I did some data mining using the Aviation Safety Network database and various sources for production numbers to come up with an Excel spreadsheet that you can download here:

http://www.filedropper.com/twinturbopopsafety

Of course this analysis is not perfect. I look at accidents/incidents versus the number of aircraft produced as it's quite difficult to know for sure how many of a particular aircraft type are still flying.

If you look at the last 5 years of data, you are least likely to die in a:

1-Piaggio Avanti (no fatal accidents yet ?)

2-Piper Cheyenne 400LS (no fatal accidents in the past 5 years)
tied with Cessna 425 Conquest I (no fatal accidents in the past 5 years)

3-Mitsubishi MU-2 (all variants combined, 0.28% of aircraft produced involved in a fatal accident in the past 5 years with 5 fatalities)


Looking at the past 5 years of data, you are most likely to die in a:

1- King Air 100/A100/B100 (2.35% of the aircraft produced involved in fatal crashes in the past 5 years with 26 fatalities)

2- Piper Cheyenne I/II/IIXL (0.66% of the aircraft produced involved in fatal crashes in the past 5 years with 25 fatalities)


I invite people to improve upon the spreadsheet with more accurate data.

-Pascal

Seems like you'd have to have hours. KA's are, I would think, likely to be flown more hours than the others listed.

BWTHDIK.

The Avanti is too cool looking to crash.

This isn't taking into account #Flying on a daily basis.

The aircraft you listed as "safest" are rarely flying.

KA, PC12 are the most "common" TP flying.

Also true. I did not look at the Turboprop singles but I will add them.

The safety statistics of the King Air 100 series does stick out very badly. In the USA they have been mostly relegated to cheap low usage private airplanes. In Canada, they are the backbone of the low-cost-per-mile charter fleet. Despite professional two-crew operations, they have been involved in 3 fatals in Canada over the past 6 years.

I have updated the spreadsheet to include the Rockwell Twin Commander series. Their safety record is even worse than the Cheyenne I/II/II XL (27 fatalities in the past 5 years). Download the new version here:

http://www.filedropper.com/twinturbopopsafetyv2

-Pascal

Aircraft flying on Flightaware right now

PC12 - 40
KA200 - 55

Avanti - 2
Cheyenne - 1
Mitsubishi - 5
Commander - 1

Airplanes aren't very dangerous when they're sitting on the ramp..... Or in a junk yard

I've added the PC-12 and TBM 700/850/900 SETPs.

The results aren't good. The loss rate on the TBMs are below that of the King Air 100 series, but basically double or more of the twins.

The PC-12 fares worse than all of the Twins.

I also made an important correction... I am using all incidents and fatal accidents since January 1st 2009, so I am going back 6 years, not 5 years. Corrected the spreadsheet accordingly.

Latest version:

http://www.filedropper.com/twinnsingleturbopropsafety

-Pascal

Is there an option to add Merlin III data to the mix?

The safety analysis with raw numbers and no flight hours attached to them is completely worthless for any meaningful discussion. Robert E. Breiling Associates publishes an analysis each year as does the Nall Report. Each of those bases the "rate" on 100,000 hours. Where they get that, or how they derive it I don't know.

I have seen some information Breiling published regarding fatal accident rates which showed (as of February 2012) that the Fatal Accidents per 100,000 hours for the the Pilatus PC 12 was .32, Socata TBM 700/850 1.66, the Piper PA 46TP was 1.77, the overall Twin Turboprop rate was .70 and Business Jets in the U.S. was .21. The same information showed the SE Turboprop rate overall at .63 and Twin Engine TP rate at .70.

The Overall accident rates (non fatal) were really interesting with the Piper at 5.11 per 100,000 hours, the TBM at 3.93 and PC12 at .81. The Twin Turboprop overall rate was 1.98 and Bizjet rate .83. The overall SE Turboprop rate was 1.89 and The Twin Engine rate was 1.98.

So, in the same way you expressed it Pascal, you have a better chance of not killing yourself in a biz jet followed by a PC12, a King Air, TBM and Piper in that order. It doesn't look to me like there is as much difference in the chance of killing yourself between twin and single turbo props as there is between the choice of which single turbo prop.

Agree - nothing to do with safety unless you include some metric of utilization.

Interesting that the PC-12 has a lower accident rate than King Airs but a slightly higher fatality rate.

Here are some stats on SETP vs TETP:

Full report using data through 2012:
http://www.westair.com/wp-content/uploa ... Report.pdf

[Accident rate] and Fatal accident rate/100K hours
KA 200 - [.91] .28
KA 300/350 - [.31] .06 (requires type rating, basically everything else equal to KA 200)
PC12 - [.74] .30
MU2 - [3.69] 1.58 (realize this isn't fair considering new training requirements)

I find it interesting that the PC-12 has comparable accidents to the TBM, but the flight hours are much higher, giving the PC-12 a huge safety margin. Do PC-12s fly 4X as much as TBM?

Looks like PC-12 is the way to go.

Better pilot training, or two pilot crews?

What Caldwell said.
Selection bias.

Pascal,

Flying involves man, mission, and machine.

Your exercise will not measure the safety of the machine, the various models of turboprops.

If you do a REALLY good job, it will measure the safety of the people and economic circumstances of how each type is flown, that is the man and the mission. The actual machine makes almost no difference.

There is a long tradition in aviation of trying to measure the safety impact of a particular model by looking at its accident history. This exercise never tries to normalize the man and mission aspects so the analysis doesn't tell you that one airplane is safer than another.

What makes one model "dangerous" is that it is used in more dangerous ways by the self selection process of the owners.

The FAA is as guilty as anybody else is making the assumption all types are used equally. This promulgates the fallacy you can measure a type's safety by looking at its accidents.

The evidence is all there.

A Cirrus is an intrinsically safer airplane than a Corvalis, yet has a higher fatal accident rate. The Cirrus attracts higher risk missions and perhaps higher risk pilots as well.

The MU2 went from worst to first by changing the man (through training) and mission (freight uses dropped), not the machine.

The Cessna Caravan has a high accident rate because it is used for night freight. The plane is about as simple and safe as a single can be otherwise.

When you look at the accidents, you see EXTREMELY FEW machine caused accidents. The vast majority are pilot caused or exacerbated by a risky situation. Most of the accident stats come from the worst pilots in the worst situations, so it doesn't measure the majority case.

The implication is that your research will indicate which airplane you should buy to be safest. But it will do no such thing. For any given man and mission, changing the machine will make very little difference. The aircraft with the worst accident history can be safely flown by a well trained diligent pilot. The best aircraft can be flown by an incompetent pilot on super risky flights. The aircraft made essentially zero difference to the risk profile of those two flights.

What you will measure is how that particular type gets used and how risky that is, not how risky the airplane itself is. You are mostly measuring PEOPLE, not PLANES.

Mike C.

Mike,

I think you make several excellent points and in general I am in agreement.

It has been argued to me that the TBM is a less safe airplane than the Pilatus because it has a tendency to torque roll when power is applied suddenly. This has occurred a number of times apparently during take offs and during go arounds. I answer that many aircraft have a lot of rolling tendency due to relatively higher powered engines and can operate safely with proper pilot training and operation (your point).

Also, it is argued that some TP's (like Pilatus) are more frequently professionally flown and thus have a lower accident rate than TBM (again in agreement with your basic point). Additionally, it is pointed out that some TP's are flown by two crew people and thus have a lower accident factor because this mitigates against pilot error.

I don't see the TBM as less safe than the Pilatus with these things considered. It looks like a pilot training and proficiency difference to me. But I may be missing something else.

You absolutely need to look at flight hours. I'd suggest taking a look at the article (a bit dated but still relevant) in Flying magazine on the value of training and what it did to the safety record of the MU-2. Article can be found here: http://www.flyingmag.com/technique/prof ... ning-works

Yes I fly an MU-2, but this article makes the (correct in my opinion) point that the MU-2 isn't inherently safer than the other aircraft, just that the mandatory training has made it so...

Dave