Reporting by news media informs the general public that shutdown of a locomotive engine (by firefighters on the scene) caused train brakes to become released. It's implied that a locomotive was no longer keeping up its main reservoir air supply, and therefore the entire train lost its air brakes and subsequently rolled away. I disagree.
As you professional and casual railroaders here can attest, that's not how the air brake system on locomotives and cars functions. There's a critical difference between "straight" brakes and "Westinghouse" brakes, as described in a Wikipedia article entitled Railway air brake, [en.wikipedia.org] (which the news media and general public would have little knowledge about).
Leading up to this tragic incident, in less than one hour 70 cars and 5 locos released and rolled away - for me, this doesn't add up as a slow loss of locomotive main reservoir pressure. Whether or not the locomotive's main reservoir bled off, that's not what was holding the train and its cars against the grade ... train braking effort would have been initiated from a service application made while stopping the train, perhaps a 20 pound reduction of train line air pressure, resulting in car brake pistons being actuated with air supplied from each car's reservoir. Hand brakes on locomotives and cars (subsequently applied in sufficient quantity as prescribed by operating rules) would have provided another static braking force at the wheels, effectively a parallel or backup brake system to hold the train in place when "bottled air" in the car reservoirs eventually bleeds off.
It's been reported that handbrakes had been applied on the locos, not the cars, so apparently bottled air in the train was temporarily keeping things from rolling off the mountain when the train was tied down. "Leaky" (or any professional railroaders familiar with air brakes), how long might it take for typical freight cars (with proper COTS work) to eventually lose charge of air in each car's air reservoir?