[dwc]

> Everything about the landing except the stuck leg seemed to be perfect.

And there's the issue: it only takes a tiny bit of imperfect to ruin a landing.

I do think that SpaceX is going to get to the point where they are recovering enough of their rockets to make it pay off. But there will probably always be a significant chance that some little thing will go wrong. Don't be fooled later by "we've landed 5 in a row" about that.

[thatswrong0]

I guess it depends on what you mean by significant. It's true that, inherently, space flight is risky business.[0] This is new technology, so of course they are going to continue to uncover weaknesses even much further down the line once they start hitting a higher success rate.

But for the tiny bit of imperfect: the same could have been said for the space shuttle, and that had people in it. And its fragility was a fundamental to its design (it rode exposed on the fuel tank with its critically important heat-absorbing tiles). And the shuttle still had a decent success rate!

I would argue that these failures are less fundamental to the design of the rocket. And the difference is that if, in the future, you have that 1/100 failure on landing, it's only a loss from a cost perspective, not a human one.

[0] http://i.imgur.com/ei3h1B7.png

[martinald]

This is very, very, early days. I'm sure you could have side the same about 'normal' flight, loads and loads of failures, then getting it to the '5 in the row', then a bunch of screwups.

Now jet flight across the oceans in often horrendous conditions is 99.999% safe (ish).

[Htsthbjig]

The number of things that could go wrong is limited. The will learn all of them.

The same happened with airplanes. Just something so simple as breaking sound barrier control inversion meant hundreds of pilots dying until someone figured it out..

Thanks to automation and telemetry, SpaceX is learning fast without killing anyone. In the past it was hard to analyze what went wrong when the pilot did not survive.