OK, I obviously did a lousy job of making my point, so I'll try again. Let's postulate that go-arounds occur every 2000 flying hours (my SLF experience suggests it's more frequent, but perhaps I've just been unlucky). Although it hasn't always been this way, today engine failures occur less than once per 100,000 hours, yet pilots spend countless hours learning how to deal with them. It's been a long time since we lost a multi-engine commercial airliner due to a single engine failure. That's because the aircraft is designed for it and the flight crews train for it.
We've had at least two major crashes this year due to botched go-arounds. It's been suggested here that having to do a go-around is a white knuckle experience - my point is it shouldn't be. Perhaps the reason for the go-around is white knuckle, but past the startle factor doing a go around should be routine. Unlike an engine failure, it's pretty much a given that a professional pilot is going to have to perform many go-arounds during their flying career, and getting just one wrong could make for a really bad day.
Maybe it means more simulator time doing go-arounds - with some failures thrown in (such as the TOGA switch not working). I'll leave that to the crew training experts. But we do need to figure this out - we can't let botched go-arounds become a leading cause of air disasters.
As a designer, I worry about failures that may occur once every 10 million or even 100 million hours - and one part of the safety analysis is to take credit for the pilot being able to successfully perform a go-around if needed. It never entered the equation that the go-around itself could cause a crash...