Stepping down in automation—the real lesson for children of the magenta line

Warren “Van” Vanderburgh was an extraordinary pilot. Twenty seven years in the Air Force, 14 times Top Gun, and 32 years at American Airlines—the sort of guy you might want to pick up a few pointers from. In 1996, Van was tasked by American Airlines to address the number of accidents, incidents, and violations that looked to be caused by “Automation Dependency.” A term probably not ever used before. In April of 1997, Van held a class at American Airlines Training Academy in Dallas, Texas, titled “Children of the Magenta Line.” The class was videotaped and is available on YouTube. Twenty three years after it was recorded, Children of the Magenta Line is still a very valuable training session and worth reviewing regularly.

Van Vanderburgh
Van Vanderburgh was an accomplished pilot and a brilliant instructor.

What is Automation Dependency and where did this issue come from? Very likely the Boeing 757 played a major role. First put into commercial operation by Eastern Airlines in 1983, Boeing delivered 1,050 757 models between 1981 and 2004. The pilots at Eastern referred to it as the Electric Jet. It was the first commercial aircraft to have a Flight Management System and Electronic Instruments. Pilots transitioning from the DC-8, DC-9, and 727 had their hands full just getting through the training program. Many washed out and others just elected to go back to the steam gauges. A rather senior 727 captain friend of mine described his first week of training as, “I felt like a dog watching TV.”

If you managed to finally check out on the 757, the next task was to figure out how to safely fly it around the system without major incident. With two brand new pilots in the cockpit of the Electric Jet and a futuristic-looking Flight Management System (FMS), it was a job just loading the thing properly and getting underway. Once airborne, the transition from actually flying the plane to knowing which buttons to push would ultimately lead to another new term: “what’s it doing now?”

This was likely repeated several times during the flight. There always seemed to be confusion as to who was to push the buttons which frequently lead to both pilots being heads down. One pilot was trying to correct an error the other pilot made, or make an entry that the other guy couldn’t figure out how to do. It was not long before the hazards of automation started to become clear.

New procedures were developed at each airline as to how to manage the various cockpit tasks. A common procedure was for the pilot flying to make the entries and the pilot not flying to verbally verify the entries; this helped substantially but the safety record still needed some work.

Enter Van Vanderburgh and the American Airlines Training Department. Van and his group analyzed accidents, incidents, and violations and determined that 68% of them were caused by automation mismanagement. They determined that pilots flying the new automated jets were becoming “Automation Dependent Pilots.” One of Van’s slides defines such a pilot as one who does not select the proper level of automation for the task and loses situational awareness which is frequently proceeded by task saturation.

In Van’s presentation he describes three levels of automation:

  • The pilot manually flying the aircraft.
  • The pilot using the flight director, autopilot, and autopilot modes to fly the aircraft for a short period of time. For example: Heading select, Flight Level Change, Vertical Speed, Indicated Airspeed, etc.
  • The pilot using the FMS to command the autopilot to fly the aircraft for hours at a time.

So, what is the appropriate level of automation?

G1000
When in doubt, step it down a level or two in automation.

This is the meat of the subject where we as pilots today, 23 years after Van’s presentation, still struggle. The basic concept is that when things go wrong or get complicated, step down a level in automation.

Consider the following…

You are on a Vnav descent and navigating in GPS, Nav mode selected. You get a reroute; type-type-enter-select new airway, enter-then direct-enter-enter. What vertical mode would you likely be in? Probably Pitch. Not good, what to do? Step down a level in automation first by selecting Vertical Speed and maybe heading if appropriate, then enter the re-route.

What about a traffic alert that requires evasive action? Would Vertical Speed be the best choice? Heading select maybe? How about stepping down three levels to hand-flying the aircraft as necessary.

A last-minute runway change with the airport insight? How much button pushing is required? Step down three levels and just fly the plane.

And think about this: there are two pilots in that 757, but how many are in our single engine or light twin, Technically Advanced Aircraft (TAA)? We are taking on the job of two type-rated pilots, technically both captains, when flying single-pilot IFR. Is the G1000/Perspective system any less complicated than the Honeywell system in a Boeing? I flew the 757 for 18 years and I can tell you that although the aircraft systems are quite different and much more complex than our typical general aviation aircraft, the FMS entries are not that much different. Also, consider that you have no one backing you up for the entries that you make. In our planes, “garbage in” does not necessarily result in “garbage out.” More like “splatter.”

Techniques to mitigate automation dependency and the single pilot issue:

  • Slow down. As pilots become more proficient with the FMS keyboard they tend to go faster with the entries. For example, Direct-enter-enter should be direct-enter-VERIFY-enter.
  • Be your own copilot. This is a very effective technique that is hard to get experienced pilots to follow. In airline operations, the pilot not flying is required to point his finger at the autopilot/flight director status display (scoreboard) to make sure that what is displayed is what was selected, then verbally state what he sees. Do this yourself, point at the scoreboard, and state out loud what you see. You will be surprised at how many times you will catch an error. (Like being in heading for five minutes when you should have been in Nav or the vertical mode somehow ended up in Pitch).
  • In VFR weather, turn your autopilot off and hand fly the aircraft using the Flight Director and FMS inputs. This will force you to hand fly what you have entered, improving your scan, and sharpen your flying skills.

Van Vandenburgh passed on a few years ago, but his very informative instructional videos, broadcasted from the American Airlines Training Academy, live on and are still available on YouTube. They include:

17 Comments

  • Vanderburgh has been in the back of my mind since I first stepped into the flight deck of the 737NG, almost 6 years ago. The dilemma between automation dependency and “fly the thing” that most airline SOPs struggle with, is as interesting as puzzling. The funny thing is that the experienced captains of today are the real children of the magenta of Vanderbugh’s days, and the new first officers talk easily with the FMC, but are missing the fundamentals of handling in an industry that fell into the automation spell and tend to discourage the use of basic flying skills. This balance that must be found by operators is in the core of Vanderburgh’s talks, that could be easily have been presented nowadays. Yes, we are managers of a multimillionaire business unit. But we are more than just that, and this “more” may save the day. Thank you, John, for reminding us all and for the very nice input into the single pilot operations environment!

  • Outstanding article! The perfect absolute example of automation dependency is the Asiana crash at SFO, where the crew ignored glaring signs of the aircraft getting too slow on final approach. They made a mistake that a student pilot would NOT make! Sad. Let’s hope this never happens again.

  • Good article, but it perpetuates the errors it is trying to warn against. In the 7th paragraph: “the pilot manually flying the aircraft” correctly lists this as the lowest level of automation but should include the phrase “with raw data only, no flight director”. The next level up includes hand-flying with the flight director. It is only with raw data, manual flying that your basic instrument flying skills and scan are exercised. Towards the end of the article the last recommendation under “techniques to mitigate automation dependency …” is “turn your autopilot off and hand fly using the Flight Director”. This is NOT the lowest level of automation! In this mode, after verifying the settings for airspeed, altitude, heading, and flight director modes, all your scan has to do is follow the flight director.

    Pilots of Technologically Advanced Airplanes frequently hand fly (with the flight director on). Yes, they are improving their control smoothness, coordination, and understanding of the automation modes, but NOT their instrument scan and basic flying skills. If you really want to improve your skills, both the flight director and auto-throttles must be off.

    After bankruptcies and mergers, I have flown with pilots from many airlines, and corporate flight departments. They all have different cultures concerning when to turn on the autopilot. One culture was above 18,000 feet, and another was at 1000’ feet above the airport. None of the cultures or training departments encouraged raw-data hand-flying. That may slowly be changing after the Boeing 737 Max accidents, the Asiana 214 crash at SFO, and the Air France 447 stall.

    The author, John Bone, has impressive ratings. I have many of the same ratings and the same 24,000 flight hours. Even though modern aviation culture does not encourage raw-data hand-flying in jets, it is not prohibited, and I flew many take-offs and landings hand-flying with only raw-data. The Boeing 747-400, Boeing 777, Airbus A320 (take-off required the flight director), are all easy and fun to hand fly. Even the highly sophisticated Boeing 787 Dreamliner was easy to hand-fly raw-data. Those flying skills must be practiced to become easy!

    If you haven’t flown raw-data recently, think about how procedures might be different, the hazards involved, what to do in an actual emergency, and develop a briefing.

    When conditions allow, turn off the flight director and auto-throttles, fly below 18,000 feet with raw data, and you can truly master the age of automation, by restoring your pilot skills, and by not being dependent on the automation working correctly.

  • My airline defines three levels of automation: 1 – raw data, 2 – use of FD, 3 – use of FMS for LNAV/VNAV. The use of autopilot is not included as it may or may not be used for each automation level.

  • It is interesting that automation has unquestionably made airline flying safer–despite the dependency. It isn’t clear that GA flying is safer, however. The principles advocate by Captain Vanderberg are sound, but the real problem is complacency, IMHO. Look at the SFO (Asiana) and BHM (UPS) accidents–automation dependency was involved, but both were complacency.

  • I had a friend who flew for a major US airline and he told me that they were required by company regulation to be on autopilot at 500 ft AGL. He also said the PIC had the option to hand fly if they felt it was necessary for any reason. I don’t know if “improving piloting skills” was a good enough reason. I never flew a jet aircraft but I do know I lost effective piloting skills after just a few weeks out of the cockpit of the 182 I regularly flew.

  • I was on a United flight into Denver several years ago, and talked to the Captain after landing, noting it looked like we did a Falcon # approach. I asked how much it was automated including auto land. He stated that they hand flew everything below 10K, which somewhat surprised me. On the other hand, in my Archer II, with a GTN-650 and STEC autopilot, I still routinely fly manually with only minimal assistance from the automation.

  • Great article, I made the transition from steam to glass and back and back to glass. All the comments are good however flight training today is about technology not aeronautical basics. Sad.
    Years ago we said three most common phrases were,
    1What did he say?
    2Was that for us?
    3 Oh s____
    Then when glass came about they were,
    1What’s it doing?
    2Why’s it doing that?
    3How do I make it stop?

    • Respectfully, I disagree. Where I currently teach, and also elsewhere that I learned, all PPL and CPL was done in steam/analog equipped airplanes. Only the IR courses are/were utilizing glass. We still require pilots to create and fill out paper nav logs without the use EFBs.

  • I listened to Cpt Van Vanderburgh presentation a few years ago, I looked him up and he just died. Pretty heartbroken that we lost him so early.

  • Great article. Sorry to hear about Captain Van Vanderburg’s passing.

    It’s a real nit that probably only an aviation historian would care about, but the 767 was developed before the 757. Technically it all started with the 767. Having said that, the 767 and 757 FMS systems were identical. The Airbus A320 FMS system was in development at the same time and it did not have the Direct- Enter-Execute type of operation. The A320 and Sperry/Honeywell Business Jet Systems were Direct-Enter. This isn’t to take away from the article at all. It was great article and is very timely. I agree 100% with the advice. I hope to see more articles from John Bone.

  • Thanks John,
    So true. It took 6-9 months of flying the line before I was fully comfortable with 757 smoke and mirrors. I filled a yellow pad with gotcha’s – that I scribbled down during flights.

    It’s an insidious addiction, the Magenta line.

    I found as departures and arrivals got more complicated, FMC automation was almost essential. For instance, takeoffs from Cologne had the autopilot and LNAV/VNAV on by 500’ to comply with noise corridor restrictions.

    “Use it or lose it,” raw data flying is almost a memory. Stepping down the automation level can be scary for pilots already in overload and past rusty with hand flying.

    There may be more talk than walk, when it comes to keeping pilots proficient in hand flying arrivals and route changes, or actually using charts.

    I think it is more a proficiency issue than complacency. It comes back to training. Simulator profiles that lead to non-FMC modes might help. Ever get a tight turn onto the localizer ending up above the glide slope at the outer marker? A late switch to the parallel runway that has a localizer only approach? Get dispatched with no FMCs? Have the auto throttle deferred? The simulator is where long lasting lessons are learned free of charge. I’m a fan of six month training programs instead of annually.

    One last idea, besides encouraging visual approaches – – Carriers should not allow pilots to bid reserve flying more than two consecutive bid periods. It’s a recipe for double trouble when things don’t go as planned – – two pilots called in late that have not flown together – – flying to a foreign airport neither had flown into – – add a reroute to another airport, bad weather and a non precision approach into a short runway.

    Proficiency matters.

    Thanks John

  • Some of the ex-WWII bomber pilots, who were the Captains I first flew with, couldn’t do an IFR NDB let-down to save their lives, literally, but pop out of cloud too high, too fast, not configured and not lined up and say ” the runway’s over there, Sir” never forgetting the Sir ! and they would straighten up and complete an immaculate hand held approach and landing. Many years later some of my students would complete an FMS approach to minima, better than I ever could, but then be almost incapable of completing the landing manually.

  • I watched all the related instructional videos from Vanderburg’s AA presentation. He’s one of the best and his sense of humor is unparalleled! Thanks for posting this. Let us never forget this great pilot and mentor.

  • As a retired USMC F/A-18 pilot, I can honestly say that my favorite activity today is flying a 1969 T-41C w/”steam gauges” and a paper chart in my hands. There’s nothing like really flying the aircraft. How many commercial pilots actually fly the aircraft anymore? We’ve all heard the story. First they got rid of the Navigators. Then they got rid of the Flight Engineers. Then they replaced the Co-pilots with a dog. The dog’s job is to bite the pilot if he touches the flight controls while managing the FMS. 😉

  • I love to teach students the I-pad. Then teach primary along with dead reckoning as it is required for the PPL. The manual E6-B computer is as old as the hills, right! The Students love their I-pads and it is great going GPS direct and following the magenta line. They often say that sectional map and dead reckoning is old and obsolete. The I-pad is great easy and accurate. I agree, as the I-pad has saved many pilots from getting disorientated in low visibility days. Hell VOR is old stuff and I am burning gas and going GPS direct is better!

    So all I do is shut off the I-pad or the panel mounted GPS.

    Yep the pilots of today fly the magenta line and God help them if the GPS or I-pad breaks…..

  • I love these videos, and have watched and dissected them numerous times. I also study accident reports, and was dismayed to discover that these seminars were used as a scapegoat in the crash of AA 587 in November of 2001, and were cited as potential contributing factors. I disagree wholeheartedly with this assessment, as the Captain at no time reccomends using the rudders in such a fashion. He does address rudder use at one point, and states that pilots may use “some rudder, not a ton, don’t just shove it in there”. IMO they were looking for something other than the PF’s incompetence to blame the accident on.

Leave a Reply

Your email address will not be published. Required fields are marked *