My point is that a 100% track record is easy to achieve, if you are allowed to reset the counter just by renaming your company or forming an alliance.

They haven't done that. 106 launches.

You're missing the point. Lockheed and Boeing had initial failures when each company was started. It's not like Boeing put up rockets on day one without a problem.

Then, when each company approaches stability with their launches, they form a new alliance. That new alliance has "zero failures", but that's only because you ignore each company's original problem s.

I see your point, but it's not exactly relevant. As a purchaser in need of putting a multi-million dollar satellite in space, I could not care less that your company had failures in its infancy. I want to see your most-recent results as evidence of your ability to put my satellite in space. Comparing SpaceX, and whatever its success rate is, to ULA, and it's 100% in 106 launches, is easy.

The string of 106 successful launches is what I'm looking at. Just because you're new to the market doesn't necessarily mean you get a pass on quality/success. This is especially true when we're talking about putting expensive objects, that took years to build, in space.

> I could not care less that your company had failures in its infancy

> Just because you're new to the market doesn't necessarily mean you get a pass on quality/success

Contradiction?

No it's not. When SpaceX gets to 100 launches with no failures you could effectively reset their counter if you want.

ULA has currently 100% in orbit mission success rate which cannot be overlooked.

Now for some it might not be an issue considering SpaceX is considerably cheaper, but if you are putting a new spy satellite into orbit your risk metrics might be different.

What I meant was if you're willing to allow failure in a company's infancy, doesn't that mean they /do/ get a pass on quality/success?

The key point is that you're evaluating both companies TODAY, and they're at different points in their evolution. The fact that ULA/etc. had failures in their infancy is fairly irrelevant because that's the distant past - the fact that SpaceX had failures in its infancy is highly relevant because they're still IN their infancy.

The "pass" on quality/success is for things that happened in the distant past because they're less relevant to current performance, not some strange pass for anything that happens "during infancy" regardless of when that was.

Probably we can all agree, if we just say "30 launch strike" instead of "100% track record". The former is still impressive, whereas the second is a bit deceiving.

But it's not misleading, ULA has a stable launch system with a 100% success rate, SpaceX currently does not.

ULA can also currently life heavier cargo than SpaceX and the most important part is that they do not still tweak the system to "improve".

When SpaceX comes with Dragon9Next and it has 100% success rate you can in effect reset the track record for that platform, you can ignore the company but you cannot ignore the platform. If Dragon Heavy turns out to be 100% reliable after a 100 launches they could be able to play in the same league as ULA but currently they cannot. Launching supplies to the ISS is fine and all, same goes for commercial satellites but some activities especially when it comes to government agencies and manned space flight play by a different rule book of risk metrics and calculations.

The NSA doesn't care about insurance money if it's satellite explodes it cares that it now cannot intercept SIGINT over the south china sea and so the US which could affect US national security, the NRO doesn't care about insurance money it cares about being able to track Russian nuclear silos to make sure that the men who are guarding them are not stealing the warheads to buy potatoes. And when we are talking about manned flight well then considering the enormous costs of training an astronaut, as well the cost of loss of life and the reputation damage that any accident will incur even a 10 times decrease in KG to Orbit costs might not be enough to go with the cheaper but the "less reliable" option.

When predicting reliability of next launch, you'll use weighted mean of historical data, not simple mean. And, weight of launches in 60s would be near 0, same for n-100 launches.

Is the cost difference between the launches usually less than the cost of building a backup satellite?

Yes. Many times yes. The cost of launches is on the order of a few hundred million.

Military satellites often cost a few billion. Losing a satellite is, by far, the biggest risk during a launch. The launch cost itself is non-trivial, but a small fraction of the overall mission cost.

> Military satellites often cost a few billion.

Wouldn't it be more accurate to say the plans for a military satellite cost a few billion?

I would assume stamping out a new copy when you have the plans costs just a few tens of millions.

I mean, if you wanted one Airbus A380, it would cost you 15 billion dollars. But once you have the plans, you can stamp out additional copies for 100 million or so.

Each one is unique, because they are always being improved.

> Military satellites often cost a few billion

This needs a reference.

Not really, but here you go: https://en.wikipedia.org/wiki/KH-11_Kennan

And no, you can't make cheap primary mirrors (in a garage).