SpaceX progress towards a 100 launch year

[Robotbeat]

SpaceX is currently on track for about 105-110 launches in 2023 given constant overall improvement in capacity consistent with the 61 launches last year and 45 launches in the first half of this year.

For those claiming that seasonal effects (hurricanes) will mean they won't meet the 100 year launch rate:
Note that SpaceX got 27 launches in the first half of 2022 and 34 in the second half. A similar ratio between first and second half of the year this year would mean they get 56-57 launches in the second half of 2023 for a total of 101-102 launches in 2023.

[Robotbeat]

We're both pointing out the same thing:  90 launches is pretty damn good.  Through all 11 pages of the thread, no one at all suggested slighting SX if they don't make 100.  I said I was skeptical (and still am), and some tried to show how I wasn't getting it that they wouldn't, but neither I nor anyone else suggested a failure if they don't. 

This whole kerfluffle is silly.  No other firm can touch SX's launch capability.  Who cares if they don't hit 100 this year; they will eventually.

Other than the possibility of a Falcon failure in the next 55 launches (which is not a small probability! about 10-25%), where does the skepticism come from? They continue to make marginal improvements in launch rate, they tend to improve launch rates in the second half of the year...

Nothing is certain, but it seems like they're right on the path to accomplish it.

[xyv]

Sarcasm is just totally lost on some folks...

drnscr was pointing out (sarcastically) that it's no hit to credibility at all to only do 90 out of a target of 100...

Yea...I had to change the original thread title because of that.  But where is 45 coming from?  We are at 43 (or 44 counting SS).  The end point of a linear fit has been intersecting the end of year line at ~85 to 86 since I start tracking this at beginning of the thread.  As of today, it sits at ~87 (eyeballing it).  I am waiting until the end of the week to update because I keep expecting a popup Starlink launch. Whenever that happens it will tilt the line a skosh lower because it is already more than 4 days since last launch.

[Robotbeat]

I get 105-106 launches so far if you say 44 launches. I also couldn’t see where 45 launches came from.

Still well above the ~88-90 you get with a straight line.

Obviously if you’re improving gradually, the flat line projection starting from Jan 1st isn’t going to change dramatically but very gradually. Again, that’s just basic principles of calculus, ie that lines work great as a local approximation to basically any smooth and well behaved function.

[xyv]

Well aware of calculus and curve fitting.  I continue to track against 'progress' towards the stated goal.  Extrapolations of non-linear functions based on noisy data are fun (and I enjoy the contribution to the discussion) but are subject to large errors.  A straight linear projection is just a great score keeping metric.

[Robotbeat]

Exponential is the simplest possible function for this sort of thing, simpler than piecewise linear. It also fits with what we expect the behavior to be and so far in this thread, it has been meeting the desired goal basically every time. But anyway.

[meekGee]

So, how many launches for the first 6 months of this year?  45?  This places SpaceX on par for 90 launches this year.  Also, hurricane season is in now through oct-nov.

Mid year is between July 1 and 2, so let's call it 46.5...  for a yearly rate of 93 in the first half.

If they improve by 15%, the second half will be 107, and so the yearly will be 100 on the dot.

15% improvement per half year, that 32% improvement per year.

But last year they did 60, so they're up at least 50% year to year.

Also last year they launched more on the second half, despite hurricane season.

It may go down to the wire...

[meekGee]

Well aware of calculus and curve fitting.  I continue to track against 'progress' towards the stated goal.  Extrapolations of non-linear functions based on noisy data are fun (and I enjoy the contribution to the discussion) but are subject to large errors.  A straight linear projection is just a great score keeping metric.

A straight line model would have (in retrospect) been a very poor predictor, unless you look only at short windows (not a big surprise there)

[spacenut]

I don't think it matters how many they get.  It is still more than anyone else. 

[DanClemmensen]

I don't think it matters how many they get.  It is still more than anyone else.

I think it is getting close to being more than everybody else combined. This is absolutely the case for total payload mass to orbit.

[meekGee]

We're both pointing out the same thing:  90 launches is pretty damn good.  Through all 11 pages of the thread, no one at all suggested slighting SX if they don't make 100.  I said I was skeptical (and still am), and some tried to show how I wasn't getting it that they wouldn't, but neither I nor anyone else suggested a failure if they don't. 

This whole kerfluffle is silly.  No other firm can touch SX's launch capability.  Who cares if they don't hit 100 this year; they will eventually.

Other than the possibility of a Falcon failure in the next 55 launches (which is not a small probability! about 10-25%), where does the skepticism come from? They continue to make marginal improvements in launch rate, they tend to improve launch rates in the second half of the year...

Nothing is certain, but it seems like they're right on the path to accomplish it.

Yes, the possible failure is the biggest issue.

There is a non-zero possibility that they'll have to stand down for some period.

There is no possibility that the launch rate will suddenly increase.

Therefore, the expected number of launches, including possible failure, is well below 100.

Thing is, we don't know what F9's reliability is, especially since these rockets are now mostly known vehicles...  And so I don't know if it's as simple as saying they're 1-in-200 reliable or similar.

[Robotbeat]

I’d say the rockets are at least 1/200 reliable as that is demonstrated, but if you include the fact they’ve not even had an engine out in over 100 missions, you could claim the real reliability could be much higher, like up to twice the consecutive successful launches. So 1/200 to 1/400 or so.

[meekGee]

I’d say the rockets are at least 1/200 reliable as that is demonstrated, but if you include the fact they’ve not even had an engine out in over 100 missions, you could claim the real reliability could be much higher, like up to twice the consecutive successful launches. So 1/200 to 1/400 or so.

Beyond that..  if a rocket flew successfully 200 times and we call it 1:200 reliable, then we're basically assuming it'll fail in the next launch, and that's as pessimistic as possible.

[niwax]

I’d say the rockets are at least 1/200 reliable as that is demonstrated, but if you include the fact they’ve not even had an engine out in over 100 missions, you could claim the real reliability could be much higher, like up to twice the consecutive successful launches. So 1/200 to 1/400 or so.

I also like to look at the data point that landings are now on a 160 or so success run, and those are an entire step down in safety factor, with multiple non-redundant single failure points. It shows that good design and operations alone can break through typical historical rocket reliability. That's a good indicator reliability for the portion with a human-rated safety factor and multiple redundancies is higher still.

[mn]

snip snip

...
There is no possibility that the launch rate will suddenly increase.
....

Yes of course it is not possible for the rate to suddenly increase since the increase is expected based on previous experience...

Of course previous experience is no guarantee but it is certainly a good indicator. Although in this case they are running up against some limits of availability so time will tell. But if they do increase it will most certainly not be sudden and unexpected

 

[Robotbeat]

I’d say the rockets are at least 1/200 reliable as that is demonstrated, but if you include the fact they’ve not even had an engine out in over 100 missions, you could claim the real reliability could be much higher, like up to twice the consecutive successful launches. So 1/200 to 1/400 or so.

I also like to look at the data point that landings are now on a 160 or so success run, and those are an entire step down in safety factor, with multiple non-redundant single failure points. It shows that good design and operations alone can break through typical historical rocket reliability. That's a good indicator reliability for the portion with a human-rated safety factor and multiple redundancies is higher still.

Exactly. You could make a case the reliability could be up to 1:1000 for Falcon 9 launching human missions due to the greater certification requirements and use of greater margin, although I’m skeptical of any reliability claims greater than twice the actual track record of maximum consecutive successful launches.

[meekGee]

Hey, for all the "flat liner" predictors:

I'm asking this without looking at the data first.

When was the last time that a half year had less than a 10% increase over the previous half, in terms of number of launches?

I do remember that when they hit 60 in 2022, the second half had more launches than the first.

Definitely first half of 2023 had more launches than second half of 2022.

How far back does this go?

[wannamoonbase]

Hey, for all the "flat liner" predictors:

I'm asking this without looking at the data first.

When was the last time that a half year had less than a 10% increase over the previous half, in terms of number of launches?

I do remember that when they hit 60 in 2022, the second half had more launches than the first.

Definitely first half of 2023 had more launches than second half of 2022.

How far back does this go?

While I agree, at a point they will be limited by the ASDS cycle times, range availability and weather.  The further they go the closer they get to those barriers pushing back.

I think they can hit 100 but some luck will be required.

[Robotbeat]

Not luck but small, continuous improvements. ASDS has margin for more turnaround, plus they’re starting to use both coasts more and to even do RTLS for Dragon missions. I thought that was already established, so I’m not sure why folks still think east coast ASDS is the bottleneck for 100 launches in 2023. They probably could do up to 200 launches per year with the existing numbers of ASDSes as almost half the launches recently haven’t used east coast droneships.

Of the last 20 SpaceX launches, only 11 have used east coast ASDSes, the rest were Vandenberg, used the east coast ground pad, or were expendable (Falcon Heavy and Starship). And they still have room for improving this or for shortening the ASDS trips by taking advantage of Falcon 9 performance improvements to have a partial boost back to reduce ASDS turnaround time.

[Robotbeat]

The fastest theoretical turnaround time for a single ASDS doing 400-500km down range is about 3 days with 7knot speeds, so theoretically a single droneship could do 100 flights per year, they have 3 droneships, plus expendable flights (Falcon Heavy), ground pads on both coasts, and Starship. So they could do well over 300 flights per year just by tweaking their current setup continuously. We aren’t near a fundamental bottleneck (even if you include weather), just coordination and operational difficulties amenable to streamlining of the process. With ground pad usage increasing, I could see an upper limit of like 500 per year as far as being limited by 3 droneships.

There are other, more important constraints, like limited niobium supplies, range availability, etc, but nothing that stops 100 flights per year.