* SpaceX: $63.5 million ($60 million * 18 total / 17 successful)
So SpaceX saves you $46.5 million per successful launch. If we assume that 1/18 SpaceX launches fail, then in the long run SpaceX wins as long as your typical payload cost is less than $46.5 * 18 = $837 million.
If we assume that the cost to replace the payload for a failed launch is less than the original payload (R&D costs are amortized, so it's likely that building a second payload is much cheaper than designing and building the first payload), then the break-even payload cost can be even higher.
For manned space flight, so far 18 people have died out of 430 that went to space. (14 Americans in two shuttle accidents, and 4 Russians in two separate incidents)
That's only slightly lower than 1/18.
The shuttle had a 40% vehicular failure rate (2 out of 5) and a 1.5% flight failure rate. That's better than 1/18, but not by that much.
Clearly, manned vehicles demand a higher level of reliability than cargo carriers. However spaceflight even after decades of experience is still pretty risky. From a life safety perspective, non-manned, robotic missions are preferable. But our drive as a species to explore and populate new territories means that we'll need vehicles safe enough to travel in, and/or people who want to do it regardless of the risk. 1/18 is still probably decent odds compared to the ones facing people 300 years ago, who set out in wooden sailing ships to explore the unknown.
With manned spacecraft, you have to factor in the chance that the crew would survive a failed launch. Specifically, in the press conference following this latest failure, it was stated that the planned crew escape system would have been capable of saving the crew. To my non-expert eyes looking at the event, there seemed to be a fair amount of time between there being an obvious problem, and complete disintigration.
At some point, yes, there is a monetary value to human life. During the early days of space flight, it was something like $1 million per astronaut. These days, it's up to something more like $50 million per astronaut (inflation would put it as more like $7 million).
One additional variable would be the cost to replace any assets lost on the 1 unsuccessful launch, although I'm not sure who ultimately picks up the tab for that.
Yeah, my back of the napkin analysis is probably missing a bunch of important stuff. I just wanted to highlight the fact that simply looking at the failure rates is not enough -- you have to consider the amortized cost of getting your stuff into space successfully. Maybe tolerating a few failures saves money in the long run.
(Of course this all assumes a replaceable payload. I am sure the analysis would be quite different for carrying people into space.)
* United Launch Alliance: $110 million
* SpaceX: $63.5 million ($60 million * 18 total / 17 successful)
So SpaceX saves you $46.5 million per successful launch. If we assume that 1/18 SpaceX launches fail, then in the long run SpaceX wins as long as your typical payload cost is less than $46.5 * 18 = $837 million.
If we assume that the cost to replace the payload for a failed launch is less than the original payload (R&D costs are amortized, so it's likely that building a second payload is much cheaper than designing and building the first payload), then the break-even payload cost can be even higher.