I wasn't going to comment on this one, but there's enough sidetracks, many interesting, that I don't see any harm. I do risk management in IT, as one of my core things, and have in industrial and public safety systems at all. Which isn't really much to say. Every engineer, every pilot, even any well raised kid has some training in risk management, whether they know it or not.
This accident seems to be to be pretty simple, I'm agreeing with many of the airline pilots here on that. And I'd add not just simple, but I'd bet almost inevitable.
We toss around the "swiss cheese" analogy to describe an accident where lots of unlikely things have to line up in an unpredictable way to cause an accident. It's a description that is pretty accurate for a scenario where multiple safety systems, all unrelated, fail. But that doesn't seem like that. Those accidents may be complicated in execution, but they're easy to understand. This one is where the experienced people just have a gut feeling that things aren't right. When I hear "this doesn't feel right" from peers in my field, I pay attention, because a lot of time they're picking up something in an almost analog way that they know isn't right but they can't put their finger on the specifics. They're not always right, but always worth listening too.
Anyway, a couple of truths are that no system's performance is ever 100% with 0 error. It's going to be +/- some tolerance, some % of the time. Plus or minus 50 feet in altitude 99% of the time, or whatever. If you make the number really big, then it might be 99.99% or whatever. But the math part says you don't get perfect. The same thing is true of any piece of moving anything built by man. It'll eventually fail. And even better, if you have a backup but there's no way to know the primary has failed, adding that backup doesn't necessarily improve your reliability ANY.
Back to this accident, simple version. In my mind it's a lot like the recent warbird crash in TX, that the pilots were more of less playing figure 8 racing on that final, but didn't know it. Before anyone says that's crazy, here's what I mean: In figure 8 racing, or in any environment where the most likely collision is at right angles, it's actually a little tough for the two objects to collide. They have to be in the same spot at the same time, and that happened here in DC. But each are potentially traveling at about 100 knots or more, and they're only about 100' long, maybe a bit less. 100 knots is something a bit over 150 ft/sec. So with zero vertical separation, they have to be on an intercept course within a second to hit each other. If my back of envelope math is right, and if there are guessing a helicopter flight every 15 minutes and a landing on that approach every 3, and if the courses lined up 100% in the vertical, each airplane flight would still have something like a 100,000:1 chance of colliding with a helicopter. Because they're approaching at right angles. The smaller they are, and the faster they go, the less chance of hitting.
But they flight paths did have vertical separation! And they did hit anyway! Why? Because the vertical separation depended on both aircraft staying in their lane, and because pilots and aircraft aren't perfect. The randomness of the altitude excursions was higher than an adequate safety margin for the vertical separation. Or in other words, the helicopter path was too close to the approach path. The accident proves it. It only took this long for it to be a problem because the right angle paths made all the other vertical separation problems unimportant.
To my knowledge we don't run things that closely anywhere else, but maybe we do.
As to the ADSB thing, I really think it's a red herring. ADSB, as others have pointed out, isn't by itself a collision avoidance system, it's a tracking thing. A pilot on final at a controlled field shouldn't be looking at ADSB in my opinion. Their focus should be on the task ahead, that's the high risk activity. The other pilot, flying about 200' above a river at 100k or whatever? if they're looking anywhere other than straight ahead, they're also looking at the wrong thing, in my opinion. But suppose someone else is looking? What evasive action does the helicopter pilot take? They're flying along a river, probably lower than the obstacles on each side. If they guess left or right, with a predicted right angle approach, they may or may not get it right. They can't go left or right much in any event. They can't go down. So they go up. Switch to the aircraft on final. Left or right or down on a final approach? I don't think so, not down and not much left or right. So they go up. It would be especially tragic if we find out that both aircraft saw each other, and both went up to avoid.
That last point puts us in a different spot, again in my view. You shouldn't put pilots in a spot to dodge traffic on final, and you shouldn't put a pilot flying 200' off the ground in a position to dodge traffic, either. Which brings us back to the first point, having that path under the approach path wasn't a good plan. Maybe in the 1950's where "eh, accidents are unavoidable" was accepted, but not now.
Long winded way to say the approach and helicopter routes weren't setup right, again in my view, but wanted to explain my reasoning, and maybe provide some math to the airline guys that could just look at it and say "yeah, that's a stupid plan" without even thinking about it.
www.ntsb.gov
