Author Topic: A bet on whether re-usable is really cheaper  (Read 24989 times)

Offline AC in NC

  • Full Member
  • ***
  • Posts: 361
  • Raleigh NC
  • Liked: 283
  • Likes Given: 335
Re: A bet on whether re-usable is really cheaper
« Reply #80 on: 04/20/2018 10:50 PM »
Nothing is proven strictly from anything except maybe mathematics, y’all. We’re talking preponderance of evidence. For everything.

"Reuse is cheaper" boils down to "price of rocket A < price of rocket B", which is a simple mathematical relation and can be proven.

Even in the case where "A" is the same rocket as "B", just with reuse.  We simply need someone to buy both and tell us how much they paid... :D

It doesn't boil down to that at all.  Customer price is not the meaningful measure of cheapness for the purpose of this discussion and any wagers therefrom.
« Last Edit: 04/21/2018 01:51 AM by AC in NC »

Online LouScheffer

  • Full Member
  • ****
  • Posts: 1837
  • Liked: 2364
  • Likes Given: 262
Re: A bet on whether re-usable is really cheaper
« Reply #81 on: 04/21/2018 01:07 PM »
I'd take the bet right now except for 2 things:

1. I think because it's crossing various governmental boundaries that it may be illegal in the US.

2. It'll be a fairer bet when it's all Block 5 boosters.

I'm amused by the idea that if re-use wasn't a good thing SpaceX wouldn't be doing it. Considering the foolishness that Elon is doing over at Tesla, I wouldn't be surprised at all if SpaceX wasn't also neck-deep in folly.

So, once they go all Block 5 and I know I won't be violating gambling laws, I'll check back on this thread and see if the bet is still on.


This looks promising.

If you have Paypal, I could simply send you the initial amount.  Then after each Block 5 launch for the next 3 years, we settle up, also by Paypal. If your PayPal account is not USA-based, then it would make more sense to do in blocks, perhaps every 10 launches, since there are international transfer fees.   But if you have, or could establish, a USA based account, then we could settle after every launch, since transfers are free.


Online Dao Angkan

  • Full Member
  • **
  • Posts: 219
  • Liked: 64
  • Likes Given: 36
Re: A bet on whether re-usable is really cheaper
« Reply #82 on: 05/10/2018 07:33 PM »
According to Musk;

Quote
Right now, flight-proven F9s are priced at $50 million USD.  New boosters are $60 million USD.

Offline Kabloona

  • Senior Member
  • *****
  • Posts: 4303
  • Velocitas Eradico
  • Fortress of Solitude
  • Liked: 2533
  • Likes Given: 524
Re: A bet on whether re-usable is really cheaper
« Reply #83 on: 05/10/2018 09:05 PM »
Also, Musk says he thinks 10 flights of a single booster are possible next year:

Quote
Musk: The first #Falcon9 Block 5 to achieve 10 flights will probably happen next year. "I think that's really a key milestone," he says

https://twitter.com/EmreKelly/status/994653260547088384

Of course, that doesn't cover what happens with all the other boosters. Several others could have multiple flights, though I think his forecast that any of them will reach 10 flights next year is overly optimistic.
« Last Edit: 05/10/2018 09:13 PM by Kabloona »

Re: A bet on whether re-usable is really cheaper
« Reply #84 on: 05/13/2018 09:15 PM »
Didn't Shotwell say that refurbishing the booster on the very first re-flight cost less than half of a new booster?
Does this account for the loss in payload mass from reusability? 
Considering the wide range of payloads and the length of their launch manifest, it doesn't appear to me that SpaceX is lacking in that respect. Besides, there's always Falcon Heavy for larger payloads.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6719
  • Germany
  • Liked: 1760
  • Likes Given: 1736
Re: A bet on whether re-usable is really cheaper
« Reply #85 on: 05/14/2018 05:43 AM »
Didn't Shotwell say that refurbishing the booster on the very first re-flight cost less than half of a new booster?
Does this account for the loss in payload mass from reusability? 
Considering the wide range of payloads and the length of their launch manifest, it doesn't appear to me that SpaceX is lacking in that respect. Besides, there's always Falcon Heavy for larger payloads.

The not so hidden assumption there is SpaceX could have built the rocket much cheaper if they had not given it the extra performance. Under this assumption you can add this hypothetical difference to the reusable rocket cost and conclude it is not worth it.

Re: A bet on whether re-usable is really cheaper
« Reply #86 on: 05/14/2018 09:47 AM »
To me there isn't much to debate, when we know the launch cost breakdown from Elon Musk himself.

From the Block 5 conference call:
Quote
We achieved the primary boost stage, that's like, half the cost of the rocket. On the order of that. [To be technically safe?], the marginal cost of launch, the boost stage is probably close to 60% of the cost. The upper stage is about 20% of the cost. Fairing is about 10%, and then about 10% which is associated with the launch itself.

The George Sowers estimate, on which many base their doubts on reuse effectiveness, assumes only 40% of the launch costs are recovered with the first stage. Elon Musk himself said the percentage is closer to 60%.
This alone makes a huge difference.

Plus, it assumes that recovery and refurbishment amount to 10% of launch costs, in the $5M ballpark. This has to be an average between RTLS and droneship recoveries, and should factor in Block 5 improvements in ease of refurbishment. I wouldn't be surprised if they managed to get to at least half of that with Block 5. So 5%.

If you put these values in the spreadsheet, retaining the other estimates by Sowers (30% payload hit, 5% as the cost of the reusability hardware ecc.) the outlook changes drastically.

Reusability becomes 'effective' by Sowers definition at the 3rd flight already, vs the 10th estimated by him.
With 10 reflights, which is the bottom line Block 5 goal, the net savings in $/kg to orbit amount (accounting for the reusability payload hit, reduced economies of scale)  to a whooping 26%. 2,5x what Sowers projected for Vulcan's SMART after 10 reflights.

Edit: even assuming recovery + ref cost at 10% of the launch costs (to me excessive) the business case closes at the 4th reflight instead of the 3rd, and the savings with 10 reflights are still huge, at 22%. It still beats Vulcan SMART by a big margin.
« Last Edit: 05/14/2018 09:52 AM by AbuSimbel »
Failure is not only an option, it's the only way to learn.
"Tradition is not the worship of ashes, but the custody of fire" - Gustav Mahler

Offline Oli

  • Senior Member
  • *****
  • Posts: 2193
  • Liked: 407
  • Likes Given: 57
Re: A bet on whether re-usable is really cheaper
« Reply #87 on: 05/14/2018 10:42 AM »

An interesting bet would be expendable Falcon 9 vs reusable Falcon Heavy.

Re: A bet on whether re-usable is really cheaper
« Reply #88 on: 05/15/2018 08:06 PM »

An interesting bet would be expendable Falcon 9 vs reusable Falcon Heavy.

Keep in mind that reusability gets more effective the bigger, more expensive the rocket hardware is, if you manage to keep your other launch costs (launch ops, payload integration ecc.) from increasing at the same rate.
This is intuitive: the recovered hardware cost/ total launch cost ratio is THE parameter that makes or breaks reusability, along with recovery and refurbishment cost. If recovered rocket hardware costs increase with its size,  while other mission costs remain low, your savings in $/kg to orbit increase.

With Falcon Heavy you recover a much bigger fraction of launch costs: the recoverable hardware cost increases a lot more than other launch expenses. The second stage, fairing stay the same; payload integration costs are the same ecc. The only added 'non recoverable' cost is due to the added complexity of dealing with three cores in transport and integration.   Furthermore, recovery + refurbishment costs probably increase less than overall launch costs, so as a fraction of the total launch costs they don't change that much vs F9.

On the other hand, with FH side boosters RTLS + centre core on droneship you have a bigger payload hit vs expendable than F9 with droneship landing, but this is balanced by the already mentioned considerations.


It's difficult to make estimates with the limited amount of info available, but I'll try.

A MORE UP TO DATE GUESTIMATE ON F9/H REUSE ECONOMY
We know that F9 expendable price is in the $80-90M ballpark, so with margins this would put expendable launch costs in the $60-70M range.
They've been selling recoverable missions at $62M, with boosters that could only be reflown once, even when they didn't know how reusability would pan out. Of the 21 F9 FT cores sold for $62M as recoverable, 3 out of 19 have been recovered but won't be reused, just 2 haven't been recovered. 16 have been reused once or will be reused once.

So, assuming they aren't/weren't selling launches at loss, I'm going to assume the $62M price matches their expendable costs and they're selling at cost at worst, actually making some profit thanks to the 16/21 Full Thrust, pre-Block 5 cores they've reflown (or will refly once).

Note: their 1st stage cost for the second launch is only driven by recovery and refurbishment costs (which, as we know, have been less than 1/2 of a new booster since the first reflight, probably decreasing with time). In the pre-block 5 era they've been offering a very limited discount for 'flight-proven' cores, rumored to be 10%, or ~$6M compared to the 20-25 millions they cut  from their costs over two launches with just a single reflight. That's why I think they could profit from the $62M recoverable launches, even if probably sold almost at cost.



So, if F9 expendable launch cost is in the $62M ballpark, given the recent  cost breakdown by Elon the actual figures would be:
- $37M for the Booster;
- $13M for the second stage;
- $6M for the fairings (a figure tossed around several times, which makes this guesstimate more solid);
- $6M for other launch costs (launch ops, payload integration ecc.)
- + additional services.

Using these assumptions, a FH expendable launch would be:
- $111M for the boosters;
- $13M for the 2nd stage;
- $6M for the fairings;
- $10M for other launch costs (a total guesstimate to account for more complex ops);

So $140M in total (maybe a bit less if I overestimated launch costs), with <10% profit margins given how they sell it for $150M. So my cost estimates seem plausible, maybe even too high.

However, the low profit margins both for FH expendable and F9 recoverable $62M price are probably mitigated by the additional services sold on the average flight.

Going by this, if SpaceX recovers 60% of the launch costs with a F9 booster, a triple core FH landing accounts for almost 80% of the overall cost.

On the Sowers spreadsheet I'm going to assume:
-40% payload hit with boosters RTLS + centre core droneship;
-80% of the expendable launch cost recovered;
-10% of the expendable launch costs (or $14M) as the cost of the added hardware for reusability;
-5% of the expendable launch costs ($7M) as the overall recovery + ref costs for the three cores ( a bit less in absolute terms than 3x what I assume for F9).

The graph below is the result: reusability is a lot more effective on FH than F9, with a lot more of the total launch cost recovered, while refurbishment costs don't increase that much relatively to the overall cost increase.

Standing by this model, at 10 reuses per core (and for the side boosters not necessarily all ten on FH), you have a whooping 40% reduction in $/kg to orbit vs expendable with Falcon Heavy.
This means that a FH B5 reusable (all 3 cores) could cost SpaceX as low as $50M to launch if they met the BOTTOM LINE of their Block 5 design goal, or 10 reuses per core.

They currently offer a reusable FH B5 for $90M, which is very conservative (actually higher $/kg to LEO than expendable) and probably oriented at profit.

Using dr. Sowers' model, my cost estimates for F9 B5 are astounding too:
with B5, if they achieve the bottom line goal of 10 reflights with an average refurbishment + recovery cost of $2,5M their launch cost would be as low as $32M for F9.

I see them offering F9 flights for $40M in 2020 (or even earlier depending on B5 performance this year and in 2019), several millions less if they manage to do >10 reuses with low ref costs, maybe even less depending on how things go with fairing recovery.

For FH they could go as low as ~$50-60M depending on the demand and on B5 reuse performance.

Final note: I tried putting BFR numbers on the Sowers spreadsheet for fun... let's just say big SHLVs get very convenient with full reuse and it bodes really well for widespread human spaceflight.
« Last Edit: 05/16/2018 03:34 PM by AbuSimbel »
Failure is not only an option, it's the only way to learn.
"Tradition is not the worship of ashes, but the custody of fire" - Gustav Mahler

Online AncientU

  • Senior Member
  • *****
  • Posts: 6207
  • Liked: 3988
  • Likes Given: 5505
Re: A bet on whether re-usable is really cheaper
« Reply #89 on: 05/15/2018 08:38 PM »
Can you set the y-axis to something more intuitive than 'reuse index'?   (Maybe $M/tonne to LEO, or base launch cost for vehicle configuration, etc.)

It is difficult to see why Vulcan would start lower than F9 by any rational measurement, indicating the 'index' likely is a made-up parameter designed to improve Vulcan's visuals.
"If we shared everything [we are working on] people would think we are insane!"
-- SpaceX friend of mlindner

Re: A bet on whether re-usable is really cheaper
« Reply #90 on: 05/15/2018 08:54 PM »
Can you set the y-axis to something more intuitive than 'reuse index'?   (Maybe $M/tonne to LEO, or base launch cost for vehicle configuration, etc.)

It is difficult to see why Vulcan would start lower than F9 by any rational measurement, indicating the 'index' likely is a made-up parameter designed to improve Vulcan's visuals.

Yep, I'll do another graph. Reuse index is $/kg reusable/expendable: if <1 reuse is 'worth it' in $/kg, which I agree isn't always the best metric.

Edit: here we go, a launch cost vs. number of uses graph. Keep in mind that the cost at 1 flight is > than the guesstimated expendable launch cost to account for reusability hardware (legs, grid fins ecc.).
« Last Edit: 05/15/2018 09:28 PM by AbuSimbel »
Failure is not only an option, it's the only way to learn.
"Tradition is not the worship of ashes, but the custody of fire" - Gustav Mahler

Online AncientU

  • Senior Member
  • *****
  • Posts: 6207
  • Liked: 3988
  • Likes Given: 5505
Re: A bet on whether re-usable is really cheaper
« Reply #91 on: 05/15/2018 11:32 PM »
Thanks.
Shows that they could charge around $40M for F9 or $70M for FH and still make buckets.
No fairing  or second stage reuse included yet.
"If we shared everything [we are working on] people would think we are insane!"
-- SpaceX friend of mlindner

Offline joek

  • Senior Member
  • *****
  • Posts: 2922
  • Liked: 681
  • Likes Given: 364
Re: A bet on whether re-usable is really cheaper
« Reply #92 on: 05/16/2018 12:05 AM »
Yep, I'll do another graph. Reuse index is $/kg reusable/expendable: if <1 reuse is 'worth it' in $/kg, which I agree isn't always the best metric.

Those plots look very familiar, as do the arguments pro and con.  The original Sowers model was flawed on several levels--among the most egregious was the use of reuse index without comparable actual costs.[1]

Are we really going to rehash this (I still have all the data/plots from that original thread if anyone is interested--or you can just go read that old thread)?


[1] I can make the Sowers model look really good if I have an LV which is 2x the cost/price of the competition due to very expensive engines.  Which makes Dr. Sower's model look good but sucks for whoever has to pay the actual launch cost.

Offline bassnfool2

  • Member
  • Posts: 8
  • Liked: 3
  • Likes Given: 0
Re: A bet on whether re-usable is really cheaper
« Reply #93 on: 05/16/2018 12:18 AM »
No one has taken the bet yet... People's doubts about how cost effective 1st stage reuse is must not be as strongly felt as argued.

Sent from my Nexus 6P using Tapatalk


Offline joek

  • Senior Member
  • *****
  • Posts: 2922
  • Liked: 681
  • Likes Given: 364
Re: A bet on whether re-usable is really cheaper
« Reply #94 on: 05/16/2018 12:40 AM »
No one has taken the bet yet...
SpaceX has taken that bet and has doubled down on it.

Offline Stan-1967

  • Full Member
  • ****
  • Posts: 537
  • Denver, Colorado
  • Liked: 282
  • Likes Given: 188
Re: A bet on whether re-usable is really cheaper
« Reply #95 on: 05/16/2018 01:29 AM »


So, assuming they aren't/weren't selling launches at loss, I'm going to assume the $62M price matches their expendable costs and they're selling at cost at worst, actually making some profit thanks to the 16/21 Full Thrust, pre-Block 5 cores they've reflown (or will refly once).


So, if F9 expendable launch cost is in the $62M ballpark, given the recent  cost breakdown by Elon the actual figures would be:
- $37M for the Booster;
- $13M for the second stage;
- $6M for the fairings (a figure tossed around several times, which makes this guesstimate more solid);
- $6M for other launch costs (launch ops, payload integration ecc.)
- + additional services.

Using these assumptions, a FH expendable launch would be:
- $111M for the boosters;
- $13M for the 2nd stage;
- $6M for the fairings;
- $10M for other launch costs (a total guesstimate to account for more complex ops);

So $140M in total (maybe a bit less if I overestimated launch costs), with <10% profit margins given how they sell it for $150M. So my cost estimates seem plausible, maybe even too high.


I think the graph would benefit from a separate "y" axis for gross margin %.   If the S1 & fairing are re-used, it seems plausible that GM% could increase as the price floor drops to $30M per launch.   However this would also result in erosion of GM$ gained from each flight.  This is likely why current re-use price ( not cost) is being maintained at a lower marginal discount for re-use.   As long as the market demand is inelastic & you have R&D cost to recover, your business does not want to drop revenue per flight.  Additionally, if SpaceX were to shrink the market size ( in $$$) by too aggressive price reduction, they may be vulnerable financially to bigger companies if they ever go public.

The interesting part to me is that nobody knows what a vertically integrated re-usable launch provider should look like on a balance sheet.  Economics says that SpaceX should be shedding jobs if it was just a re-usable rocket company, but that does not seem to be the case.   If SpaceX were to size & price themselves as low as possible, without the BFS/BFR entanglements, I suspect they could operate a profitable ( 35%-40% GM) business at $15-20M per launch with F9R

Offline bassnfool2

  • Member
  • Posts: 8
  • Liked: 3
  • Likes Given: 0
Re: A bet on whether re-usable is really cheaper
« Reply #96 on: 05/16/2018 01:56 AM »
No one has taken the bet yet...
SpaceX has taken that bet and has doubled down on it.
It's obvoius to me that reusing 1st stages is cost effective for SpaceX... The evidence is that they keep doing it. Your right, they have placed their bet and I believe they are already counting their winnings.

Since the bet was offered by a pro-reuseability poster who has taken the position that reuse is cost effective... Someone taking the bet would be betting against reuseability.  No one here who argues that reusing falcon 9 1st stages is cost ineffective has been willing to put their money where their mouth is.



Sent from my Nexus 6P using Tapatalk


Offline joek

  • Senior Member
  • *****
  • Posts: 2922
  • Liked: 681
  • Likes Given: 364
Re: A bet on whether re-usable is really cheaper
« Reply #97 on: 05/16/2018 02:21 AM »
... No one here who argues that reusing falcon 9 1st stages is cost ineffective has been willing to put their money where their mouth is.

Nor would they at this point, because it's a fool's bet.  But please let's skip the rehash of ULA's SMART.  That's noise which we dispensed with long ago; nothing new presented in this discussion.

Re: A bet on whether re-usable is really cheaper
« Reply #98 on: 05/16/2018 08:35 AM »
Yep, I'll do another graph. Reuse index is $/kg reusable/expendable: if <1 reuse is 'worth it' in $/kg, which I agree isn't always the best metric.

Those plots look very familiar, as do the arguments pro and con.  The original Sowers model was flawed on several levels--among the most egregious was the use of reuse index without comparable actual costs.[1]

Are we really going to rehash this (I still have all the data/plots from that original thread if anyone is interested--or you can just go read that old thread)?


[1] I can make the Sowers model look really good if I have an LV which is 2x the cost/price of the competition due to very expensive engines.  Which makes Dr. Sower's model look good but sucks for whoever has to pay the actual launch cost.

I'm going to read your posts on that thread, but to me what you say here doesn't make the model 'flawed'.
It shows a feature of reusability that is intuitive and true: it's more beneficial for expensive systems.
That's why the 'added cost for reusability' doesn't really matter, and why you can make big rockets really affordable if you nail reusability. With it the initial hardware cost matters way less.

I know that reusability being cheaper is a no-brainer, this was an attempt to roughly see how much for SpaceX's  rockets, using a model which may be conservative but whose basic math is solid, or at least I haven't seen anyone disprove it.
« Last Edit: 05/16/2018 09:11 AM by AbuSimbel »
Failure is not only an option, it's the only way to learn.
"Tradition is not the worship of ashes, but the custody of fire" - Gustav Mahler

Offline meberbs

  • Full Member
  • ****
  • Posts: 1705
  • Liked: 1534
  • Likes Given: 376
Re: A bet on whether re-usable is really cheaper
« Reply #99 on: 05/18/2018 06:23 AM »
I want to take the bet. But how do you send me the money? I am afraid you're just being rhetorical and hypocritical.

I can bet you any money that other company leaders want to take your bet too.
If reuse really save money. Why nobody is rushing to follow suit? Why nobody is filing for bankruptcy yet?

Any normal person can see the point here, that there is no evidence that reuse save significant money. Except those r* fan*s.
If you bother to read the thread, you'll see that the OP already suggested using paypal for the fund transfers. He is not being rhetorical. I'd be offering the same bet since it is basically free money, but it would be unfair to steal his idea.

Before taking this bet you might want to actually do a minimal amount of research. ULA and Arianespace are both pursuing reuse, although  in a more limited form. Blue Origin and China are both going for SpaceX style reuse. That only leaves Orbital, Russia and NASA as major players without reuse plans. I don't think Russia really could support the development with their current space budget, and NASA is stuck doing whatever congress says, which usually has nothing to do with cost effectiveness.

Space launches are booked years in advance, while the first re-flight of a booster was only a year ago. There hasn't been time for this to significantly impact competitors yet.

Tags: