Risk analysis covers a lot of ground. John Denver lost his life because of a fuel valve that was in an awkward spot. People have flown perfectly good jet airplanes into the sides of mountains on bright, sunny, clear days while screwing around with the flight computer. More people than you can shake a stick at have gotten killed by simply running out of gas.
Not a lot of guys want to go to jail over an airplane, but people dent up airframes every day. Aviation is a great sport, hobby, and profession, but it is also very unforgiving. If I had to pick one, I'd take the angry guys over unfamiliar complex systems, especially in hard weather.
You could argue the point either way, but since he's still alive he's obviously played the odds well. I wonder, however, if all super-repo men are of the same caliber.
But that doesn't mean that taking off solo in a plane you know only from manuals is always going to result in an accident, either. Even DUI drivers "usually" get where they're going without incident.
Reading the article (yeah, with a little salt applied), this is a guy who faces down armed resistance. So the appropriate question of risk isn't "is an accident possible?", it's "Is an accident more likely than the other risks assumed?". There, I think he's on much firmer ground.