An emergency application originating because of the derailment is not a UDE. I can think of several cases where I did NOT place the train in emergency. At low speeds on poor track - it is sometimes better to just set the bracks hard and keep pulling on them. You might end up with a couple more cars on the ground, but keep them upright. The last instance I had a tourist train when an empty semi ran a crossing. I knew I would probably win that argument, and didn't want to chance knocking any standing passengers off their feet. The only thing broken that day was the horn handle. Another time I was running the main at 45 mph with 20+ old rotten coal cars of sand and gravel. A school bus had stopped for the crossing, but a loaded semi of sugar beets was going to pass them into the crossing. I would have hit calf first knocking the semi away from the bus, but had every expectation of piling up the train on top of us all. The truck stopped and backed up at the last minute. If I had gone into emergency, I wasn't going to stop short anyway and didn't want to chance dumping the train.

Other times, the brakes don't go into emergency when they should. I rocked an empty car off the track on the rear of a train this summer - and then drug it for 10 miles with one truck on the ground. I stopped for a sun kink, and by then the railroad knew something was wrong because of the bonds knocked off at a highway crossing had the flashers still on.

A UDE is an undesired emergecy - generally originating from a non-derailing event in the train. It can be caused by the air hoses parting (low or poorly connected hoses), a faulty emergency portion of the brake valve ("kicker"), slack changes during a brake pipe reduction or any time the brake pipe or hose fails. I've been on many trains with kickers - one was one engine and something like three cars, but the AAV on the ABDW was cycling when it shouldn't. Every time you touched the train air the brakes went into emergency. I never did like ABDW. The newer valves (ABDX and DB20/60) were specifically designed to stabilize the emergency feature, and older valves are also being modified to improve their performance.

When an emergency brake application occurs, for whatever reason, the engineer has basically lost control of his train and you just have to ride it out. You can bail off the engine brakes, and depending on the engine might still have dynamics or a few more seconds of power - but for the most part you have no control over how and when the slack will change. The brake valves are designed to create the brake cylinder pressure in controlled steps hoping that the slack will adjust. I've never dumped a train during an emergency brake application, but it is a real possibility that varies with each situation.

When you make a mistake with the throttle and air, you rarely have time to correct it and just have to suffer the consequences. Years ago I screwed up trying to power brake 10 empties on the head end past the conductor to make a cut. I had just made another reduction when he said that'l do, and I closed the throttle doing about 5 mph. I knew I was in trouble the moment the head end immediately stopped. I yelled on the radio for the conductor to stay in the clear and a few seconds later got blasted forward about 30 feet when the rest of the train - all loads - came off the hill against the empties and engines but luckily everything stayed ont the rail.

At higher speeds, the slack changes seem to be far easier to control (assuming you don't get too crazy or stupid, and the engines are behaving). At 10 mph if there is a substantial speed differential the slack will kick you - hard. But at higher speeds tat is a much smaller percentage of the train's average speed and the train seems to able to absorb it better.

Modern trains have another consideration - DPUs. Will the remotes drop out immediately? Plus on a manifest freight they usually have some loads blocked on the rear for the DPUs to shove against. And as mentioned, where in the train the emergency originates does affect slack.