Accepting the Risk of a Derailment

According to an article in The Wall Street Journal, based on documents from the National Transportation Safety Board investigation into the 2014 crude oil train derailment in Lynchburg, Virginia, CSX Corp. knew of a flaw in the section of track where the derailment occurred.
On April 29, 2014, a track inspection had revealed the flaw, and CSX decided to replace a 40-foot piece of track on May 1.  The accident, which caused an estimated $1 million in damage, occurred on Wednesday, April 30.

The track inspection indicates that CSX was monitoring the risk of a derailment, and their decision to replace the track segment (given the inspection result) shows that they were reacting to the increased risk (indicated by the precursor: the internal flaw).

Until the NTSB releases its final report, we can propose scenarios that illustrate the difficulty of risk management:  The decision to continue using that line before the replacement was done suggests that someone at CSX was willing to accept the derailment risk.  Risk mitigation has costs, and greater risk aversion costs more.  Perhaps the cost of closing that line (with the consequent disruption to shipping and revenue) for two days was too large.

Another possibility is that those who detected the flaw and scheduled the track replacement failed to communicate the increased risk to the those responsible for the operations on that track.

When New Horizons Halted

As the New Horizons spacecraft flies by Pluto today, it is collecting and sending back to Earth data from its many sensors.  But this success almost didn't happen.

On Saturday, July 11, The Washington Post had an article about the crisis that occurred just a week earlier (July 4).

The story illustrates a couple of key ideas in decision making and risk management.

First, the loss of contact occurred because the spacecraft was programmed with a contingency plan: if something goes wrong, then go to safe mode: switch to the backup computer, turn off the main computer and other instruments, start a controlled spin to make navigation easier, and start transmitting on another frequency.  A contingency plan is a great way to manage risk.

Second, fixing this situation required the New Horizons operations team to manage an "issue," not a "risk," because the problem had already occurred (it was not a potential problem).

Finally, after diagnosing the problem and re-establishing contact with the spacecraft, the team had to make a key decision: whether to stick with the backup computer or switch back to the main computer (which had become overloaded, causing the crisis).  Here, they displayed some risk aversion (not surprising considering the one-shot chance to observe Pluto): they went back to the main computer because they "trusted [it], knew its quirks, had tested it repeatedly."

Congratulations to all of the engineers, scientists, and technicians who designed, built, and operate the New Horizons spacecraft!