Given the history of DC's train control system (which requires manual operation far more than it should) and similar "loss of train" problems in the early days at BART, the investigation is heading where I suspected it would. The computer(s) basically forgot (or didn't see) that there was a stopped train, and tried to drive another one through it. Resulting in a violation of physical laws of spacetime. There will be many contributing factors as the story develops, but I suspect this will be confirmed as the primary cause.

BART avoids this by running trains on much greater headways than theoretically can be done, leaving plenty of room for braking. That limits capacity, of course, but BART doesn't need all of it. DC doesn't have that luxury - they have to use the system as designed, and occasionally it hiccups - with Really Bad Results this time. Might be interesting to see if there are records of how many times in normal operation somebody has to hit the "mushroom" because the computer got confused.

BTW - question for somebody on the inside perhaps - are the "1000 series" cars at DC similar to or the same as BART's originals? They were both apparently built by Rohr (one of those attempts to get aerospace people into transit/rail) and rebuilt or at least refurbished by Alsthom. I suspect DC runs theirs harder in worse weather resulting in more wear, but has BART had similar issues with performance? Besides those 2 systems, did anybody else ever get cars from Rohr or was that (like Boeing with the ill-fated Standard LRVs) a single shot?