MacDaddy,
10 to the minus 9 fatal accident risk per hour seems a bit ambitious.
Say NGs fly 5000 hours a year? There are 4000 of them (now, not through their lifetime). They've been around for 15 years. There's been three reported incidents ("recently" you say, not over the lifetime). To be generous, 15x5000x4000/3 = 10 to the 8. So, you have to assume that the least sophisticated crew anywhere won't steer off a cliff or into a colleague if this happens to them in more than about 1 in 10-100 incidents?
The 10^9 is a case that can be argued, but it doesn't seem open and shut.
I guess if the typical flight time is 3 hours, then it makes the odds per takeoff a bit better.