Metro-North crash and the push-pull argument

The tragic crash of the New York Metro-North train, which derailed coming into a sharp curve last weekend and killed 4 passengers, is likely to renew concern about the push-pull system.  The train was cab-car forward, with the engine at the rear.  It looked from the photographs like a Gennie, which is the GE-Genesis model used by Amtrak on its long-distance lines (and different from our usual Pacsurfliner F59s).

In discussions with non-train-riding friends on line, I had to explain that no, the engineer is not in the locomotive facing backwards when it is pushing.  People find it scary to think there's no one in the locomotive, but I reminded them that when they see a big freight with 4 engines, the followers aren't staffed either.

I also had to explain that trains live in a one-dimensional world, and can't simply turn around.  Instead, they must find a wye and jockey back and forth to reverse direction.  This isn't very practical for most commuter runs, a complication compounded by "stub" stations like LA Union, which dead-end in one direction.

Why did the Metro North train enter a sharp 30mph curve at 82mph?  The news today suggests that the engineer had dozed off.  I thought the throttle used a "dead man's hand" system, where a loss of tension would slow it down;  if it does, the engineer didn't loosen his grip while he dozed.  By all accounts he was a solid engineer.  It's a tragedy however it happened, and our thoughts are with the injured and mourning.