Abort options for Starship and Starship/SuperHeavy

[chopsticks]

I think we've managed to boil this down to the point where the two sides of this argument aren't going to be moved by the others' points any more.  However, I'm willing to bet 1,000,000 Internet Toldja So Points that NASA won't crew-certify Starship for crew launch/EDL any time before 2035 without an escape system.  And I'm willing to bet 250,000 ITSPs that SpaceX won't either.

Yep, I agree. And this is important, because the requirements from NASA (or other customers) will influence the design. If no one wants to fly on the thing because it's considered too unsafe, SpaceX is going to do something about it.

[Negan]

Use the capsule system for launch and my aircraft idea for landing.

[sebk]

I think we've managed to boil this down to the point where the two sides of this argument aren't going to be moved by the others' points any more.  However, I'm willing to bet 1,000,000 Internet Toldja So Points that NASA won't crew-certify Starship for crew launch/EDL any time before 2035 without an escape system.  And I'm willing to bet 250,000 ITSPs that SpaceX won't either.

I agree with your ideas, though I suspect some sort of pressure will be brought to bear to allow crew on this vehicle at some point.  But it's an inherently bad architecture.  Let this thing launch cargo and launch crew on something with abort systems - like F9 and Dragon.

I know, I know - upper-stage reusability.  Whatever, it's still an inherently unsafe architecture, reusable or not.

You idea is inherently unsafe.

Capsules with parachutes are inherently unable to achieve a high flight rate. And you absolutely do need a high flight rate to get safety to even get into the ballpark of motorcycle riding or general aviation (the general aviation is merely 2x safer than motorbike riding, BTW).

If you want to have a high flight rate with a capsule then you need a powered landing. At that point you have the same primary issue as the architecture you claim to be inherently unsafe.

So maybe use space planes able to glide to a landing? Space planes are either inherently not passively stable in hypersonic regime or they have horrible subsonic flight properties so they sink like a brick, or both. In effect inherently not safe.

BTW. By your thinking jet (turbofan) transport airplanes are inherently unsafe, too:
* it can't just stop in the air
* even worse, past certain speed it can't even just stop on a runway, it's committed to fly,
* it's supported by inherently dynamic and turbulent atmosphere
* runways are few and far between, the rest of the planet is not conductive to a safe landing
* it flies at 40000ft and 500knots (oh, so scary big numbers)
* its engines are super complex, and highly stressed, and store large amounts of rotational mechanical energy.
* mere collision with a flock of birds could knock it off the sky
* etc...

Yet that "inherently unsafe" thing is the safest means of travel bar city busses and elevators.

The "inherently unsafe" argument is without legs, and is generally a handwaving. It's not even remotely close to a legitimate engineering.

[chopsticks]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works. As a result, we compare it to the closest analogues that are familiar with, ie. other spaceflight architectures (with abort systems) and aviation (with no abort system). And then there's the Shuttle (crewed space vehicle with no abort), and we know how that worked out. The problem I have with comparing Starship to airliners is that airliners will always win because there are so many of them and have been for so many years that the data gap is a gulf that cannot be spanned.

Remember, *spaceflight* is inherently unsafe.

The questions is, how can we make it safer? If Starship accomplishes what SpaceX wants it to, there's a decent chance that by the sheer flightrate of the thing it might easily surpass the safest capsule (or other crewed) system flying today or ever. But that will never get to airliner numbers.

[Lee Jay]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

[Lee Jay]

I think we've managed to boil this down to the point where the two sides of this argument aren't going to be moved by the others' points any more.  However, I'm willing to bet 1,000,000 Internet Toldja So Points that NASA won't crew-certify Starship for crew launch/EDL any time before 2035 without an escape system.  And I'm willing to bet 250,000 ITSPs that SpaceX won't either.

I agree with your ideas, though I suspect some sort of pressure will be brought to bear to allow crew on this vehicle at some point.  But it's an inherently bad architecture.  Let this thing launch cargo and launch crew on something with abort systems - like F9 and Dragon.

I know, I know - upper-stage reusability.  Whatever, it's still an inherently unsafe architecture, reusable or not.

Capsules with parachutes are inherently unable to achieve a high flight rate. And you absolutely do need a high flight rate to get safety to even get into the ballpark of motorcycle riding or general aviation (the general aviation is merely 2x safer than motorbike riding, BTW).

If you want to have a high flight rate with a capsule then you need a powered landing. At that point you have the same primary issue as the architecture you claim to be inherently unsafe.

I don't particularly like capsules, but these arguments are nonsense.  They were nice in the 60s.

So maybe use space planes able to glide to a landing? Space planes are either inherently not passively stable in hypersonic regime or they have horrible subsonic flight properties so they sink like a brick, or both. In effect inherently not safe.

I disagree.  Keeping them stable in a hypersonic regime isn't challenging and landing a plane that has a high sink rate is quite easy compared to landing a high glide ratio glider.  This is because a steep angle makes a spot landing simple while a shallow angle results in a wide margin for error along the path of flight.  This is, in fact, why gliders have spoilers.

BTW. By your thinking jet (turbofan) transport airplanes are inherently unsafe, too:
* it can't just stop in the air
* even worse, past certain speed it can't even just stop on a runway, it's committed to fly,
* it's supported by inherently dynamic and turbulent atmosphere
* runways are few and far between, the rest of the planet is not conductive to a safe landing
* it flies at 40000ft and 500knots (oh, so scary big numbers)
* its engines are super complex, and highly stressed, and store large amounts of rotational mechanical energy.
* mere collision with a flock of birds could knock it off the sky

None of this nonsense has anything whatsoever to do with my argument.

[TheRadicalModerate]

I think we've managed to boil this down to the point where the two sides of this argument aren't going to be moved by the others' points any more.  However, I'm willing to bet 1,000,000 Internet Toldja So Points that NASA won't crew-certify Starship for crew launch/EDL any time before 2035 without an escape system.  And I'm willing to bet 250,000 ITSPs that SpaceX won't either.

I agree with your ideas, though I suspect some sort of pressure will be brought to bear to allow crew on this vehicle at some point.  But it's an inherently bad architecture.  Let this thing launch cargo and launch crew on something with abort systems - like F9 and Dragon.

I know, I know - upper-stage reusability.  Whatever, it's still an inherently unsafe architecture, reusable or not.

I'd be very surprised if there isn't an F9/D2-mediated phase to cislunar flight before Starship gets to full certification.  In any sane world, this would be a no-brainer.  But there's a sort of Catch-22 of insanity lurking around this:

1) F9/D2 + Starship + tankers is probably too expensive to split across four private customers.

2) The death throes (or the longer vegetative state) of SLS/Orion means that NASA's gonna drag their feet on using F9/D2/LSS together, even though it's vastly less costly, until they get shamed into it by private cislunar flights.

So SpaceX kinda needs to be able to launch 12ish at a time, without the extra $200M-250M for the F9/D2, in order to get NASA to relent.  So sooner is better.

If everybody were fine with missions of no larger than 4 for ten years or so, then there'd be watchful waiting while F9/D2 did its thing, accumulating data on cargo Starship launch/EDL safety, to see if it can be certified more-or-less as-is, and then you'd get crews of 12+ sometime in the mid-2030's.  But if they need to get to 8-10 people before then--or if the accumulated safety data looks bad--they're going to have to invest in the escape kludgery.

PS:  It does occur to me that the way to slice the Gordian Knot of private pricieness is for SpaceX to subsidize a couple of private D2-mediated cislunar flights.  They can call them "experimental" and get Jared to lead them, using the negatively-paid test pilot model he seems to have adopted, but at a price he can afford.  They could even get him to aerocapture one into LEO, if they're feeling comfy with the hypersonic regime.  Then NASA will sign on to the D2-mediated model, and SpaceX can wait for the architecture to evolve enough to get crew-certified for the whole nine yards.

[volker2020]

This discussion has taken some strange turns. Just to establish some base lines again:

* the only goal for SpaceX to build SpaceShip is to be able to transfer 1 million people to Mars.
* they have a very short timeline, being the lifetime of Elon Musk, because I don't think that this project would survive his death.
* so the question has never been, do they use SpaceShip to launch a lot of people, but when.
* there are error vectors in StarShip that can not be migrated, and there is the clear risk of failure, most likely killing anyone on board.
* but any kind of transportation has this error vectors, so the only question is, can they be reduced to a level that is acceptable. But there is no general rule, what is acceptable. Acceptable is a personal choice, not an objective one, otherwise certain hobbies would be prohibited.
* migrations to the error vectors, that make it impossible to use StarShip for mass transport will not happen, unless they are done in a prototype level (and I think that they will never happen, because SpaceX will try to root out this problems when flying cargo).

[sebk]

I agree with your ideas, though I suspect some sort of pressure will be brought to bear to allow crew on this vehicle at some point.  But it's an inherently bad architecture.  Let this thing launch cargo and launch crew on something with abort systems - like F9 and Dragon.

I know, I know - upper-stage reusability.  Whatever, it's still an inherently unsafe architecture, reusable or not.

Capsules with parachutes are inherently unable to achieve a high flight rate. And you absolutely do need a high flight rate to get safety to even get into the ballpark of motorcycle riding or general aviation (the general aviation is merely 2x safer than motorbike riding, BTW).

If you want to have a high flight rate with a capsule then you need a powered landing. At that point you have the same primary issue as the architecture you claim to be inherently unsafe.

I don't particularly like capsules, but these arguments are nonsense.  They were nice in the 60s.

You call reality nonsense. Unfortunately for you reality doesn't care.

Capsules by necessity have very low lift to drag and thus steep reentry. That means less total heat pulse but much worse peak heating. Which in turn means higher thermal stress of the heatshield. Which in turn means none of the TLR>3 reusable heatshields would hold. In effect capsule effectively means ablative heat shield which is incompatible with a high flight rate.

But even if currently low TRL solutions were developed, you still have a parachute which is needs at least a refurbishment:large parachutes, unlike those made for individual humans, are complex devices with pyros and delay charges, go through high stresses and are not quickly reusable.

So maybe use space planes able to glide to a landing? Space planes are either inherently not passively stable in hypersonic regime or they have horrible subsonic flight properties so they sink like a brick, or both. In effect inherently not safe.

I disagree.  Keeping them stable in a hypersonic regime isn't challenging

They need active controls to be stable in hypersonic regime. If you lose control, they turn ballistic in a few seconds. Once they are ballistic, they can't withstand structural loads and they fall into pieces (Columbia lost control and in 35s fragmented due to g-overload; at that point there were no indications of overheat except the parts around the left wing which was being penetrated by re-entry plasma for a dozen minutes already).

and landing a plane that has a high sink rate is quite easy compared to landing a high glide ratio glider.  This is because a steep angle makes a spot landing simple while a shallow angle results in a wide margin for error along the path of flight.  This is, in fact, why gliders have spoilers.

That would be true if you talked about descent slope of few degrees. But here were are in several dozen degree descent slope. Shuttle was initially supposed to be a lifting body. As the size grew lifting body became infeasible, because it would have even worse sink rate and so high landing speed that no existing runway would be long enough (maybe with the exception of that Edwards one, but this is the only one in the whole worlds).

Moreover lifting bodies have odd shapes making fuel tanks and other pressure vessels to have suboptimal mass which makes the whole thing mass inefficient. Which means large vehicle size compared to even small payload which in turn means large landing speeds even for moderate size missions. So you're committed to a single gliding landing approach at a high sink rate with no abort options (there's no mass budget for those, it has been eaten by oddly shaped pressure vessels, high speed landing gear, etc...) 

BTW. By your thinking jet (turbofan) transport airplanes are inherently unsafe, too:
* it can't just stop in the air
* even worse, past certain speed it can't even just stop on a runway, it's committed to fly,
* it's supported by inherently dynamic and turbulent atmosphere
* runways are few and far between, the rest of the planet is not conductive to a safe landing
* it flies at 40000ft and 500knots (oh, so scary big numbers)
* its engines are super complex, and highly stressed, and store large amounts of rotational mechanical energy.
* mere collision with a flock of birds could knock it off the sky

None of this nonsense has anything whatsoever to do with my argument.

Yes, I'm being facetious, but this is exactly how your argument is built.

For example, you wrote about high reentry temperatures. This is big number scare of the same kind as claiming FL400 is high and 500 knots is fast. Because in the very same plane you have jet engines with fast moving parts at even higher temperatures, intricately orchestrated active cooling, exotic materials, etc. In fact while reentry tech has been succesfully implemented decades ago by most large world's powers decades ago, good turbofans are West-only. 

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Replace highly complex operating propulsion with highly complex guidance depending of chaotic and dynamic atmosphere and you have a high sink rate unpowered spaceplane returning. Returning lifting bodies "inherently fly" as well as a human in a wing suit, i.e. they have one and only one shot at an approach and landing. Gliding spaceplane (pretty narrow to begin with) margin of error is eanten away by the uncertainty of atmospheric conditions on which it critically depends to execute the landing.

And the spaceplane is also passively unstable in the hypersonic part of the EDL. Lose control there and game over.

[meekGee]

The idea is that when the heat shield is partly compromised, an underlying steel structure will last a lot longer than an underlying aluminum one.

Which is irrelevant if they both fail.

Yes, obviously. I assume the discussion is about the cases where details means one or the other has a larger chance of surviving.

The temperature at the structure will not be that of the outer skin, so the ability to withstand elevated temperatures really matters.

Since the skin is the structure, I don't know what this means.

Not on the side covered in TPS, the glass coating on the TSP is the outer skin with regard to heating.

Similarly if for whatever reason the peak temperature reached is higher than planned, temperatures will rise on the inner side of tge heat shield, and again Stainless will perform much better than Aluminum.

Which, again, doesn't matter since neither one will survive entry temperatures if the heat shield is compromised or if the entry is otherwise off-nominal.

This is the same as saying that reentries are impossible and meteorites do not exist since temperatures during reentry are far above the vaporization temperature of all known materials...

Yup exactly on all. LJ was just trying to intentionally misunderstand.

[Lee Jay]

I agree with your ideas, though I suspect some sort of pressure will be brought to bear to allow crew on this vehicle at some point.  But it's an inherently bad architecture.  Let this thing launch cargo and launch crew on something with abort systems - like F9 and Dragon.

I know, I know - upper-stage reusability.  Whatever, it's still an inherently unsafe architecture, reusable or not.

Capsules with parachutes are inherently unable to achieve a high flight rate. And you absolutely do need a high flight rate to get safety to even get into the ballpark of motorcycle riding or general aviation (the general aviation is merely 2x safer than motorbike riding, BTW).

If you want to have a high flight rate with a capsule then you need a powered landing. At that point you have the same primary issue as the architecture you claim to be inherently unsafe.

I don't particularly like capsules, but these arguments are nonsense.  They were nice in the 60s.

You call reality nonsense. Unfortunately for you reality doesn't care.

Capsules by necessity have very low lift to drag and thus steep reentry. That means less total heat pulse but much worse peak heating. Which in turn means higher thermal stress of the heatshield. Which in turn means none of the TLR>3 reusable heatshields would hold. In effect capsule effectively means ablative heat shield which is incompatible with a high flight rate.

And powered landing fixes this how, exactly?

But even if currently low TRL solutions were developed, you still have a parachute which is needs at least a refurbishment:large parachutes, unlike those made for individual humans, are complex devices with pyros and delay charges, go through high stresses and are not quickly reusable.

And you think that's less work that turning around a ship that's 100 times larger?

So maybe use space planes able to glide to a landing? Space planes are either inherently not passively stable in hypersonic regime or they have horrible subsonic flight properties so they sink like a brick, or both. In effect inherently not safe.

I disagree.  Keeping them stable in a hypersonic regime isn't challenging

They need active controls to be stable in hypersonic regime. If you lose control, they turn ballistic in a few seconds. Once they are ballistic, they can't withstand structural loads and they fall into pieces (Columbia lost control and in 35s fragmented due to g-overload; at that point there were no indications of overheat except the parts around the left wing which was being penetrated by re-entry plasma for a dozen minutes already).

Columbia was missing a wing.

STS proved that this is not challenging with 1970s technology.

and landing a plane that has a high sink rate is quite easy compared to landing a high glide ratio glider.  This is because a steep angle makes a spot landing simple while a shallow angle results in a wide margin for error along the path of flight.  This is, in fact, why gliders have spoilers.

That would be true if you talked about descent slope of few degrees.

It's true anyway.  These vehicles can roundout and flare.

But here were are in several dozen degree descent slope. Shuttle was initially supposed to be a lifting body. As the size grew lifting body became infeasible, because it would have even worse sink rate and so high landing speed that no existing runway would be long enough (maybe with the exception of that Edwards one, but this is the only one in the whole worlds).

Dream Chaser is a lifting body, and it has the same glide slope as Shuttle.

Moreover lifting bodies have odd shapes making fuel tanks and other pressure vessels to have suboptimal mass which makes the whole thing mass inefficient. Which means large vehicle size compared to even small payload which in turn means large landing speeds even for moderate size missions. So you're committed to a single gliding landing approach at a high sink rate with no abort options (there's no mass budget for those, it has been eaten by oddly shaped pressure vessels, high speed landing gear, etc...) 

Which is irrelevant to my point.

For example, you wrote about high reentry temperatures. This is big number scare of the same kind as claiming FL400 is high and 500 knots is fast.

No, I was stating that, in the event of a failed heat shield, the fact that underlying structure is stainless isn't going to prevent burn through.

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Replace highly complex operating propulsion with highly complex guidance depending of chaotic and dynamic atmosphere and you have a high sink rate unpowered spaceplane returning.[/quote]

Okay, that's not even close.  STS and X-37 have proven that the later is very simple and reliable compared to the former.

Returning lifting bodies "inherently fly" as well as a human in a wing suit, i.e. they have one and only one shot at an approach and landing.

As does a powered landing, but gliding inherently works without the need for fuel, pressure, ullage, startup, propulsion, TVC, and so on, and it has a far wider margin for error (a long runway versus short "chopsticks".

Gliding spaceplane (pretty narrow to begin with) margin of error is eanten away by the uncertainty of atmospheric conditions on which it critically depends to execute the landing.

And the spaceplane is also passively unstable in the hypersonic part of the EDL. Lose control there and game over.

All of which were overcome in the 70s and were proven to be reliable by STS and X37.

[chopsticks]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

If this were to happen, this architecture would empirically be safer than any other crewed spacecraft, even with all of its complexity. What do we say then?

(Actually, for this thought experiment let's ignore the architecture altogether. It could a spaceplane, capsule, a bouncy ball, anything - as long as it fullfils the purpose of this post)

[DanClemmensen]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

If this were to happen, this architecture would empirically be safer than any other crewed spacecraft, even with all of its complexity. What do we say then?

(Actually, for this thought experiment let's ignore the architecture altogether. It could a spaceplane, capsule, a bouncy ball, anything - as long as it fullfils the purpose of this post)

I understand your point, but that's not how probability math works. If the probability of a failure is 1/270, then probability of success is 1-(1/270), and probability of 270 successful missions with no failures is (1-1/270)^270. This is 0.367. In other words, for every three sets of 270 flights, we would expect two of them would include at least one failure. Can a real failure analyst or statistician please step in here?

[JayWee]

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

We don't need to imagine. Starship is designed for Starlink V2 first and foremost. 30k V2 satellites require ~300-600 launches.
Read that again - yes, Superheavy/Starship has currently that many launches on its manifest already.

So putting people on only after couple hundred of launches is eminently possible and likely.

[Lee Jay]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

If this were to happen, this architecture would empirically be safer than any other crewed spacecraft, even with all of its complexity. What do we say then?

(Actually, for this thought experiment let's ignore the architecture altogether. It could a spaceplane, capsule, a bouncy ball, anything - as long as it fullfils the purpose of this post)

Then I would say the same thing I've said above - it may be possible to eventually make it safe enough for paid, professional astronauts (possible but doubtful to me), but it will never be safe enough for the general public, which is it's express reason to exist.

[InterestedEngineer]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

If this were to happen, this architecture would empirically be safer than any other crewed spacecraft, even with all of its complexity. What do we say then?

(Actually, for this thought experiment let's ignore the architecture altogether. It could a spaceplane, capsule, a bouncy ball, anything - as long as it fullfils the purpose of this post)

I understand your point, but that's not how probability math works. If the probability of a failure is 1/270, then probability of success is 1-(1/270), and probability of 270 successful missions with no failures is (1-1/270)^270. This is 0.367. In other words, for every three sets of 270 flights, we would expect two of them would include at least one failure. Can a real failure analyst or statistician please step in here?

Naively, yes.

But then tell my why Dragon-2 has 1:270 even though it's flown what, a mere dozen flights?

This is like showing up at a gun fight with a blunt pencil.  Or apples/oranges if you prefer that kind of metaphor.

There's no reason why over 3-4 years of actual flight the same statistical analysis can't be applied to Starship.  Using real data, which narrows the confidence intervals considerably.

If you've only flown a dozen flights, the 1:270 probability of LOC is a wild guess.  Kind of like the wild guesses about reliability of SRBs or re-entry reliability on a certain prior spacecraft design that a certain someone upthread claimed "solved all these problems in the 1970s"

The sooner Dragon-2 is retired in favor of something with actual departure count in the 100s, the safer for everyone who flies to space and back.

[Lee Jay]

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

We don't need to imagine. Starship is designed for Starlink V2 first and foremost. 30k V2 satellites require ~300-600 launches.
Read that again - yes, Superheavy/Starship has currently that many launches on its manifest already.

So putting people on only after couple hundred of launches is eminently possible and likely.

You're assuming those launches and landings are all successful.  What do you think the odds of that are?

[DanClemmensen]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

If this were to happen, this architecture would empirically be safer than any other crewed spacecraft, even with all of its complexity. What do we say then?

(Actually, for this thought experiment let's ignore the architecture altogether. It could a spaceplane, capsule, a bouncy ball, anything - as long as it fullfils the purpose of this post)

I understand your point, but that's not how probability math works. If the probability of a failure is 1/270, then probability of success is 1-(1/270), and probability of 270 successful missions with no failures is (1-1/270)^270. This is 0.367. In other words, for every three sets of 270 flights, we would expect two of them would include at least one failure. Can a real failure analyst or statistician please step in here?

Naively, yes.

But then tell my why Dragon-2 has 1:270 even though it's flown what, a mere dozen flights?

This is like showing up at a gun fight with a blunt pencil.  Or apples/oranges if you prefer that kind of metaphor.

There's no reason why over 3-4 years of actual flight the same statistical analysis can't be applied to Starship.  Using real data, which narrows the confidence intervals considerably.

If you've only flown a dozen flights, the 1:270 probability of LOC is a wild guess.  Kind of like the wild guesses about reliability of SRBs or re-entry reliability on a certain prior spacecraft design that a certain someone upthread claimed "solved all these problems in the 1970s"

The sooner Dragon-2 is retired in favor of something with actual departure count in the 100s, the safer for everyone who flies to space and back.

Of course. You need a carefully-constructed statistical failure model, using all of that advanced statistics that I forgot immediately after taking the final exam (grade = 100%) back in 1973. The professor asked us what we had learned. I responded that I now knew that I should hire a professional statistician whenever non-trivial statistics were involved. Statistics are a way to quantify what you do not know, and a good model will yield better guesses and a purely qualitative analysis.

My point stands: 270 non-fatal launches does not equate to a failure probability of better than 1:270.

[InterestedEngineer]

I feel like the argument of the architecture being inherently unsafe is mostly based on the fact that it's unproven and that there is no empirical data on how this approach works.

No it's not.  It's based on the idea that it has few abort options, probably no pad abort options, and everything has to go basically perfectly for it to land, including both highly-complex operating propulsion and landing in the right spot plus or minus some small number of meters while being caught by a tower with two swing-arms.  It doesn't inherently fly like a glider does and it doesn't inherently land safely like a capsule under parachute does, each of which has a far higher margin for error.

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

If this were to happen, this architecture would empirically be safer than any other crewed spacecraft, even with all of its complexity. What do we say then?

(Actually, for this thought experiment let's ignore the architecture altogether. It could a spaceplane, capsule, a bouncy ball, anything - as long as it fullfils the purpose of this post)

I understand your point, but that's not how probability math works. If the probability of a failure is 1/270, then probability of success is 1-(1/270), and probability of 270 successful missions with no failures is (1-1/270)^270. This is 0.367. In other words, for every three sets of 270 flights, we would expect two of them would include at least one failure. Can a real failure analyst or statistician please step in here?

Naively, yes.

But then tell my why Dragon-2 has 1:270 even though it's flown what, a mere dozen flights?

This is like showing up at a gun fight with a blunt pencil.  Or apples/oranges if you prefer that kind of metaphor.

There's no reason why over 3-4 years of actual flight the same statistical analysis can't be applied to Starship.  Using real data, which narrows the confidence intervals considerably.

If you've only flown a dozen flights, the 1:270 probability of LOC is a wild guess.  Kind of like the wild guesses about reliability of SRBs or re-entry reliability on a certain prior spacecraft design that a certain someone upthread claimed "solved all these problems in the 1970s"

The sooner Dragon-2 is retired in favor of something with actual departure count in the 100s, the safer for everyone who flies to space and back.

Of course. You need a carefully-constructed statistical failure model, using all of that advanced statistics that I forgot immediately after taking the final exam (grade = 100%) back in 1973. The professor asked us what we had learned. I responded that I now knew that I should hire a professional statistician whenever non-trivial statistics were involved. Statistics are a way to quantify what you do not know, and a good model will yield better guesses and a purely qualitative analysis.

My point stands: 270 non-fatal launches does not equate to a failure probability of better than 1:270.

Then by trivial logic going all the way back to Aristotle, a dozen Dragon-2 flights does not equate to a failure of probability better than 1:270

No fancy statistics required.

[chopsticks]

Ok. But let's go on a thought experiment. Let's imagine a scenario where Starship executes a higher number of successful uncrewed landings than the LOC requirement for Dragon (I think it's 1:270?) So if Starship lands successfully 271 times consecutively, it therefore has exceeded NASA's requirement for crew safety.

We don't need to imagine. Starship is designed for Starlink V2 first and foremost. 30k V2 satellites require ~300-600 launches.
Read that again - yes, Superheavy/Starship has currently that many launches on its manifest already.

So putting people on only after couple hundred of launches is eminently possible and likely.

No, we don't know that yet because Starship hasn't actually flown those flights. We don't know what the outcome will be.