The risky moments: when decisions go bad

Everyone who writes about aviation safety eventually comes around to the subject of risk management. The FAA wants CFIs to teach it using checklists, which is hardly realistic. There is simply not a formulaic answer to any question here.

The simple truth is that risk management can be done only through a deal the pilot makes with self. Just like you can tell a kid a thousand times not to venture out on an apparently frozen pond, you can tell a pilot a thousand times not to do this, that or the other hazardous thing, but in the end each individual has to come to a conclusion about what is risky, what is not, and what level of risk is acceptable.

Risk assessment tool

Managing risks is not as simple as a checklist.

Because bad decisions lead to accidents, we can speculate that a pragmatic pilot might be the best at dealing with risks. A good understanding of the risks and of the fact that, when improperly or recklessly done, flying can be extremely dangerous is also required. Finally, a strong sense of self-preservation is a definite risk-management asset. No checklist will ever take the place of these things.

It is a fact that the risk is lowest when flying a well-maintained simple airplane on a clear and calm day. Beyond that, the risk increases. I would hasten to add that, given our rules, airplanes, and pilots, the safety record plateau we have reached in recent years is likely to remain in place. Which is another way of saying that general aviation flying isn’t going to get any “safer.” That doesn’t mean that each individual pilot can’t improve his personal safety potential.

The best way to learn about risks without getting hurt is to look at the mistakes other pilots have made and learn from them. These are chronicled in the accident reports issued by the NTSB. In each case as well in a couple of generalized areas I will try to identify the point where the pilot went too far and even fancy footwork and a burst of brilliance would not have saved the day.

The pilot of a Mooney stalled and spun in, apparently while making a steep turn to try to patch up an overshoot of the turn to final. This happens and is easily addressed by never exceeding 30 degrees of bank below 2,000 feet. When the decision is made to “bend” an airplane around at low altitude it is likely to be bent, literally. The moment the pilot decides to try to salvage a bad approach is when risk peaks.

A Cirrus SR-22 pilot was flying an ILS at night in IMC and having some problems with aircraft control. The first approach attempt was botched badly enough that the pilot requested another try. After the request, the airplane entered a steep turn, the pilot transmitted “I need some help” and control was apparently lost for the airplane was completely out of control when it hit the ground.

You might say that the bad decision was made when the pilot opted to take off on a flight that he was ill-prepared to handle. I think, though, that if the pilot had asked for a time out, with a heading and altitude to fly on autopilot for a minute or two, while he got his bearings, the outcome might have been better. The risky moment might have passed. The accident reports contain a number of cases where a pilot persisted after an unsuccessful instrument approach and crashed on a subsequent approach. That can also be true in visual flying.

The pilot of a Cessna 172 did one go-around and crashed on a second approach to his home airport in good weather. It was an easy approach in a docile airplane but this pilot’s mind may have been elsewhere. The pilot had recently gotten a divorce, his business had failed and he was facing arrest on a felony charge. Just going flying with all those troubles in mind was the pilot’s big mistake.

Risk dice

Flying is a roll of the dice, at some basic level.

I came across other examples of a second try being a fatal mistake.

One involved a Cessna 501 Citation flown by a type-rated private pilot. I was not aware that any training institution would rate a private pilot in a jet but apparently one did. If the pilot had insurance, which some don’t on older turbine airplanes, I don’t understand the thinking of the underwriter, either.

The weather was perfect and this pilot was headed for a 5,001 foot long runway with which he was not familiar. The airport was in a mountainous area. The surface winds were light. There were four passengers on board.

The airplane was too high on the first visual approach to the runway. The visual illusions in rough terrain can lead to such if the pilot is inexperienced in this area.

When the pilot decided to try again, the risk went out of sight as the pilot apparently tried to “make” it work. The next approach was steep and the nosewheel contacted the surface first, about half way down the runway. The main gear then touched, the airplane bounced, the sound of engine noise increased, the right wing contacted the runway and the airplane flipped and burned off the right side of the runway. For some reason, thrust reverse had been selected for the right engine only and this would have complicated an already uncontrolled touchdown. There was no established mechanical malfunction.

As kids we learned that if at first you don’t succeed, try and try again. The problem with this in airplanes is that the risk increases greatly when you do that. Maybe the saying should be modified to suggest doing something more easily managed than to keep trying that which bested you the first time.

Many folks say that all the high-tech equipment that is now available makes flying safer. I am not saying that it increases risk though that is possible as was illustrated by a pilot flying an A36 Bonanza.

This relatively low time pilot (343 hours) with limited instrument flying experience (32 hours in two years) was on a long cross-country flight. His Bonanza had both Nexrad and lightning-detection equipment. Using this, and talking with controllers about the weather, the pilot decided to fly through an active line of thunderstorms.

Nearing the weather, the pilot reported that a cell had filled in but there was still a gap in the line about 10 miles north. The pilot attempted to fly through that gap and nothing further was heard from him.

The airplane broke up in flight, with the wreckage spread over a mile on the ground. An overlay of the Bonanza’s track on a weather radar plot showed that the pilot tried to fly through a Level 5 (heavy) thunderstorm cell. That almost never works.

Radar imagery in the cockpit is fine but there is enough of a delay in what the pilot sees to make use of it quite risky in a dynamic thunderstorm situation. When a relatively inexperienced pilot makes a decision to penetrate an area of weather using this information, the point of no return can be quickly passed.

As airplanes get older, expensive maintenance becomes a larger part of the risk management challenge. Yes, you can get a fine and complicated twin for not much money, but can you afford to care for it properly? In recent times there has been a rash of accidents in older 400-series Cessna twins to illustrate this. A recent look at Trade-A-Plane showed a number of 421s available for under $200,000 with almost as many (39) for sale as 310s (46). A 421 would today sell new for well north of $1.5 million so the temptation to buy that much luxury for under $200,000 would be strong. It would be a great family airplane, too, which was the case in the accident I am about to tell you about.

Cessna 421

Those older piston twins offer a lot of bang for the buck, but they also demand serious maintenance.

The 33-year old 421 was cruising at Flight Level 210 in good weather when the pilot declared an emergency. He had a rough-running right engine and was diverting to a suitable airport that was only 10 miles away. The pilot subsequently reported that he had shut down the right engine.

The airplane was descending out of 17,000 feet as it arrived over the divert airfield where the pilot circled as he descended. Everything appeared to be unfolding smoothly as the pilot entered a downwind leg. The airplane was about 2,200 feet AGL when abeam the runway threshold.

Then the pilot started giving altitude away and was down to 600 feet when three miles from the approach end of the runway. Two hundred feet per mile would be a shallow approach slope and the airplane was either on, or turning onto final when it rolled inverted, crashed and burned. The pilot, his wife, and their five children perished. The pilot not only gave away altitude, he gave away airspeed and then everything else as well.

This approach and landing should have been a piece of cake and one could only speculate that the situation might have involved less risk if both engines had failed. Then the pilot could have glided serenely in for a dead-stick landing.

As an aside, the left engine had been overhauled by a fine shop right before the pilot bought the airplane. The right engine had been overhauled by a less well-known shop well before that, long enough ago that the failure of the cam gear due to lubrication distress should not have been related to that overhaul. The 421 engines have a great need for expensive maintenance, and a lot of it.

In a far more recent event, a 32-year old Twin Commander turboprop 690 crashed on its second approach to a Tennessee airport in bad weather, according to a press report. It was also reported that there was a family of four on board. From the appearance of the wreckage, control of the airplane was obviously lost.

Now, I am going to segue into a discussion of why I think some of the notions that pilots harbor can lead to those risky moments. Those last accidents in twins illustrate something I have been discussing for years.

Statistically, twins are no safer than singles and by some measures they have proven to be less safe. Yes, if a twin pilot is sharp, he can handle an engine failure. Apparently most twin pilots don’t do well at this and your chances of being done in by an engine failure are higher in a twin than a single. It is almost as if a pilot can buy a twin and fly it safely until an engine fails. Then, boom.

There are valid reasons to buy a twin. Compared with big-engine singles, they climb better, go faster, and have more payload. I think, though, that ego satisfaction outranks those things as the reason pilots like twins. Flying and arriving in a twin is simply good for any pilot’s ego. I always loved to fly them.

G1000 MFD with weather

The latest avionics don’t make flying safer, but they are a lot of fun.

The riskiest part of twin flying comes when a pilot does something in a twin that he wouldn’t do in a single. All Part 23 propeller airplanes are created equal and viewed as equals by things like weather and flight envelopes. A pilot is operating in a high-risk zone if this is not acknowledged.

Any other thought about “buying” safety is a risky one. The Cirrus experience clearly demonstrates this. No manufacturer has been as aggressive in offering the latest high-tech equipment and the Cirrus is the only certified airplane you can get with an airframe parachute as standard. Yet the Cirrus fatal accident rate is no better than that of like airplanes even though the much younger Cirrus fleet flies with all the goodies. That is no reflection on the airplane but it does say a lot about the risks that pilots take in the airplane.

I always put the latest high-tech gadgets in my airplane, autopilot excepted. I liked the simple ones of those best. But despite all the good things, including vertical profile radar, weather in the cockpit, ground prox, and traffic information, I never thought there was a safety advantage over the basic nav/com equipment it was originally equipped with except in one area. Traffic. I did value that. I did, however, enjoy the hell out of flying with the rest of the good stuff.

An extremely risky moment comes when a pilot flies with this thought: “I think I can make it.” Overconfidence is not good but a basic high level of confidence is required of a pilot. We need to know what we are doing and do it well. If there is any doubt, don’t.

The final risky moment I’ll share with you applies to motorcycles, hot cars and pickup trucks as well as airplanes. It comes when a person thinks or says, “watch this.”


  1. Louis Sell says:

    Thank you again Richard for a wonderful article that talks about practical things.
    I’ve read many stories here about all the wiz bang glass boxes and wondered how many really knew that the weather display they were using may be 20 minutes old. In fact, not long ago there was an accident where a pilot flew into bad weather while using what turned out to be 20 minutes old information.
    I also admit that after flying a long cross country, in poor weather, my instrument approach was dismal. I called center and ‘fessed up and ask to go missed and fly in a hold until I could get myself back together. It took me about 15 minutes to get everything screwed back together and then I made a very good approach. How good? Well 200 and 1/2 was really tight. Fortunately, I had business there because it was no better than 100 and 1/8 for two days in fog.

  2. Linda BR says:

    Mr. Collins, once again you are right on target. I would only add that pilots also get themselves into trouble when they get away with a close call and fail to put that incident onto their I’ll-never-do-that-again list. In my humble opinion, many accidents are caused by pilots who are known around the field as “accidents waiting to happen” or by those who use their close calls as hall passes to go out and push their risk beyond their abilities (or that of their equipment.)

  3. Steve Phoenix says:

    Not much to argue with here. I continue to take issue however, with the aviation magazine editors insistance that publishing and dissecting accident reports in every issue “so that we can learn from them” is the correct thing to do. It is very peculiar to the aviation business. This feature is not seen in motorcycle, car or boat magazines; all activities with the same fatal mistakes repeated over and over. As Richard has noted, aviation safety has reached a plateau from which it is unlikely to move. That plateau, by the way, is about even with rat racing around on a Ninja motorcycle. A safety level publicly acceptable for solo riders, but probably not for a family of five. In any case, reaching a safety plateau, just like reaching a learning plateau when learning to fly, may be the juncture at which a light comes on and a sudden insight takes place. What is that insight? I don’t know, but I’m pretty sure it’s not dissecting the same accidents over and over. Maybe it’s like, ah hah, all private pilots should be limited to well-maintained, simple 2-place airplanes on clear and calm days.

    • Duane says:


      We don’t really know that “aviation safety has arrived at a plateau from which it is unlikely to move”. We do know that we’ve reduced the GA fatal accident rate dramatically from what it was a couple generations ago (from near 10 per 100,000 flight hours to near 1 the last couple of years). The rate obviously can’t get to zero due to human fallibility. But there are many things that could still significantly cut the fatal rate further.

      Given that many if not most of today’s pilots tend to be old you-know-whats, who tend to be least somewhat resistant to changing habits or technology, it’s also not surprising that we’ve at least temporarily plateaued. However, if we can entire a new generation of young pilots raised to live 24/7/365 with modern technology, I would expect the accident rates to benefit over the next 20 years.

      We do know that new aviation technology – specifically GPS navigators, cockpit weather displays, as well as a few other boxes – have likely had at least something to do with the big reduction in GA fatals we’ve already experienced, even if the rates have plateaued the last couple of years. Yet most of our GA aircraft today are legacy birds that are often too expensive to retrofit with the latest technology – thanks largely to FAA over-regulation of cockpit technology.

      Thankfully, the newly-mandated simplification of FAA Part 23 aircraft certification should make it a lot easier and cheaper to retrofit our old legacy birds with the latest in safety gear, and make new birds cheaper to buy as well.

      For another matter, we may well have reached the upper limits of current pilot training … but that only means we may need to radically change how we teach people to fly. Aviation groups such as AOPA and NBAA are already working on revised training standards. Cheaper and more capable flight simulators would help keep us current, and again, younger pilots accustomed to playing video games are more likely to take advantage of simulator training than are many of the old codgers who now make up such a large proportion of our pilot population.

      I can think of at least one instance of “new technology meets new training methods” that could significantly improve aviation safety:

      In the old days where almost nobody had AOA readouts in GA aircraft, we taught airspeed control as the principle means of operating aircraft to avoid stalls and spins. It’s a true statement – if you always control your airspeed for every possible combination of bank angle, gross weight, wind shear effects, and basically every single corner of the flight envelope, great, you’ll never stall and spin into the ground. Yet despite that fact, stall spin accidents still make up about a quarter of all GA fatals.

      So what if we follow Icon’s lead and install cheap retrofit AOA indicators in our aircraft that read out as a cross-section of a wing with respect to green, yellow, and red arcs in every single airplane … that being an extremely intuitive control input display … and then train (or retrain) all of our pilots to fly exclusively on AOA, regardless of the flight regime, and forget trying to (ineffectively) rely on airspeed control? I think there is a very good chance that we could cut way back on stall-spin accidents over the time.

      We’ll never get to zero fatal accidents, of course .. but I could see us getting down to less than 0.5 fatals per 100,000 hours, thereby perhaps approaching the safety record on our highways (which also has improved dramatically in recent years due mostly to improvements in automobile technology).

      The human noggin, unfortunately, does not seem to evolve as fast as our technology does.

      In the meantime, I and likely most pilots are quite happy to learn from the mistakes of others in the comfort and safety of our homes. Keep the aviation accident analyses coming, please.

      • Louis Sell says:

        I would like to believe accident rates can be further reduced and agree all the new electronic devices can be used to advantage.
        However, the pilot population is riddled with people of great egos. People who formed very successful businesses out of nothing. People who are fed daily how smart they are and worshiped by their patients and staff.
        Many of these people can afford expensive aircraft with all the bells and they just can’t comprehend the need for continued training and they go out into conditions over their head.
        Some of us know some of these people. Some just hear about their mistakes, Kennedy (?). A doctor who provided recurrent instrument training and yet when he flew, with his family, into real instrument conditions crashed.
        Frankly, as our airplanes age, we construct more kit units which can be abused and incorrectly used, I am not confident our safety percentage will not become worse.

        • John Zimmerman says:

          As people fly less, they are naturally less proficient. Given the state of the overall economy for the past 5 years, and the lack of life in general aviation in particular, I think this may be another factor in our plateau. There’s simply no substitute for getting out there and flying.

          • Louis Sell says:

            Re: your last sentence–show to Hal–this needs to be printed and given to each student and posted on the wall of every FBO !!!!

  4. Cary Alburn says:

    Not sure I’d agree that the risk of GA flying is equivalent to rat-racing a crotch rocket. In 2012, there were 432 GA deaths in 271 fatal GA accidents–very large to the families who suffered, but pretty small compared to the number of GA aircraft in the US (some 220,000). The overall accident rate was less than 7 per 100,000 flight hours. All of this comes from an August 2013 NTSB press release. I doubt that there are similar statistics available for rat-racing crotch rocket riders, but given the choice, I think I’ll stick with my airplane. :)

    But assuming a plateau for overall accidents and fatal accidents, individually each of us can learn from the mistakes of others. I think that’s what Dick is saying. None of us will ever live long enough or fly enough hours to make all of the possible mistakes, and I’d just as soon learn from others’ mistakes than try to make them all on my own.

    We fly in a very dynamic environment which makes all of us students–we can’t possibly know all there is to know about every situation that we might find ourselves in. But we can apply what we’ve learned from others’ errors, and perhaps minimize our own.


  5. Louis Sell says:

    While it is sometimes troubling to read accident reports, I think it serves the purpose intended. That purpose is to get people thinking about similar or somewhat similar errors they made in the past.
    I am not afraid to admit to my past errors in judgment and I did learn from them.
    I’ve also learned from listing to and reading others admissions and those from instructors.
    In my opinion, this on line magazine is a learning tool for novices as well as experienced pilots.

    • Steve Phoenix says:

      Actually, I wasn’t pointing to this magazine in particular, but most of them in general. Think FLYING, SPORT AVIATION, AOPA, etc. They all carry at least one accident analysis each month.

      My point is, I can open up any monthly motorcycle magazine and see nothing but the keeness of all the new bikes and the fun of riding. There are no turn offs. There are no accident analyses of riders that go through a corner too fast and hit a tree. It happens all the time, but they apparently don’t see the need to repeat the lesson. Everbody knows that if you slide off a tree lined corner it’s going to hurt so why repeat it?

      But maybe you have answered my question. The aviation magazine editors include an accident analysis each month because it sells magazines. For some perverse reason the pilots like it (I,m not so sure that non-pilots are quite so enthralled if they happen to pick up the rag and start reading). My next question would be, how many accident reports of VFR pilots flying into IFR weather, IFR pilots losing control, tackling thunderstorms, etc do you need before you get it? My theory is that we got all that information in basic training, but are choosing to ignore it due to external pressures we have not been taught to resist. I know that the new training methods are supposed to address that, but my own attempts at teaching these things has not been entirely satisfactory. I find it impossible to simulate that a person’s business is going to fail unless he makes this next trip to get a client or that the kid’s play is on Monday and must be at school. All the simulations I have tried like that have failed because the trainee easily recognizes that it a contrived situation and is able to respond with the correct procedure exactly because there is no pressure; no fear. In conclusion, I think we’re barking up the wrong tree and Richard’s conclusion that safety has plateaued tends to support that (and I trust his statistics more than most). I think there’s got to be a better way that not only improves the safety situation but also does not turn off prospective pilots.

      • Liad B. says:

        I respectfully disagree.

        I don’t read the NTSB reports because they are fun, I reads them because they teach me something new every time.

        The main difference between airplanes and motorcycles boils down to the party at fault; Even if you are the best biker in the world, you are still in the hands of the drivers around you. I have some scars that can testify to that.

        In an airplane it’s the exact opposite, you and you alone makes the decisions, and if you make the wrong ones you become your worst enemy. Attention to details and learning from others is the only way to beat those internal demons that make us do stupid things in the air, so, bring it on, I will take all the accident reports I can get.

        My 2 cents.

  6. John Sheehan says:


    Excellent insights, as always; they are thought-provoking and instructive. How do we get these ideas out to more people?

    However, I believe that six accidents and one+ fatalities per 100,000 hrs is not an irreducible minimum statistic. Better instruction, recurrent training and emphasis on risk assessment techniques may be the answer. Or, maybe just retorquing the nut holding the yoke?

  7. Robert says:

    Good article – the AAIB database is invaluable in helping investigate scenarios which need risk management.

    Modern avionics need up dated training programmes which are properly integrated into basic and IR training. Some of the cases cited should have had a better outcome if the pilot had been trained to manage the aircraft’s up to date flight management systems.

    Although most of my time has been on multi engine, I agree that a well maintained recently manufactured SEP is a priori a safer proposition than an ageing MEP, especially one with many expensive systems like a 421.

    A relatively new T206H is likely to cover 80% of a 421 missions, with excellent modern avionics, simple systems and in the event of an accident in controlled conditions, around a seventh, if not less, kinetic energy to dissipate of that of a cabin class pressurised piston twin.