbers are misleading too. (Off the top of my head) There are over 30 different types of Bonanza's out there. Notice how he categorized most of them by model type which is 1 or 2.
 
J-3 is also an older design whose fleet size may have a lot of inactive aircraft.

I'll bet the J3 has a lower usage rate. A Cub won't register in the stats if it never leaves the hangar.

Another WAT[1] is that J-3 pilots may be a more skilled pilot subset. Must be good at flying a tail dragger, used to low/slow/seat of the pants flying, many used for backcountry and/or short field ops., etc. Either you get real good at it or get real dead. Looks like they're good as a whole.

Interesting that the Mooney is well below average in accident rate, but when Mooney drivers do crash, that fatality rate is higher than the mean.

[1] Wild assed theory
 
Looking it over the one that stands out for me is the Bonanza fatality rate. Wowzers. That's an airplane that usually isn't flown by low timers, and yeah, probably many do more IMC than some of the others, but holy cow man.

That fatality rate is pretty high.

Some of that I can see where it comes from -- there's a pretty tight correlation with landing speed and fatality rate on your chart.

Hit things fast, you die. But everyone always says they want no-flap landings in everything because they "feel like they have more control in gusts" and what-not. (Another topic, we've covered that one before.)

But... that Bonanza rate bugs me a bit.

Especially considering they're another group that has upped their game on training and proficiency quite a bit through type organizations since my early days of flying.

And this is all 2000's era data.

What the heck are you Bo drivers doing?! ;)
But if I'm reading the numbers correctly, the Bo accident rate isn't higher, the fatality rate is. This would leave me to believe that the pilots aren't (particularly) screwing up, but the high performance of the Bo is killing people. Not sure you could change that.

And (again) if I'm reading things correctly...what is up with the 172 R&S? Holy cow!

Oh, and good work @wanttaja ! Looks like quite a time consuming project.
 
Another WAT[1] is that J-3 pilots may be a more skilled pilot subset. Must be good at flying a tail dragger, used to low/slow/seat of the pants flying, many used for backcountry and/or short field ops., etc. Either you get real good at it or get real dead. Looks like they're good as a whole.

Interesting that the Mooney is well below average in accident rate, but when Mooney drivers do crash, that fatality rate is higher than the mean.

[1] Wild assed theory

The reason for this is Mooneys are primarily long distance traveling machines. As are a lot of Bonanzas. So a higher percentage of accidents are weather and IMC related. These are fatal at a way higher rate. Which explains why the fatal rate is higher...


Sent from my iPhone using Tapatalk
 
I would be more interested in seeing accident rates by engine type. Very few crashes are a result of air frame failure so what brand they happen in is of little importance to me. fatality rate however is an interesting statistic for comparison between manufacturers.
 
What is consistent is that the fatality rate per accident is higher for faster airplanes. Makes sense.
 
What is consistent is that the fatality rate per accident is higher for faster airplanes. Makes sense.

That, and mission, as someone else stated. Mooney, Bo, Comanche, Lance/Toga, Cirrus, these are going places airplanes.
 
That, and mission, as someone else stated. Mooney, Bo, Comanche, Lance/Toga, Cirrus, these are going places airplanes.

And what's interesting with that is that imho weather is the most preventable by the pilot kind of accident. Either by making a no go decision, or steering well clear, or to an extent with equipment like making sure you have an AI backup, stormscope, nexrad etc...


Sent from my iPhone using Tapatalk
 
But if I'm reading the numbers correctly, the Bo accident rate isn't higher, the fatality rate is. This would leave me to believe that the pilots aren't (particularly) screwing up, but the high performance of the Bo is killing people. Not sure you could change that.

And (again) if I'm reading things correctly...what is up with the 172 R&S? Holy cow!
My apologies...I should have made clearer the reason for separating the 172 R&S and the 182 S&T.

I compute a Fleet Accident rate; the average number of accidents per year for a given aircraft. I prefer this as being something that be computed directly from available records. Most analyses compute it on the basis of flight hours. Both approaches use the fleet size as part of their process, the problem is, the flight-hour method must also use a prediction of how many of the aircraft are active, and their hours per year. IMHO, the FAA's active-aircraft determination process is flawed.

So...since I don't try to weed them out, the fleet size value has some unknown percentage of inactive aircraft in it. As long as I'm comparing aircraft of similar vintage, this is not a big deal...both probably have the same inactive percentages. If they're of similar missions, the flight hours should be able the same, too.

This does get to be a problem when computing the fleet rate for an aircraft of more-recent introduction. The Bonanza, for instance, has a much longer "tail" than the Cirrus... it's been around sixty years, vs. just 18 for the Cirrus. There are going to be a lot more inactive Bonanzas that will make its Fleet Accident Rate artificially lower.

The Cessna 172 and 182 presented a unique opportunity. Cessna shut down the production lines in the 80s, re-starting them in the '90s. And they restarted them with NEW MODEL NUMBERS. So it's possible to eliminate the long tail of inactive aircraft by analyzing exclusively those models that first appeared in the '90s. Since they came out almost the same time as the Cirrus, it allows a more-direct, more accurate comparison.

So the resurgent 172 has an accident rate nearly twice the Cirrus. Why? Just guessing, I suspect it's due to being flown as a trainer, and is a common mount for low-experience pilots. I analyzed 172 accidents as part of my homebuilt accident analysis. Homebuilts see about 38% of their accidents due to pilot miscontrol, while a whopping 60% of 172 accidents have this for a cause. Median flight hours for the homebuilts was about 1,000 hours, while the Cessna group was 460, less than half that. The Cirrus training programs might be helping, too...I've got a year-by-year rate summary for the Cirrus I may post in a bit.

As far as the fatality rates are concerned, my homebuilt analysis gets me to speculate that the issues are three-fold. First, simple physics says if a plane hits the ground at a higher speed, the occupants will be subjected to higher impact forces. MV^2 is a *****.

Second, as others have pointed out, airplanes that are used for transportation are more likely to experience weather conditions. Continued flight VFR into IMC conditions is a relatively low percentage of the accidents, but a very high mortality rate.

The last...well, my data seems to indicate that wing position is a major factor. A high-wing aircraft puts major structure around the occupants' heads. I suspect this is protecting the occupants better.

Ron Wanttaja
 
Just from the glance on Nall report: 64-74% of the acidents are pilot-related. That by itself already gives a bit of a food for thought.
Well, when they can't find another cause, PE is the default, and is almost always a "contributing factor".

Ex:
"Cause: the departure of both wings during level flight in VMC conditions. A contributing factor was the pilots failure to maintain altitude."

Yes, I'm a NY cynic.
 
Well, when they can't find another cause, PE is the default, and is almost always a "contributing factor".

Ex:
"Cause: the departure of both wings during level flight in VMC conditions. A contributing factor was the pilots failure to maintain altitude."

Yes, I'm a NY cynic.
True though...I've seen many a report that has statements like "a contributing factor was the pilot's failure to control the burning aircraft"...
 
...well, my data seems to indicate that wing position is a major factor. A high-wing aircraft puts major structure around the occupants' heads. I suspect this is protecting the occupants better...
Hmm, hadn't noticed that but now I see...guess I'll have to take solace in that my plane, which is a low wing making it more dangerous, is therefore also "cooler" (because we all know dangerous things like motorcycles and smoking are cool).
 
If you're computing accidents per hour, then you'd expect the Mooney to be higher since it's faster, and goes further per unit time.

But maybe it's counterbalanced by Mooney pilots flying the same distances in shorter time frames ;)


Sent from my iPhone using Tapatalk

I'd say on average that the Cirrus models are faster as a total fleet. Most Mooneys flying today are the 140-160 knot C-J models.
 
As a DA40 owner, I was interested in your report of Diamond's accident and fatality rate. I looked through the NTSB database and found 7 of the accidents for that period were foreign registered, all but one of them fatal. Since you are using the number of U.S. registered aircraft to compute these rates, then I would think that dividend should be the number of U.S. registered aircraft that have had accidents. If so, the U.S. registered accident rate would be .55% (52/7/1349) and the fatal percentage would be 17.3% (9/52). Also, with a rate for all Diamonds you are comparing a motor glider, a two-seat trainer, a four-seat trainer/cross-country aircraft and a twin trainer/cross-country aircraft.
 
I'd say on average that the Cirrus models are faster as a total fleet. Most Mooneys flying today are the 140-160 knot C-J models.

I'd not be so quick to be sure. We have SR20s in the mix too. And plenty of M20J, K, M, R and TNs have been built... A J is faster than 160... there are also missiles and rockets and turbos installed in earlier airplanes.

I wonder how easy it would be to use a snapshot of say flightaware data to get a sample of type and cruise speeds...




Sent from my iPhone using Tapatalk
 
chute saves
Would be interesting to see the chute pull stats from Cirrus, and the "hours flown" would be telling. Cirrus may share chute pull stats, but I don't see anyway you could get hours flown, unless someone like Foreflight was willing to let you API plugin to their database from the logbooks and from there, even though it's not the whole population, you could draw some conclusions within some level of confidence
 
I'd not be so quick to be sure. We have SR20s in the mix too. And plenty of M20J, K, M, R and TNs have been built... A J is faster than 160... there are also missiles and rockets and turbos installed in earlier airplanes.

I wonder how easy it would be to use a snapshot of say flightaware data to get a sample of type and cruise speeds...




Sent from my iPhone using Tapatalk

Flown in several Js. All settled in around 155-160 knots. I know the book number is a touch faster.
 
Flown in several Js. All settled in around 155-160 knots. I know the book number is a touch faster.

The ones I've been in flight plan 165 at a non economy power setting...


Sent from my iPhone using Tapatalk
 
I'd say on average that the Cirrus models are faster as a total fleet.
Would also depend on how it is measured... hours flown vs numbers flying. There are a couple schools out there using SR20s as trainers... they don't make the best trainer (my personal opinion) but that would be something to note. Most of the Cirrus accidents I've read about (at least what has been publicized) have been 22s
 
Would also depend on how it is measured... hours flown vs numbers flying. There are a couple schools out there using SR20s as trainers... they don't make the best trainer (my personal opinion) but that would be something to note. Most of the Cirrus accidents I've read about (at least what has been publicized) have been 22s

And very few Mooneys are used for other than traveling - it's pretty rare to attempt PPL training in one :-o


Sent from my iPhone using Tapatalk
 
And very few Mooneys are used for other than traveling - it's pretty rare to attempt PPL training in one :-o
Yes and I personally have never seen a Mooney to rent
 
So....Mooney's are awesome. Anything else you fly?o_O

Hey you read the stats too... trying to be objective and factual at all times... except when expressing preferences. I've flown Cessna, Piper, Cirrus, Mooney and a Bo, oh and a Schweitzer 300... and a couple light sports. Owned Piper and Mooney.

If you haven't flown in a modern Mooney, come by ill take ya up!


Sent from my iPhone using Tapatalk
 
Hey you read the stats too... trying to be objective and factual at all times... except when expressing preferences. I've flown Cessna, Piper, Cirrus, Mooney and a Bo, oh and a Schweitzer 300... and a couple light sports. Owned Piper and Mooney.

If you haven't flown in a modern Mooney, come by ill take ya up!


Sent from my iPhone using Tapatalk
No problemo.....I probably wouldn't fit. :lol:
 
As a DA40 owner, I was interested in your report of Diamond's accident and fatality rate. I looked through the NTSB database and found 7 of the accidents for that period were foreign registered, all but one of them fatal. Since you are using the number of U.S. registered aircraft to compute these rates, then I would think that dividend should be the number of U.S. registered aircraft that have had accidents.

You're right. Other makes have international crashes as well, but as a foreign-manufactured aircraft, the Diamond may be more significantly impacted. I'll re-run the analysis for all aircraft using only US-registered aircraft. When done, I'll update the original graphic and re-post the image.

Of the seven foreign accidents you mention, two are N-Numbered....

Also, with a rate for all Diamonds you are comparing a motor glider, a two-seat trainer, a four-seat trainer/cross-country aircraft and a twin trainer/cross-country aircraft.
The problem is, with the elimination of the non-US aircraft, the Diamond will right at my threshold for performing the analyses at all. I generally don't post analyses if there are less than 50 accidents; a change here or there can make quite a bit of difference. Looking at just DA-40s, for instance, we're at the point where a single accident more or less can have a significant impact on the results.

Might increase the number of years being sampled from. Have to take a look at some fleet sizes for Cirrus and Diamond. Cirrus had 2771 aircraft at the end of 2006, so I might right the analysis over ten years instead of seven. No one should have a hissy, then if the total number of accidents increases on the table....

Have to wait a couple of hours. My wife says I have to vacuum.

Ron "Retirement is hell" Wanttaja
 
As far as the fatality rates are concerned, my homebuilt analysis gets me to speculate that the issues are three-fold. First, simple physics says if a plane hits the ground at a higher speed, the occupants will be subjected to higher impact forces. MV^2 is a *****.

Second, as others have pointed out, airplanes that are used for transportation are more likely to experience weather conditions. Continued flight VFR into IMC conditions is a relatively low percentage of the accidents, but a very high mortality rate.

The last...well, my data seems to indicate that wing position is a major factor. A high-wing aircraft puts major structure around the occupants' heads. I suspect this is protecting the occupants better.

I agree with the first two, I wonder if the wing location isn't so much a structure thing as making it a BIT harder to flip the high wing all the way over while still moving at high rates of speed. Most high wing accidents I've seen where a wingtip is dragged, a large number of them skid along and plop back down on the wheels. On light low wings (especially with nothing but a canopy above the shoulder line) it seems much more likely that digging a wingtip in, results in a cartwheel that flips the airplane over the nose, sometimes crushing the cabin in the final stop/fall.

Yes and I personally have never seen a Mooney to rent

We were lucky enough to have an M-20C on the line at a club I was in long ago, which is when I got all my Mooney time. It was a rugged little thrify go-places machine that wasn't too difficult for a low timer to fly, and the manual gear was about as robust as any retract out there in light aircraft.

It's a shame more of them don't end up on leasebacks, but the owners usually buy them to go places and want the flexibility to leave "right now" without the scheduling hassles of a club or school using them. With the schools constantly wrecking the 172RG fleet and rebuilding them for the Commercial retract requirements, I'm somewhat surprised more of them don't look into acquiring older Mooneys like the M-20C.

You'd probably be able to keep it in the air with a better dispatch rate (marginal, but better) than a 172RG and less students would miss flopping that big handle from the back seat to the front for "gear down". It's a visceral experience and less likely to be missed than flipping a gear handle in the panel.

They're a little doggy in climb but heck, they'll make pattern altitude each time around for laps around for Commercial students. And they'd have to plan the power off 180s a little different. They sink like a brick as soon as they get too slow.
 
@denverpilot I don't have any Mooney time but that's one of the planes I'd like first hand experience in. I find the Mooneys (like the Cirrus) to be a differentiating type plane, IE, people either love or hate them, so I'd be curious to fly it first hand. But to be honest I can't picture myself "hating" any plane, I personally like flying the Cirrus as much as I do the Skyhawk and the Archer, it just depends on the mission for that day and who's coming with. A Mooney is fast and is a retract.. my only retract time is in an Arrow, and the club's Arrow was decidedly *not* a fast plane
 
OK, the updated figures...which aren't that different from the earlier ones, but there are a few telling items. Version 2 has two basic changes: First, it only counts N-Numbered aircraft. Second, I ran the start of the analysis back by three years to gain a longer duration (one month less than ten years). So the total accidents for each type has increased.

rate3.jpg

At the bottom, I added two rows with the results of the ORIGINAL analysis, so y'all don't have to try to click back and forth.

The Fleet Accident Rates don't strongly change...but elimination of the non-US-registry aircraft did lead to a ~10% reduction in Cirrus' rate.

One of the major differences is in the fatality rate: The original had about 25% for the Diamond, which some questioned. This new version has it down to 15.7%.

The weird thing is, the nationality of the aircraft doesn't factor into the fatality rate. It's the number of fatal accidents divided by the total number of accidents; the number of US-registered aircraft isn't a factor.

However, looking at the Diamond accident results, I was struck by an interesting fact. There were seven accidents non-N-Numbered accidents in the NTSB accident database. Five were fatal....way out of proportion. I suspect the foreign fatal accidents have a better chance of ending up in the US database, so eliminating them was probably a good move.

Ron Wanttaja
 
Interesting stuff. Thanks for posting OP.
 
...Also, with a rate for all Diamonds you are comparing a motor glider, a two-seat trainer, a four-seat trainer/cross-country aircraft and a twin trainer/cross-country aircraft.
And the results for "Beech" include everything from Skippers to Musketeers to Bonanzas to Barons to King Airs.

Only 23 DA-40 accidents in my 10-year window, not enough to be accurate. It does come out to a 0.31% accident rate and a 30.4% fatality rate.

Ron Wanttaja
 
It also let Cirrus take the lead on fatal percentage, and let the Bo drop to second place.

And you'd think the 210 would match fatal rates with the Bo and Cirrus with its much higher accident rate, but it doesn't. Interesting.

I wonder if we're seeing "210 pilots don't fly enough and wreck on the runway but don't die doing it" in those numbers...? Gear ups? (I guess that goes for the Bo, too. It's a whole category of accident the Cirrus can't even have...)
 
Back
Top