Derailing Yourself

Lessons from the Amtrak 188 tragedy

Posted Mar 05, 2019

On May 12, 2015, Amtrak passenger train 188 derailed just north of Philadelphia at 9:21 pm. A 98-ton locomotive and seven 50-ton cars went off the tracks at the Frankford Junction, one of the sharpest curves on Amtrak’s Northeast Corridor. Of the 258 people on board, 8 died and 185 were taken to hospitals. It was one of Amtrak’s most serious train accidents. 

The speed limit for the Frankford Junction curve was 50 mph. Amtrak 188 had reached a speed of 106 mph when it entered that curve.

National Transportation Safety Board
Source: National Transportation Safety Board

How could this have happened? It makes no sense. So this essay is an exercise in explanation, to move the reader from an initial state of disbelief to an eventual realization that this could happen to any of us. To move you from a view that this was a bizarre event to a view that it was all too plausible.   

I will be drawing on some recent research I have done on the activity of explaining events. One of my findings was that, in order to effectively explain a surprising event to another person, I try to identify what problematic assumption that person is making, and then using this diagnosis I design my explanation to revise or replace that assumption.

In this case, I think the assumption is that the train engineer had become irrational — that he knew the curve was coming up and yet accelerated rather than slowing down.

According to the 2016 National Transportation and Safety Board report, that assumption is wrong. The 32-year-old engineer, Brandon Bostian, showed no trace of alcohol or drugs in his system when he was tested following the accident. He was not texting or using his cell phone. He was not deranged or suicidal.

The NTSB report concludes that Bostian had gotten confused about where he was: He thought he had already passed the Frankford Junction curve, and needed to accelerate to keep on schedule. That’s why he opened up the throttle, only to make the horrifying discovery that he was just entering the curve.

And now I have to add a story, a 4-part story to describe how this loss of situation awareness occurred. 

The first part of the story is that Bostian’s attention was diverted because he had heard a radio report that another train, near the Frankford Junction, had been “rocked” — that someone on an overpass had dropped a rock on the locomotive, sending glass into the face of the engineer — and the engineer had issued a warning about dangerous activity in the area, something for Bostian to worry about. Bostian was also aware that a train crew might be on the tracks inspecting potential damage on the other train that had been rocked, and so these crew members might be at risk when Amtrak 188 passed by.

The second part of the story is that Amtrak 188 had just passed a smaller right-hand curve, and there was a similar small right-hand curve after the Frankford Junction, so it was possible that Bostian, distracted by the rocking incident, confused that first curve with the second small right-hand curve right after the Frankford Junction curve. Bostian might well have believed he was past the Frankford Junction curve, had just come round that second curve, and was free to increase his speed to 110 mph.

But why would Bostian make that confusion — why couldn’t he see where he was?

The third part of the story is that the accident occurred in the evening, at 9:21 pm, so Bostian couldn’t readily observe any landmarks.

But why didn’t he just glance down at his display to see where he was on the track? And here we come to one of the most interesting parts of the story, the 4th part: There wasn’t any display.

That’s right: Amtrak doesn’t provide its engineers with a display showing their current location and the upcoming stretch of track, and the Federal Railroad Administration does not require such displays. As automobile drivers, we have GPS devices showing roads and landmarks. You would think that with a train on an unmoving track, it would be child’s play to design a GPS-based display. But Amtrak seems to have determined that such a display is not necessary, because engineers have memorized the routes they are taking. True enough, except that it doesn’t allow for engineers to become confused, to momentarily forget where they are.

I don’t know the reasoning and analyses that went into the decision not to provide engineers with real-time displays of their current and upcoming situations, but I speculate that it involves the “job as envisioned” mindset — reviewing how one expects the job to be done, and not get sidetracked with additional considerations. 

In contrast, Human Performance Specialists, including Human Factors professionals, have trained themselves into the “job as performed” mindset, imagining what can go wrong, how people can become distracted, and what resources they might need to recover and to adapt. Some designers get stuck on the concept of how a device is supposed to work, whereas others immediately begin imagining how it might malfunction. 

This “job as performed” mindset takes experience in order to draw on previous kinds of breakdowns and use them as analogs. It also takes an orientation toward speculation as opposed to calculation. It’s a mindset that might have gotten a location display installed in Amtrak 188 so that it didn’t accelerate to 106 mph going into the Frankford Junction. 

And now that we are done with the story, are you convinced that this was a plausible explanation, and that if you had been controlling Amtrak 188 it might have happened to you? Think about the times you are driving home at night, on a familiar road, and get distracted by a news item on the radio or another drive making a risky left-hand turn in front of you, and for an instant you wonder, "Where am I?" as you search for cues to re-establish your sense of location, cues to alleviate your momentary dislocation. You don't need to spend more hours memorizing your route. Instead, you need markers to help you re-orient. And Brandon Bostian didn't have those markers.

[Note: This essay is based on an article by Matthew Shaer (2016) and on the 2016 NTSB report.]


Share, M. (2016). The wreck of Amtrak 188. The New York Times Magazine January 31, 2016, 49-55.

National Transportation Safety Board (2016). Railroad Accident Report: Derailment of Amtrak Passenger Train 188, Philadelphia, PA, May 12, 2015. NTSB/RAR-16/02, PB2016-103218.