Talk:Space Launch System/Archive 4

RfC for SLS Launch Cost
Should the "cost per launch" figure in the article infobox be changed? – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:09, 7 July 2020 (UTC)
 * Small note to any readers: discussion on this topic also includes the next section as well as the three previous sections, starting here Leijurv (talk) 17:12, 21 July 2020 (UTC)
 * Pinging . Would like to lay this issue to rest for good. I personally am satisfied with the current compromise, but would like to get a broader subset of the community involved. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:24, 7 July 2020 (UTC)
 * The cost estimates for the SLS cannot be summarized in an infobox. The best we can do is show the large range of available estimates, and hope that the interested reader will dive into the details and make up their own minds. Eggsaladsandwich (talk) 17:26, 7 July 2020 (UTC)
 * I totally agree, this is why the “At an estimated cost of over $2 billion per launch for the SLS once development is complete" is the most accurate. Because while we don't know how much an SLS launch will cost, we damn sure know that it's gonna be more than $2 billions!
 * I don't see why we can't just put the $900 million as "Marginal launch cost", which is what it is! It's more precise and doesn't confuse the two concepts (cost and marginal cost) which are two different things! I'm ready to even accept this confusion (because I'm sick and tired of this nonsensical debate), but I'm sure some one else in less than a month is gonna bring it up again because it's such an evident issue!
 * The $500 million figure is just "utter fiction" as N2e put it. I don't think there is even a debate here! This figure was the target price when SLS started but is not even in the ballpark of today's estimates. Pinging Moamem (talk) 18:17, 7 July 2020 (UTC)
 * That's a wise thought Eggsaladsandwich. Perhaps the biggest problem is that infoboxes should, per wiki guideline I believe, restate info that is already well explicated, and well sourced, in the prose in the article body.  It seems that this article, as it currently stands today, is trying to do all the heavy lifting in the Infobox itself.


 * If someone doesn't get to it first, I'll try to write up a few encyclopedic statements of prose that are supported by the various cost per launch citations I'm seeing given in the infobox. Will do it such that each statement says no more than the source supports, with adequate context, without undue emphasis on any one of the many cost estimates that exist, maintaining neutral point of view. Then, the infobox can simply summarize the range of launch costs that is supported in the article prose.  N2e (talk) 18:13, 7 July 2020 (UTC)
 * This confusion is not innocent some people with vested interest are trying to make this program look good no matter the truth and this is not limited to Wikipedia. IMO the best way to dissipate any confusion is to separate the different figures and be as precise as possible :
 * Launch cost : Over $2 billion once development is complete (est.)
 * Marginal launch cost : $900 million (est.)
 * Launch cost target : $500 million per launch (objective in 2011)
 * We could lose N°3 IMO because it's not in current plans. If we don't want to leave any confusion that's the way to do it.
 * I've see the exact false figure "$500 to $2 billion" cited in serious publications. This stuff drive policies and should not be allowed to stand just to please one user or another. Moamem (talk) 19:12, 7 July 2020 (UTC)
 * Thank you for your kind words and your offer to do the heavy lifting on this. I skimmed the references, and some of them get into complexities about "Real Year" costs, projected inflation, etc. Trying to come up with a summary was more work than I was willing to do. Eggsaladsandwich (talk) 19:18, 7 July 2020 (UTC)


 * Keeping the range in the infobox and having a section that goes into deeper detail about the various estimates sounds like a good idea. I support it. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 21:16, 7 July 2020 (UTC)


 * Support keeping the range per Jadebenn. This is what I want in the cpl parameter of the infobox: "$900 million–$2 billion (est.)". --Soumya-8974 talk contribs subpages 05:46, 8 July 2020 (UTC)
 * One of the decadals cites $500M as the cost and the other cites $750M. I really think we shouldn't be picking and choosing. Defeats the whole point of it being a range if we do. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:39, 8 July 2020 (UTC)
 * One cites $500 million with nothing to back it up, the other cites $650 million (not 750) saying it's an "assumption". We'e not picking and choosing, not all sources are created equal that's all. I can find sources citing over $4 billion, should we include those too? Moamem (talk) 17:05, 8 July 2020 (UTC)
 * If "not all sources are created equal" then your $2B wouldn't be there either, because the only things that claim it are a policy document and a journalist who explicitly identifies as an opponent of SLS. It's a package deal, mate. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 22:56, 8 July 2020 (UTC)
 * A policy document? It's the OMB warning congress that their funding is not going to be sufficent. May I remind everyone that "The Office of Management and Budget (OMB) is the largest office within the Executive Office of the President of the United States (EOP). OMB's most prominent function is to produce the President's Budget, but OMB also measures the quality of agency programs, policies, and procedures to see if they comply with the president's policies and coordinates inter-agency policy initiatives." That's the office overseeing pretty much all the executive branches budgets and programs! It's a neutral party! It is by far the strongest source of the bunch! Besides the OIG said that NASA would save $1.5 billions if they choose to launch Europa Clipper on Falcon Heavy instead of SLS. How is it possible to save that much if SLS cost $500 million or even 900!??? Please square that circle. Moamem (talk) 04:11, 9 July 2020 (UTC)
 * Please do not bring you irrational hatred of Eric Berger into this conversation. He was just been proven right when you were claiming the opposite.
 * This was also the compromise I proposed. Even tho the $900 million is the "Marginal launch cost" not the "launch cost" and we could just add this figure separately for more clarity...Moamem (talk) 17:05, 8 July 2020 (UTC)
 * That was how things were before. Then you insisted that the entire yearly program costs needed to be included (and you claimed that was "launch cost" because there just happens to be one launch in a year), and a compromise was put forward to placate your concerns. I also strongly recommend you review WP:USTHEM before commenting further. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 01:48, 9 July 2020 (UTC)
 * Finally! Thank you for admitting that you were talking about Marginal cost all this time, you could just have said so in the beginning and could have had a whole different conversation but I guess muddying the waters was a feature not a bug! I'm creating a new section debating whatever of not "launch cost" should be equal to "marginal launch cost"
 * Soumya-8974 is definitely not suggesting to go back to the figures you forced despite now 8 people saying otherwise. All that he said is that the $500 million is fiction. that's it! Moamem (talk) 04:10, 9 July 2020 (UTC)
 * I am male, not female. Please use "he" or "they" to refer to me. --Soumya-8974 talk contribs subpages 04:47, 9 July 2020 (UTC)
 * Soumya-8974 One of my female friend has the same name Soumya, so I just supposed... Sorry. Corrected.Moamem (talk) 05:04, 9 July 2020 (UTC)

Pinging participants from previous discussions. Sorry about this, but made it clear he was not satisfied with the current compromise, so it appears we need to reopen the issue. I know this discussion is a bit of a mess, but please weigh in. Nothing's going to get done if it's just the two of us going back and forth over and over. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:45, 9 July 2020 (UTC)

As stated, moving this back to the RfC. While I can answer your question, I believe you're still missing the point. The compromise we arrived at gave us both figures we disagreed with. The $500M figure is just as well-sourced as your $2B one, and the report it's in explicitly calls it launch cost. You may feel that there is some sort of significant difference between the so-called "marginal launch cost" and proper launch cost, but that's just it, your point of view. Wikipedia is explicitly supposed to advocate a neutral point of view, see WP:NPOV. Rehashing the debate we had a month ago is fruitless, as we covered these same points before. I believe it'd be more productive to accept the solution are offering, which is to have a section that breaks down how each is calculated, and keep the existing range in the infobox. You may personally feel that the $500M figure is unrealistic, just as I feel the $2B figure is, but that's the point of compromise. If there are reputable sources for both, and there's no consensus on which to pick and choose, than the only option is to show them all. And yes, that means that if you could find reputable sources for an even higher high-end estimate, it would go there too. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 17:27, 9 July 2020 (UTC)
 * Agree. The budgetary arcana around the different kinds of launch costs cannot be meaningfully explained in a single infobox line. The budget section of Space Shuttle entry has more than 400 words in it, and "(t)he exact breakdown into non-recurring and recurring costs is not available..." even though that program ended 9 years ago. We should move on from this arid discussion about the infobox and get to arguing interesting questions like should Shuttle development costs be included in the SLS development cost. Eggsaladsandwich (talk) 18:36, 9 July 2020 (UTC)
 * I don't see what's complicated about adding "Marginal Launch cost" in the info box unless there is an agenda to make this program seem less expensive that what it is!
 * The $500 million is a different issue since that number is obviously false as 8 different people have pointed out in less than 9 months (and no $500m and $2b figures are not equally unrealistic, I can give you a detail breakdown of this figure. Can you get close to doing the same?). I unwillingly accept to keep this "fictional" $500 million figure if we can add the $4 billion figure that is actually quite realistic : https://arstechnica.com/science/2019/11/nasa-does-not-deny-the-over-2-billion-cost-of-a-single-sls-launch/ Moamem (talk) 22:06, 9 July 2020 (UTC)


 * The $500 million and $2 billion figure are equally unrealistic. With (optimistically) 10 launches over the next decade and program costs being what they are, $4 billion is a more realistic number.  If we include Constellation program costs then that number can't be much lower than $5 billion.  If by divine intervention they double that number of launches without increasing costs then $2 billion becomes the lowest possible cost per launch.  This is mostly Boeing we are talking about so things going that well is simply never going to happen.SandowTheHeretic (talk) 13:19, 14 July 2020 (UTC)


 * Therefore, we should use "$500 million–$4 billion (est.)" on the infobox, right? --Soumya-8974 talk contribs subpages 09:05, 12 July 2020 (UTC)
 * If there's a credible source that has $4B as the launch cost (regardless of how dubious I personally find that figure), I guess so. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 17:06, 12 July 2020 (UTC)
 * I gave a source in the line above That's what Soumya-8974 was referencing - Moamem (talk) 17:32, 12 July 2020 (UTC)

Objectively SLS is going to cost a lot more than 4 billion. But nasa can't say so. Zegfred (talk) 19:20, 12 July 2020 (UTC)

My take is this: The ONLY relevant launch cost number for NASA, Congress, the American Taxpayers and indeed all other relevant partners is the total launch cost, that is the total program cost divided by the number of launches. This is the case since it is the cost that NASA chose to incur INSTEAD of deciding to launch their satellites and humans on commercial launch vehicles. So far, the SLS has cost 20,3 billion dollars with 0 launches. We will only know the cost per launch when the program finishes, but we can make fairly good predictions (+- 30%) as of now. If SLS funding continues until 2030 @ 2B a year with a total of 10 launches, the cost per launch will be 4 billion dollars. Fewer launches mean more per launch, more launches mean less per launch, but it is very safe to say that the cost per launch will be between 3,5B and 5,5B. TheSkalman (talk) 18:04, 12 July 2020 (UTC)
 * Hey man. Finally some fresh meat! While I personally agree with you that the launch cost should include all costs. We agreed that since the final dev costs and the number of launches would not be known until the end of the program we will exclude them in our launch cost calculations!
 * In my opinion :
 * The launch cost of an SLS rocket is “Over $2 billion once development is complete” as the White House OMB (a neutral party overseeing almost all budgets and is therefore the best source provided) has said and that we already knew for some time just doing basic arithmetic.
 * The $900 million figure is the Marginal Cost which is a totally different thing and is completely ignoring the fixed costs which are most of the costs associates with government programs in general. Should be discarded or explicitly described as Marginal Launch Cost.
 * The $500 million is the aspirational figure that was touted at the beginning of the program but we know today is a complete fiction. Should not even be debated, the engines alone cost $400 million for god’s sake!
 * Which I think are all reasonable opinions but Jadebenn thinks that marginal cost is the same as actual launch cost and that $500 million per launch is an acceptable figure because some source said so and all source being created equal we can not qualify the reliability of a specific source and its biases.
 * So what is your opinion on the launch cost excluding Dev costs and the use of marginal cost as the SLS launch cost? - Moamem (talk) 18:42, 12 July 2020 (UTC)

If you want to cite the marginal cost because it makes SLS seem cheaper just call it marginal cost. but it's not the launch cost. Zegfred (talk) 19:20, 12 July 2020 (UTC)

Yes, the launch cost should be changed. Especially having a number of $500 million in there just makes no sense, considering that one RS-25 engine alone costs $146 million (https://arstechnica.com/science/2020/05/nasa-will-pay-a-staggering-146-million-for-each-sls-rocket-engine/), and the SLS needs 4 of them, leading to an engine cost of $584 million. Having this Wikipedia page say that SLS might be as cheap as $500 million per launch is just incorrect.

In general, I have to say that this is a discussion that shows shortcomings of how Wikipedia works. If I come to Wikipedia, I want to get neutral information, not information that is cherry picked by people who try to make something look better than it really is. It seems one user here (Jadebenn) has a lot of influence on preventing a higher, more correct launch cost from being mentioned. That user is not neutral though. He is the most active admin of the SLS group on Reddit (https://www.reddit.com/r/SpaceLaunchSystem/), where he only allows positive articles to be posted, while even light criticism is only allowed in a special "Paintball" thread meant for criticism, and some of the people who often express criticism about certain aspects of SLS he just banned. The SLS subreddit has over 6000 subscribers, and when reading through the comments, it looks like a lot of the people active there are employed working on SLS, and that obviously gives them a strong incentive to not look at SLS neutrally. I don't know if Jadebenn himself also is employed working on SLS, but he seems to have some strong incentive to spend a lot of time on the internet with trying to suppress criticism about SLS. Someone like him should not have any influence on the Wikipedia page about SLS, it's not good that he currently seems to have a lot of influence here. Skytie (talk) 10:22, 13 July 2020 (UTC)
 * There is a fairly logical conclusion about Jadebenn. He is most likely a paid publicist working for Boeing or a related contractor.  It is easy to look back at this page's history and see that he treats curating this page as his job.  Probably because it is.  His attempts to badger people into leaving his domain alone by using his interpretation of wikipedia rules amount to schoolyard bullying. Yes, he should be banned.  I don't know that it is likely to change anything since he is likely to simply create a new account and continue his behavior.SandowTheHeretic (talk) 13:35, 14 July 2020 (UTC)
 * I wish I got paid for this. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 17:30, 14 July 2020 (UTC)

After weeks of debate Jadebenn finally admitted that the figures he's presenting are the marginal cost while I was always talking about the "normal" cost :

I think this is the gist of the disagreement, and solving this would definitely close this issue. So first let's get some definitions :

So basically it's the cost of building one more unit in an already functional production line. This completely ignores fixed costs and any savings provided by economies of scale. Basically if you produce a billion units or one the marginal cost is the same.

I gave an example earlier : If Banana makes 10 Jphones at $1000000 apiece and Apple makes a million Iphones at $1000 apiece. You do not get to say that they cost the same because they have the same marginal cost! No one thinks like that for any other thing!

So bringing it back to the matter at hand, what Jadebenn proposes is not only ignore all the fixed costs of an SLS launch (Ginormous NASA personnel, Installation, maintenance, support...) but he also want us to treat a low volume (well one!) production as if it was mass produced!

Talking about Falcon 9 and Atlas V he said:

He thinks that it's somehow unfair for SLS to be compared apple to apple to commercial rockets (or any other thing ever produced by the way) because it does not benefit from economies of scale. therefore any cost advantage given by mass production should be totally ignored. This is wrong on so many levels that I don't think it's necessary for me to explain them.

For me and for everyone IMO cost is cost, and that include fixed costs... It just seems evident...

Pinging Soumya-8974, Eggsaladsandwich, Jadebenn, N2e - Moamem (talk) 05:01, 9 July 2020 (UTC)
 * There's an RfC currently open where any of your concerns should be placed, preferably while assuming good faith and without personal attacks. I would once again recommend you please review WP:USTHEM and stop seeing this as a "battle" between you and me. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:04, 9 July 2020 (UTC)
 * In addition, you really shouldn't be dragging in off-site statements I've made, especially ones taken out of context. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:39, 9 July 2020 (UTC)
 * Please stop patronizing me. There are no personal attacks in the comment above. All the quotes I gave are from this exact page, where do you see off-site comments? Please address the arguments instead of me of the form they take.
 * I created a new section because YOU admitting to using marginal cost and not real cost changes this debate from one about sources and figures to one about semantics. It's way easier to debate semantics! - Moamem (talk) 14:31, 9 July 2020 (UTC)
 * I am sorry you see it as patronizing, but it feels like you're unfamiliar with the code of conduct, and I'm trying to point out what I see as consistently uncivil behavior.
 * As I've made abundantly clear before: I see no difference between "marginal launch cost" and launch cost. To me, launch cost is what NASA pays to add another flight to the launch manifest. But you're missing the point, which is that we both agreed to do a range of estimates, as is done on the Space Shuttle article, since we were unable to come to an agreement on this matter.
 * I will not comment on this matter further outside of the RfC, where this discussion belongs. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 16:03, 9 July 2020 (UTC)

Let me highlight this quote of yours:

No you were not "abundantly clear". But thanks for finally doing so. The talk about estimates and references is irrelevant if they estimate and reference different things. So I won't be participating in the discussion about those until we come to an agreement about what we're actually talking about. Marginal Launch Cost? Launch Cost? Any of them? My opinion is that they could not be more different from each other : For example if the SLS program excluding development cost $2 billion a year for one launch a year. And the marginal cost of adding one more launch is $900 million. What is the launch cost if you launch once a year? Twice? N times?
 * 1) Marginal cost is the cost of making one more unit.
 * 2) Cost is Fixed cost + variable cost / number of units
 * 3) No one thinks that cost do not include fixed costs. Does anyone think that besides you?
 * 4) Marginal cost could be irreverent if we only launch once a year which seems to be the case for SLS at least for the next decade!
 * For me the cost of launching once a year is $2 billion. Twice is 2+0.9/2 = $1.45 billions. N times is (2+(N-1)*0.9)/N billion dollars.
 * For you, it cost as much to launch SLS once a year than a million, which in my opinion is ludicrous. In your world view economies of scale do not exist or at least are not related to cost at all. - Moamem (talk) 16:51, 9 July 2020 (UTC)


 * As previously stated, further discussion should take place in the RfC. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 17:08, 9 July 2020 (UTC)

Yes absolutely. 500 million is sifi. marginal cost is nothing like total cost. 2 billion a launch is probably a nasa lowball estimate. I will read the whole post and give my full opinion later,Zegfred (talk) 19:13, 12 July 2020 (UTC)

Ok, as an outsider, it looks like the debate is around the meaning of cost and how to ensure that is relevant to various scenarios.

The reality is SLS is run by NASA so cost means something different there than if they are purchasing the service from a third party. However, you may want to compare those costs to what a third party charges.

In my opinion, both marginal and total cost should be used. Why? Marginal cost gives insight into costs if the program were scaled up to 2, 3, 4 or more launches per year. Total cost gives insight into what it cost to provide the service and capability.

The way I have seen this done is to map out planned launches by year, break out fixed and per launch costs and give a total planned cost for that year.

I think the goal here should be to provide the maximum amount of information and trust the consumer of that information to parse it as appropriate. Timlograsso (talk) 20:47, 12 July 2020 (UTC)


 * Hi, and thank you for your input. I see no issue with citing Marginal launch cost along with the normal launch cost. But these to numbers should be separated. What Jadebenn is doing is trying to equate these 2 numbers like if they were the same thing. If we merge them (jadebenn considers them to actually be the EXACT SAME THING) it's not being exhaustive but being confusing (on purpose IMO). Do you agree that while both figures should be cited, they should be in 2 different entries? - Moamem (talk) 22:16, 12 July 2020 (UTC)

'''Soumya-8974, Eggsaladsandwich, Jadebenn, N2e, TheSkalman, Timlograsso, Zegfred, I think we're reaching a consensus here : keep both Launch cost and Marginal launch cost but on 2 separate lines. Something like the following :''' "Launch Cost : Over $2 billion once development is complete (est.)"

"Marginal Launch cost : $900 million (est.)"

How does that sound? - Moamem (talk) 22:25, 12 July 2020 (UTC)
 * That sounds like a good idea to me. The marginal launch cost is worth mentioning, since SLS may (or may not) be used of the Europa Clipper mission and has been suggested for other planetary missions. For those applications, the total launch cost (fixed per year cost plus per flight cost) isn't relevant, since NASA is making the decision to pay the fixed costs for other reasons. For the planetary missions, it really is a "since human spaceflight is already paying for the fixed costs, could we have an extra launch for our robotic spacecraft?'' issue. But for human spaceflight and Artemis, the decision to accept the fixed costs was all about those programs, so the total cost is more relevant for them. Giving both numbers covers both of those applications. Fcrary (talk) 23:26, 12 July 2020 (UTC)
 * Please move the discussion to the RfC. This fragmentation complicates things. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 23:30, 12 July 2020 (UTC)
 * Sorry man, I'm not going to bury this in your section. I know that this tactic worked on Reddit, but enough is enough. If you want to participate in this conversation your welcome. Otherwise since editors seem to unanimously agree with this format, I'm going to wait for one last comment and go ahead with the changes. - Moamem (talk) 23:53, 12 July 2020 (UTC)
 * It's not "my" section. An RfC is a dispute-resolution method, added to a list of other RfCs that are accessible to neutral, 3rd-party, editors who wish to participate. I'm sorry you're impatient with how long the process is taking, but RfCs, as a rule, aren't very quick. There's no set time limit - they go on until an agreement is evident to all editors involved. Usually it takes about 30 days in my experience. Sometimes more, sometimes less.
 * Anyway, you can't just circumvent the RfC and claim "consensus" because one or two people said something that's vaguely supportive of your statements. That's bypassing the whole dispute resolution process. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 00:05, 13 July 2020 (UTC)


 * It's not one or two! It's everyone except you, I'm counting a dozen at least, you're the only one defending this position! And they are not "vaguely supporting my position", most are saying to leave the $900m out altogether or that even putting the cost at 4 o 5.5 billions. This is a very consensual position! Now that it's obvious that you're unreasonable or trying to push an agenda you're trying to accuse me of canvassing or even publishing your personal info (like I don't even have any! I don't know you personally!). Sorry but this has been dragging on for months! Enough is enough! The consensus is obvious you do not have veto power here! A single Editor can't block a sensible position (even an optimistic one) that garnered consensus! One more contributor and that's it! Don't worry I'll quote those supporting my position before making the change! You're welcome to do as much.- Moamem (talk) 00:21, 13 July 2020 (UTC)
 * You keep trying to paint me as some rogue contributor, but it is emphatically not just me who has reservations about this. and  both made statements supporting a different solution than the one you're offering, to give an example. You're ignoring them entirely.
 * It's also not acceptable to steamroll through changes because you think something's "taken long enough already." That's tendentious editing, and a good-faith assumption of unfamiliarity with Wikipedia policies and customs can only go so far when you've had at least a month to familiarize yourself with them. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 00:31, 13 July 2020 (UTC)
 * Also, I never once accused you of publishing personal info. I have no idea where you got that from. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 00:40, 13 July 2020 (UTC)
 * Look, it's clear to me that some people involved in this discussion have strong personal feelings about the subject. One is getting very picky and lawyer-like about the process for a RfC. Let's not get that way. First, I don't see a good reason to be so formal about this. Second, if you do have strong personal opinions about the subject, you should drop out of the discussion and recuse yourselves. A strong personal opinion on the subject of an article means you are not unbiased, and should not be involved with editing the article or debating how it should be edited on the talk page. Fcrary (talk) 01:41, 13 July 2020 (UTC)
 * Sorry, but fragmentation discussions are a pain for everyone. I reply to a comment under the section that comment was posted in. That's [i]my[/i] way of dealing with fragmented discussions. If you don't like that, then get other people to stop posting comments in sections you consider inappropriate. Fcrary (talk) 01:33, 13 July 2020 (UTC)
 * Yeah, I also see how it could be useful. I just don't want it to be confusing. Thanks. One more opinion maybe? - Moamem (talk) 23:53, 12 July 2020 (UTC)
 * A suggestion: Is there a way to accurately divide the cost of SLS into these three buckets: Program cost per year, Development cost, Marginal cost per launch? Perhaps with those three values, a reader could get a more accurate sense of what "cost" means for something like this. While three different numbers is a little complicated, with proper copy and citing, it might be the clearest and most neutral way to present it. Currently the three numbers are Project cost, Cost per launch, and Cost per year. That's pretty unclear. For example, on reading that sidebar for the first time just now, I don't know how much of the "Project cost" is the "Cost per year" of previous years, I don't know if the "Cost per year" includes the cost of each launch, I don't know if the "Project cost" is a total to date or a project total for the entire duration of the program, I don't know when the "Cost per year" starts, etc. If we have a "total cost to date" number, let's call a spade a spade and write out those words, instead of a vague "Program cost". Same goes for "total cost (lifetime) (estimated)" and such. Leijurv (talk) 03:35, 13 July 2020 (UTC)
 * I think the best and clearest way to divide the costs would be: 1) Cost to develop the SLS launch vehicle, 2) the fixed cost per year to build and launch the SLS, regardless of the the number launcher per year and 3) the additional (marginal) cost of adding an additional SLS flight. In other words the total cost for the entire SLS program could be written as:
 * Cost = Development_Cost + Annual_Fixed_Cost_per_year * number_of_years + Marginal_Cost_per_launch * number_of_launches. But extracting referenceable numbers for those costs may not be possible. I think the current suggestion is to list Annual_Fixed_Cost_per_year/number_of_launches_per_year + Marginal_Cost_per_launch as the total cost per launch and also list the marginal cost per launch. Fcrary (talk) 05:23, 13 July 2020 (UTC)
 * since which year? When does/did the clock start? Leijurv (talk) 05:25, 13 July 2020 (UTC)
 * I really don't know, since SLS has heritage from the older Ares design work. But I think we're talking about the operational costs. So once SLS has been developed, that's either Annual_Fixed_Cost_per_year/number_of_launches_per_year + Marginal_Cost_per_launch as the total cost per launch and/or the marginal cost per launch. And, yes, I am ignoring things like the development of the new upper stage, which would be in development in parallel with the first SLS Block 1 flights. NASA accounting is, as I said, murky. — Preceding unsigned comment added by Fcrary (talk • contribs)
 * Then could we have,  , and  , no? Is this upper stage a part of SLS, or is it a payload of SLS? I think the current suggestion is to list Annual_Fixed_Cost_per_year/number_of_launches_per_year + Marginal_Cost_per_launch as the total cost per launch Can you explain how this is accurate? Shot in the dark, but is the idea that SLS can only be produced at a certain rate (given that annual fixed cost) therefore it's correct to include it as manufacturing cost? Leijurv (talk) 17:51, 13 July 2020 (UTC)
 * My 2 cents: IMHO it should be done consistently with all other launchers. In which case Marginal Launch Cost would be the one you are looking for (usually it's done this way due to lack of other information, though consistency is the key anyway). Stuff like development costs, fixed costs, infrastructure maintenance, etc. absolutely should not be included in the launch cost field as it isn't in any other launcher. I would find it ridiculous to target SLS specifically with an aim to give it as high value of the field as possible when it obviously departs from the way it's handled everywhere else. SkywalkerPL (talk) 11:16, 13 July 2020 (UTC)
 * Thanks for your input. I'm curious about this assertion that all Launchers have Marginal cost as their launch cost. Which launcher are we talking about exactly? Space Shuttle? Saturn V? Falcon 9? Atlas V? Delta IV? Coz as far as I know NONE of them has Marginal Launch cost as their launch cost, most include dev cost in their launch cost (STS) and some even profits and investments in future products! (commercial launchers) Can you provide some examples of what you mean? Cheers - Moamem (talk) 14:58, 13 July 2020 (UTC)
 * Support After spending a few hours reading through the sources (see my lengthy comment below), I now agree with this. I support Launch Cost of $2 billion (I am neutral on the "Over" and support "once development is complete" as there are no launches currently and there will be more development costs before the first launch), and I support Marginal Launch Cost $900M (my figure of $1B marginal was a rounding, $900M is also a valid rounding to one more significant figure). Leijurv (talk) 01:03, 14 July 2020 (UTC)

I consider myself an SLS fangirl. Jade I don't think you should try to support SLS based on price. Yes the Space Launch System is going to cost around two billions a launch and that's fine. Space is expensive, going to the moon is even more expensive, and that's OK because it is worth it. I don't see an issue with what momem is proposing you just need to highlight the unique capabilities that SLS provides. Until a commercial alternative eventually materializes, SLS is the only concrete vehicle to reach to MoonSLSgal (talk) 22:00, 13 July 2020 (UTC)
 * With all due respect, this isn't a question of being a "fan" or not of SLS. This is a content dispute over a figure in the info box. One that would've probably not escalated so far if appropriate Wikipedia policies had been followed by all users involved. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 00:43, 14 July 2020 (UTC)

Can you please explain why you keep attempting to change the disputed figure while an RfC is in progress? The status quo should be maintained until a course of action is agreed upon. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 00:43, 14 July 2020 (UTC)

High level clarifying question: Regarding this range of costs, what is the situation? In other words, why are we presenting a range? Do we have conflicting sources ranging from 0.5B to 2B? Is this a value that has changed over time? Are these external estimates? Those are the questions I have coming in, as to why we can't find the most accurate number. Looking into the sources now, I see:
 * The NASA Administrator saying I do not agree with the $2B number, it is far less than that. I would also say that the number comes way down when you buy more than one or two. And so I think at the end we're going to be in the $800M to $900M range - I don't know, honestly. We've recently just begun negotiations on what number three through whatever - we don't have to buy any quite frankly, but we intend to. But we're looking at what we could negotiate to get the best price for the American taxpayper, which is my obligation as the head of NASA. I do not see this as clearly supporting either end of this range. I also think that the And so I think at the end can't be discounted here, especially right after a comment on increased volume decreasing cost beyond the first few. It comes across to me as speaking off the cuff / speculating on what the price could be after these negotiations take place. In my opinion, I don't think this should be cited as any form of upper bound, but it is a good source as a lower bound, implying that each SLS costs at least that much today, and likely more.
 * The Europa report in May 2019. This seems quite clear. I see NASA officials estimate the third SLS Block 1 launch vehicle’s marginal cost will be at least $876 million. There is further discussion of alternatives, and the tradeoffs due to SLS offering a faster transit time, but that is irrelevant here. This is, however, 1. an estimate 2. a projection to the third launch 3. a marginal cost. This is fine however, as it would be the cost to the Europa mission of choosing this option (I believe). I could see this being cited, however I would ideally prefer the NASA officials estimate directly.
 * The Habitable Exoplanet report, section 9-11 9.4.1 basis of estimate, (which is page 281) (found here). I see The launch vehicle costs of $650M FY20 ($925M RY) for the SLS Block 1B and $300M FY20 ($429M RY) for the Falcon Heavy were assumed. I also see The SLS block 1B was assumed to cost $650M and the other launch vehicles were assumed to cost $300M based on NASA guidance. Technology development costs were not included in the estimates for this architecture tradespace sensitivity study a little later on page 297. The only entry in Appendix J - References that I could see that might be related was NASA. 2018. Space Launch System (SLS) Mission Planner’s Guide. Exploration Systems Development (ESD). which I believe can be found here, however, I do not see the $650M number there. Regardless, the number appears to be an estimate from 2018. (while the document was initially released in 2017, it states in the header that it was completely rewritten in 2018, and the citation in habex does say 2018). Assuming the source data is from 2018, I believe this source could be outdated compared to some of the newer ones below. Additionally, the executive summary states The Habitable Exoplanet Observatory, or HabEx, has been designed to be the Great Observatory of the 2030s, a successor to the Hubble Space Telescope (HST) with enhanced capabilities and community involvement through a competed and funded Guest Observer (GO) program., from which it appears this is planned to launch a ways in the future. As well as Between now and the expected launch of HabEx in the mid-2030s (page 49), I conclude that this figure is for the far future.
 * The Origins report. The launch cost ($500M for the SLS launch vehicle, as advised by NASA Headquarters) is also included. NASA GSFC’s Resource Analysis Office (RAO) independently estimated the mission cost using different methodology. RAO and CEMA are firewalled from each other, but they both referred to the same MEL and mission schedule. Looking at the bottom of page 13 of the report, I see that launch is scheduled for 2035. I believe this could be another aspirational number for that reason. I also could not find where this advised cost is coming from.
 * An Ars Technica article by Eric Berger. It appears to cite the next source which says NASA Europa Mission. The bill requires that NASA use the Space Launch System (SLS) rocket to launch the Europa Clipper mission. The Administration is deeply concerned that this mandate would slow the lunar exploration program, which requires every SLS rocket available. Unlike the human exploration program, which requires use of the SLS, the Europa mission could be launched by a commercial rocket.  At an estimated cost of over $2 billion per launch for the SLS once development is complete, the use of a commercial launch vehicle would provide over $1.5 billion in cost savings.  The Administration urges the Congress to provide NASA the flexibility called for by the NASA Inspector General and consistent with the FY 2020 Budget request. The Ars Technica article adds on top of this: The White House number appears to include both the "marginal" cost of building a single SLS rocket as well as the "fixed" costs of maintaining a standing army of thousands of employees and hundreds of suppliers across the country. Building a second SLS rocket each year would make the per-unit cost "significantly less."  This was said by Kathryn Hambleton, a spokesperson of NASA, in response to Ars Technica's inquiry. This is interesting to me... why would the launch cost not include the wages of the employees and suppliers who build the rocket? I may be misunderstanding something, but it seems to me as if the statement is "If we could make twice as many rockets employing the same people, the per unit cost would be less". The article itself goes on to describe the development costs to date, and calculates a figure that divides the development cost among the launches of the rocket. I discount this because we already have a separated development cost figure in the infobox, and it seems natural to separate that from operations. So, discounting the article itself and just looking at the appropriations letter and the NASA spokesperson's response, I see that there is some tension between what cost deserves to be considered part of the rocket, and what is a fixed cost. The reason is that (it appears) the NASA employees and suppliers would exist regardless, therefore increasing the cadence of construction would result in a lower unit cost. I am not too convinced by this hypothetical brought up by the spokesperson, and the reason is how far the schedule has slipped already. (from the article itself: The first flight of SLS has slipped multiple times: first to 2019,[129] then to June 2020,[130] then to April 2021,[131] and most recently to November 2021.[20]) I think the spokesperson is being diplomatic and stating essentially "if it were cheaper it would be cheaper". I don't think we can ignore the cost of labor involved in constructing this rocket. If the labor became more efficient in the future, the cost would go down, but has this happened?
 * Appropriations letter itself: previous bullet point
 * The SpaceNews article, found here. It says NASA has been reticent to provide cost estimates for a single SLS launch, although ballpark figures have been around $1 billion. NASA, in its fiscal year 2020 budget request, said launching the mission on a commercial rocket, such as a Delta 4 Heavy or Falcon Heavy, “is estimated to result in over $700 million in savings compared to use of an SLS rocket.” This provides some explanation, but does not say why NASA has been reticent. Why is it the case that in order to find a cost for this launch vehicle, so much digging has to take place?? A NASA Office of Inspector General (OIG) report in May played down the cost savings between SLS and commercial alternatives largely because it offered a much lower cost estimate for the SLS. That report estimated an SLS launch at $876 million, versus $450 million for commercial alternatives. However, in a follow-up letter in August, OIG said NASA could save up to $1 billion by launching Europa Clipper on a commercial vehicle versus the SLS, adding that NASA needed to decide how to launch the mission in the next few months in order to procure a launch vehicle in time to support a potential 2023 launch.. This is interesting at first glance, but looking at the actual follow-up letter, it appears that the $1B figure is an all-inclusive figure that also takes into account the cost of storing the spacecraft (at least 2 years at a cost of $3 to $5 million per month until an SLS becomes available), so it is not quite the case that SLS would cost $1B more on its own. 2 years times 3 to 5 million per month arrives at $72M to $120M. This is likely a part of: NASA recently added $250 million in Headquarters-held reserves to the project to address these storage and related personnel costs. So, of the $1B, $250M is this, leaving $750M for SLS versus an alternative. This is still significantly greater than the $876M versus $450M (difference of $426M) figure from the earlier OIG report, but still roughly in line with the FY2020 budget request which said that there would be a $700M savings by switching away from SLS. Looking a little later in the letter on pages 4 and 5, I see Our May 2019 report highlighted three main differences between launching the Europa mission on an SLS versus a commercial rocket: cost, transit time to Europa, and availability for a 2023 launch. The SLS is the most expensive launch vehicle option by a factor of three. In fact, the JCL analysis conducted by the SRB for Clipper showed the SLS would cost about $700 million more than a commercial vehicle option. Off-setting a small part of the higher cost is the fact that the SLS is powerful enough to carry Clipper directly to Europa in about 2.4 years whereas a commercial vehicle will need to use a planetary gravity assist trajectory for a transit time of at least 5.9 years. Bit of arithmetic, but if SLS costs $700M more than the alternative, and is 3x as expensive, this means that the alternative is $350M and SLS is $1.05B.
 * Finally, the budget request for 2020, which allocates $2.5B to SLS for this year.

In summary. I put the most weight on what a NASA spokesperson said directly, in reply to the inquiry from Ars Technica, relating to the cost of labor. Drawing on the spacenews article, seems to me as if NASA is providing a sticker price aspirationally, in a manner that doesn't include fixed cost of maintaining its employees. This is supported by the NASA spokesperson's quoted statements. But all the labor that goes into building the rocket does count as a cost, even if it isn't "billed" to the projects that want to launch such as Europa and Origins etc. The accounting seems murky here and all based on estimates of varying age, but the clearest figure that cuts through the noise (without including development costs) does appear to be $2B per launch, which includes cost of labor to build the rocket (why wouldn't it?). We also see marginal costs (not including fixed salaries etc) in a fairly tight range: $800M to $900M as a lower bound aspirational figure from the administrator, and $876M to $1B from Europa. I discount the two estimates for what SLS might cost for a launch in 2035. In summary, I see these sources as supporting, with all development costs left aside, a $2.5B annual cost, a $1B marginal cost of launch, and a $2B cost per launch. Leijurv (talk) 21:37, 13 July 2020 (UTC)


 * WOW! What an effort! I agree with almost everything you said! But man, wow! - Moamem (talk) 01:55, 14 July 2020 (UTC)
 * One thing I'd like to clarify is where the ultimate source of this $2B figure is, anyway. Does Ars Technica ever describe where they get it from? Is it an estimate? Or is it based on yearly programmatic budget? If it's the latter, I have some points I'd like to make in contention, but I'd like to hear your interpretation first. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 02:02, 14 July 2020 (UTC)
 * I believe I said that, didn't I? As I said, I don't agree with some of the later conclusions of the Ars Technica article (it's too speculative), however it is the source on the NASA spokesperson's statements to the author, so we can at least use those parts. Let me copy/paste the start of my fifth bullet point:
 * An Ars Technica article by Eric Berger. It appears to cite the next source which says NASA Europa Mission. The bill requires that NASA use the Space Launch System (SLS) rocket to launch the Europa Clipper mission. The Administration is deeply concerned that this mandate would slow the lunar exploration program, which requires every SLS rocket available. Unlike the human exploration program, which requires use of the SLS, the Europa mission could be launched by a commercial rocket.  At an estimated cost of over $2 billion per launch for the SLS once development is complete, the use of a commercial launch vehicle would provide over $1.5 billion in cost savings.  The Administration urges the Congress to provide NASA the flexibility called for by the NASA Inspector General and consistent with the FY 2020 Budget request.
 * So, as I wrote, Ars Technica is citing the next source that we also cite, which is this. The relevant passage is At an estimated cost of over $2 billion per launch for the SLS once development is complete, the use of a commercial launch vehicle would provide over $1.5 billion in cost savings. We already say this in the text of the article proper: A letter from the White House to the Senate Appropriations Committee revealed that the SLS's cost per launch is estimated at "over US$2 billion" after development.[95]. I see no justification to consider this number to have been based on yearly programmatic budget, I don't know where that's coming from. It is a cost per launch figure for SLS, after development. As I say in the above comment, this includes the cost of labor to build the rocket (why wouldn't it?). This is, again, supported by the NASA spokesperson quoted statements. I suspect you disagree, if so I'm all ears, why shouldn't those aspects be included? (referring to the costs of maintaining a standing army of thousands of employees and hundreds of suppliers across the country. This is very clearly to me a part of how much it costs to build SLS. Again, if they could build more rockets using the same labor force, the cost of labor per rocket would be cheaper. But that's a hypothetical that is not actually happening. The whole idea of the "marginal" cost is, to my eyes, a giant hypothetical because while we can imagine "well WHAT IF they built just one more" it doesn't arrive at an accurate conclusion because it (by design) doesn't scale up the fixed costs. If up to two SLS rockets can get built a year with these fixed cost facilities, factories, suppliers, and employees, then the marginal cost of one more SLS is sure going to include 6 months of those fixed costs, plus the cost of procurement of materials. This isn't like some car factory that can just as easily build 1 more or 1 fewer car on any given day with no change in labor costs. Each one here is a massive undertaking, and the analysis of a marginal cost by its very nature assumes fixed costs remain fixed. They don't. Simple test: if we wanted 100 SLS rockets, what would actually happen? We can't wait some half century at the current "fixed" cost per year of salaries and suppliers. To acheive this in, say, 5 years, there would need to be 10x more factories and facilities... and that fixed cost suddenly becomes less fixed. But again, that's another hypothetical. We have a letter to Congress that rests upon an analysis of these costs, and a NASA spokesperson tried to damage control by explaining that in a hypothetical that if they could make more with less, more would be made with less. Leijurv (talk) 03:01, 14 July 2020 (UTC)
 * You can actually see in my first comment on this talk page yesterday that I thought that SLS could be split out into "program cost per year" and "marginal cost per launch", but it truly cannot due to the limitations of the construction of SLS: everything that's a so called "fixed" cost, such as factories, salaries, and suppliers, is actually a cost that will scale up with the rate of production (as was betrayed by the NASA spokesperson's comments). Keep in mind that the NASA spokesperson did not dispute the $2B figure, but instead explained the hypothetical of labor making more rockets while being paid the same resulting in cheaper rockets... Want to build twice as many a year? You need a second factory and workforce. Want to build twice as many with the same fixed costs per year? It'll take twice as many years. These are marginal costs in fixed costs' clothing. The only thing that I'd put in a "fixed cost" bucket is the development cost prior to the first one. Leijurv (talk) 03:10, 14 July 2020 (UTC)
 * That's not really how it works. Most of the people involved's salaries, and the infrastructure which has to be maintained really are fixed. If they build no SLS rockets in a given year, those costs still have to be paid. They can't fire people (or shift them to other jobs) and then expect them to come back the next year. To maintain the ability to build SLS, NASA has to pay those costs, regardless of the number of launch vehicles they build per year. The current workforce and facilities can build up to one, possibly two, per year. If they wanted a higher production rate, they'd have to hire more people and expand facilities. But just to be able to maintain the existence of the SLS program, they have fixed annual costs, whether they fly zero, one or two per year. — Preceding unsigned comment added by Fcrary (talk • contribs)
 * Basically, I don't buy this. Yes, in a hypothetical scenario where NASA drags their feet and builds SLS slower than they are capable of, then the cost per SLS would increase. This is just... how it works. Isn't this how the manufacturing of everything, ever, works? "If we slowed down our assembly line to half speed, our salary costs would stay the same, but each item we produce costs us twice as much in labor"? I don't think SLS deserves special treatment where we take the cost of labor out of the cost of manufacturing.
 * And regardless, can you say more about If they wanted a higher production rate, they'd have to hire more people and expand facilities.? I'm not sure why you believe that a higher production rate could be acheived by hiring more people and expanding facilities, yet don't believe that a lower production rate could be a acheived by laying off some people and reducing facilities.
 * Now, obviously, that's a bit of an exaggeration because we can't simply describe this as a linear . It isn't linear. The analysis of binning the cost into what's marginal and what's fixed breaks down, for the reasons I mentioned above. A fixed cost is not fixed if it changes with the production rate, can we agree on that? And as you said, the current workforce is capable of possibly two per year, and if they wanted a higher production rate, they'd have to hire more people and expand facilities. That sounds correct to me. Therefore, labor is not a fixed cost. On the other hand, we cannot cleanly move labor to the other side of the equation. As in: I'm not claiming we can "fix" this "model" and make it perfectly accurate by moving labor from the left side to the right side where it gets multiplied by number of rockets. It doesn't scale down infinitely to zero, because of speed. If a gap in production developed, where no SLS rockets were desired, so none were being constructed, we'd need to reevaluate how to present the cost information. But is this happening? It looks to me like production is going as fast as possible for these Artemis missions. Given this reality, it isn't perfectly accurate, but it is quite close to it, to say that labor scales linearly with the number of rockets constructed. The reason is that we don't have these gaps in production; they aren't happening. SLS is being built at the speed it's being built, and they are certainly not outpacing outside demand I'm sure we can agree. There is no such outside force of low demand slowing them down.
 * For that reason, I argue that between the hypothetical scenarios of "additional rocket needs to be built" and "gap in production", the latter can be pretty much entirely discounted because it simply is not happening, while the former is much closer at hand as we see whether or not future Artemis missions (or other SLS launches) will occur. I'd also like to point out that in the "gap in production" scenario, the cost per SLS would increase, not decrease. So I don't see any reason to twist the "cost per launch" number downwards by removing labor from the equation. We have a very likely hypothetical and a very unlikely hypothetical, in the likely one we have a quite-close-to linear relationship between rocket count and labor cost, and in the unlikely one the cost per rocket balloons even further into the billions. To put it in fully concrete terms: we can write a single "cost per launch" number that is accurate to reality, but we cannot write a single "cost per launch" number that is accurate to any and all potential future changes to production. So which number is that?
 * Having looked at all these sources, I believe the explanation is that what we're seeing is just NASA accounting being unintuitive. It's important to keep in mind that the figures presented to Europa HabEx etc are purely internal numbers. If Europa wants to launch on SLS, they need to give this much of their budget to make it happen. But, as was stated earlier on this talk page, SLS production is happening anyway. If the "SLS team" doesn't bill the "Europa team" for labor, does that mean the labor is free? Nope! It still occurred, and NASA, the parent organization, is still footing the bill. These prices would never be offered externally (and if they were, no one would take them up on it, as evidenced by Europa clamoring to get off of SLS). The benefit to the Europa team of switching away from SLS is $700M but that doesn't mean that the benefit to NASA as a whole (really, the goverment as a whole) is just $700M; we have a much higher level analysis than just one NASA subteam that says that launching Europa on SLS will cost $2B all told, which is $1.5B extra compared to commercial. (On top of which, the two lower numbers, Origins and HabEx, are for 2035). So, I honestly have serious doubts on how reliable the internal figures (e.g. Europa) are for this reason, in comparison to the letter from the administration that criticized the high cost taking everything into account. Leijurv (talk) 22:35, 14 July 2020 (UTC)
 * Looking at the positions people have taken over the duration of this thread, it seems pretty clear that there is only one person that believes that the cost per launch (no matter how you try to isolate costs) will be less than $2 billion. When you get down to it, the current trend in this discussion is for everyone to try and reach some kind of compromise with Jadebenn.  His stance is and absolute fringe position.  This is like giving a nazi an equal voice in an race equality page.  Their position would be aberrant to begin with and suppressing it is entirely justifiable.  Better then 90% of posts in talk seem to be in opposition to at least some of what he is saying.  If you go back and look at the edit history then there are dozens of people that have tried to edit the cost figure to be $2 billion or higher.  I think Moamem has done a pretty good job trying to wrangle this CF toward resolution but I think he has fallen into a trap of behaving like a reasonable person with someone that is intrinsically unreasonable.  Compromising the costs to be $900 million to $2 billion does not reflect what the evidence supports.  $2 billion is a far more realistic lower boundary and there are plenty of sources that support that figure.  An upper boundary closer to 4 billion is similarly more realistic. Stop giving ground to someone that won't compromise.SandowTheHeretic (talk) 02:06, 17 July 2020 (UTC)
 * I believe you can express this without comparing Jadebenn to a Nazi. I'd remove that bit, perhaps. Leijurv (talk) 04:24, 17 July 2020 (UTC)

Okay, now that things are a bit calmer, I'd like to point out a recording of a NASA official breaking down a similar $1.5B budget estimate, cited in another OMB paper, who essentially states it's likely a $2B program cost per year estimate over a single flight per year. The objection I have with this is that SLS program costs do not directly correlate to "costs needed to run SLS." I'll given an example: eCryo is a cryogenic technology development program that receives part of its funding from the SLS program. This is because the technology could be useful on a future variant of SLS, but whatever slice of the program cost that is (could be anywhere from a few million to a few hundred million), is 'completely' unrelated to launch costs. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 06:59, 17 July 2020 (UTC)
 * That's a very interesting argument to me. A couple things. Can you clarify further, are you saying that you think it's unfair that SLS is paying for other seemingly unrelated development, or are you saying that it's an accounting error that the SLS is marked as paying for those programs and it isn't truly?
 * If I might make a tongue in cheek example: if SpaceX decided to mark up the Falcon 9 by +100% to cover development costs for the next rocket, would you also say that this unrelated development program that receives part of its funding from the Falcon 9 platform doesn't "count" as a Falcon 9 launch cost, and so we should stick with the old "correct" launch price that just covers the Falcon 9?
 * What is the connection between this NASA official's methodology and the one used by the White House to arrive at the figure cited in their letter? Just a similar end result...?
 * I'll also reply to something you said earlier up in the thread, to save time: I think if all those fixed costs were directly traceable to things needed to launch SLS, you'd have a stronger argument, but they're not. As previously mentioned, it's not just the guy that mows the lawn of the VAB or the salaries of the flight controllers at launch control, but general R&D programs like eCryo that have applications beyond SLS, or refurbishment of multi-user infrastructure such as an engine test stand, etc. To give a real example I've heard from someone who works at MSFC, the SLS program pays for technicians to provide machine shop services. When SLS doesn't need them, other NASA programs are allowed to make use of their services. Now that the SLS program is considering cutting those services, the other programs are scrambling to pick up the tab or find other alternatives. That's a minor example, but I think it illustrates the point I'm making quite well: SLS program costs pay for more than just SLS. You can't use that figure for "launch cost."
 * I have to disagree here. If making 1 SLS takes 6 months of The Rocket Factory's time, and The Rocket Factory needs to maintain its facilities and pay wages to all the workers (yes, even including whoever mows the lawn), then the cost of 1 SLS does absolutely include 6 months of The Rocket Factory's costs. Even if The Rocket Factory was previously used for other rockets. Even if after SLS, The Rocket Factory will need to find other ways to cover its costs. This analogy also applies to the machine shop. If SLS is paying for the machine shop, it doesn't matter if you think this is unfair, it doesn't matter if they might stop, it doesn't matter if they didn't used to, it doesn't matter if they let other NASA programs hitch along for free. The cost is still there, and SLS is paying for it. This is one of the costs associated with building this SLS! I fundamentally strongly disagree with the idea that, if SLS is as bloated as you describe and paying billions for things that aren't necessary for launch, that we should somehow sweep this under the rug and subtract those out, and not provide the true full cost per launch. This goes in parallel with what I was saying above about Europa (forgive my imprecise language, but): If the "SLS team" doesn't bill the "Europa team" for labor, does that mean the labor is free? Nope! It still occurred, and NASA, the parent organization, is still footing the bill. Could you respond to some of what I was saying about including cost of labor?
 * Let me take your machine shop example to a bit of an absurd extreme to make a point. SLS pays people to work on constructing SLS. These people also support their families with this salary, even though the families are not helping SLS AT ALL! (this is unfair, SLS budget should only support SLS, let's subtract out every portion of salary that is shared with family members). And now that SLS is considering laying these people off, the family is scrambling to pick up the tab or find alternatives.
 * I don't think we can pick and choose which "parts" of the SLS budget are truly part of "a launch". It's a never ending stream of exceptions. If NASA budgets this amount of money per launch, then that's how much a launch costs.
 * I'd also like to reply to this other thing you said:
 * I also think a lot of the dispute comes down to the difference in the way a commercial rocket's price is calculated and the way NASA calculates SLS price. Moamem is not wrong to say that a rocket like Falcon 9 or Atlas V includes fixed costs in its launch pricing. However:
 * One, those rockets have many missions per year, so the fixed costs are minimized in a way that SLS's are not.
 * Two, SLS is not a commercial rocket, NASA does not need to recoup its costs.
 * Three, even if you still ultimately believe the same accounting scheme should be used, there remains the issue that we do not have an actual figure for pure SLS launch costs that both includes SLS fixed costs but excludes items unrelated to launch. This is the previously mentioned BOLE/EUS/RS-25E/eCryo problem. In terms of sheer practicality, that $500M to $900M is the closest thing to the actual launch cost we've got.
 * I see this the opposite way as you, evidently. If this rocket has significantly higher fixed costs, then that is all the more reason to include them in the cost of launch. The extremely low volume of the SLS makes it all the more disingenuous to exclude the high fixed costs, it is not a rounding error. If a rocket's fixed costs were minimal compared to the cost of launch, that would be reason to discount them. The only way I can see it the same way as you is if I wanted to paint the SLS in as desirable a light as possible... Imagine we wanted to make a new International Prototype of the Kilogram. Huge development costs, many people's salaries, etc. The actual manufacturing is fairly speedy and cheap, but it stands upon millions of dollars of research. Let's say we make two of them, and won't make any more. Which price should we say was the cost of replacement, per kilogram? Hint: cost / 2.
 * The exact same thing goes for NASA not needing to recoup its costs. Since this is paid for with public taxpayer money, it is all the more important to accurately describe how much this space launch system is costing them. If they're really playing it as fast and loose as you describe with the budget free flowing into unrelated endeavours, then that's extremely questionable and the least we can do is keep them honest and put an accurate figure in this article. Yeah if it were a private company, we could simply put up the price they charge per launch and be done with it, but as we both know, it isn't that simple internal to NASA.
 * Regarding everything about the RS-25E etc: Then how do you explain the $2 billion figure being stated as "after development is complete"? Leijurv (talk) 07:40, 17 July 2020 (UTC)
 * I don't think we can pick and choose which "parts" of the SLS budget are truly part of "a launch". It's a never ending stream of exceptions. If NASA budgets this amount of money per launch, then that's how much a launch costs. NASA doesn't budget $2B per launch. Again, they split things into program costs and launch costs. This has actually come up in a couple of places, such as most of the decadal reports currently cited, and in terms of what a commercial user of an SLS should be charged in the event of an SLS-launched lunar lander.
 * I understand the point you're making about the factory/machine shop/VAB not being severable from making a completed rocket, which is why I brought up the example of eCryo. NASA is not a company, as we all know, and they're expected to spend budget on general technology development projects. The SLS program is funding part of eCryo because the technology it develops could have future applications for EUS, but it's not exactly comparable to, say, ACES, where the objective of the development program is to directly produce useful rocket bits. Do you see the distinction I'm making? They're putting money into a general technology development program because it could benefit them, but it's not directly transferable to SLS. This happened in the Shuttle days too. Some of the tech did end up making its way to STS. Some of it didn't.
 * There is no real distinct point where "development finishes." Looking at the future, we can see a continuous plan of replacement of obsolescent parts/upgrades for further capability. RS-25Es replace the RS-25Ds in the mid-2020s, BOLE SRBs replace the RSRMVs in the late 2020s/early 2030s. And that's just the stuff that's planned right now.
 * Ultimately, I've come to agree with the view that we should keep a range of costs, because it seems clear to me there's no directly comparable figure. You can argue the so-called "marginal cost" is leaving out costs that should be considered, and you can argue the so-called "program cost" is including things that shouldn't. There's no neat answer. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 08:49, 17 July 2020 (UTC)
 * What is curently being placed into these program vs launch buckets? And why didn't I see this breakdown in any of the sources that I painstakingly went through in my initial comment ^?
 * what a commercial user of an SLS should be charged in the event of an SLS-launched lunar lander This is not going to be equal to cost of launch since SLS is operating "at a loss" of over $20B (just looking at the development cost number).
 * I'm not being inconsistent here: in an extreme example (e.g. numbers such as this), I would not trust the "cost of launch" number provided by a commercial launch provider as truly representing how much it costs them to launch, if they were operating at a loss this extreme. In that scenario I would advocate for us to find an analysis of how much they are truly spending on each launch.
 * I undestand your eCryo example. SLS is giving funds to a development program that could, if successful, help the upper stage of SLS, but might not pan out. Yes? Then I'll modify my example: if SpaceX decided to mark up the Falcon 9 by +100% to cover development costs for the next rocket a new upper stage, would you also say that this unrelated development program that receives part of its funding from the Falcon 9 platform doesn't "count" as a Falcon 9 launch cost, and so we should stick with the old "correct" launch price that just covers the Falcon 9?
 * I don't think I accept your premise that the $2B number that is explicitly "after development" is actually a falsehood, isn't after development, and is a division of program cost. (and even if I did, that probably is pretty accurate still). On what basis do you think that this number is wrong? I see your statement that development will never finish, don't you think it's possible that they could have taken this into account? And even if there's development into the early 2030s, please remember that the two lowest numbers currently cited are for launch in 2035, so if they can project that far into the future, why can't the White House (if we take it as a given that the way the defined "after development" was chronological instead of the other meaning of "after" which is more of "discounting development" or "without development")?
 * This happened in the Shuttle days too. Some of the tech did end up making its way to STS. Some of it didn't. If I give you twenty billion dollars to build and launch two rockets to the moon and you spend half of it on ideas that don't end up panning out, the money was still spent. It is disingenuous to say each rocket cost $5B when the program really spent twice that. The accurate figure would be $10B per launch.
 * If I take that as a given that development will never truly stop, that is, yet again, all the more reason to fold it into the cost of launch. If SLS will spend $1B on 1 rocket and $1B on developing the engines / boosters for the immediate next rocket, every year, for the next two decades, then the cost of launch of SLS is $2B. If we never truly end development and enter production then the only fair thing to do is explain this to the reader. Could you give your thoughts on my example of the kilogram, as well as What is the connection between this NASA ...?
 * At the moment, what range do you think would be the most accurate and truthful?
 * Let me give an example. My current thinking is we should say: Development cost: $18.6B (as of 2020), Cost per launch: $2B (after development), Launches per year: 1, Cost per year: $2.5B (2020)
 * I believe launches per year is a needed figure in any event, to accurately demonstrate to the reader the burn rate in comparison to the launch rate. Even if we tone down the "cost per launch" to that internal number projected 15 years in the future, the reader can grasp what's going on if they read "Cost per launch: $0.5B. Launches per year: 1. Cost per year $2.5B". Wouldn't that be better? They can clearly tell WHY this (too low) cost of launch figure is so different from cost per year, clearly other things are being folded in.
 * You can argue the so-called "marginal cost" is leaving out costs that should be considered, and you can argue the so-called "program cost" is including things that shouldn't. There's no neat answer. I no longer really believe that there is such a thing as marginal cost for SLS. It defies that economic analysis tool (idk what to call it) because every individual rocket is a very nonlinear step function. You can't differentiate cost with respect to number of rockets and arrive at a meaningful figure (the marginal cost), like you can with other things e.g. the example of a car factory making thousands of cars a day can just as easily make one more or one less. Especially since we're talking 6 to 12 months of people's salaries for every single one. Leijurv (talk) 17:37, 17 July 2020 (UTC)


 * Let me ask a simple question of both Jadebenn and Leijurv. The Space shuttle took off 134 times and the program cost ~$209 billion in 2010 dollars.  How much did each launch cost on average?SandowTheHeretic (talk) 16:17, 20 July 2020 (UTC)
 * Interesting... Dividing 209 into 134 (or 211 into 134, taking the 211 number from Space Shuttle) gives 1.56 billion. Then how come the cost on Space Shuttle ranges up to 1.64 billion? That's a little strange / interesting. Anyway. I would say that the average space shuttle launch cost yeah about 1.5 billion. There are a few differences, such as that constructing a space shuttle is quite clearly and presently reusable (at least partially), and they were reused many many times, which brings us a bit closer to having a reasonable figure for marginal launch cost. I have not looked at all at any of the space shuttle sources, but off the top of my head I suspect we could more cleanly divide what was initial construction and what was refurbishment? I also see in the article actually In 2009, NASA determined that the cost of adding a single launch per year was $252 million (in 2012), which indicated that much of the Space Shuttle program costs are for year-round personnel and operations that continued regardless of the launch rate. This is interesting because the Space Shuttle could turn around in 54 days (all time best) or later on 88 days (source), so I actually sort of believe this. Like, the reusability combined with fast turnaround time leads me to believe that an analysis of this cost is a bit closer to the "car factory" example I gave and less like the SLS. Maybe they could truly Just Add Another Launch. I don't know what the constraints were though; was it budget? Not enough payloads wanting to launch? On the other hand, the total cost number including the TRUE cost per launch (all inclusive) seems pretty good to me. SLS will not be reused, making our job simpler. "How much did we spend, all told, per launch" is a pretty good criteria for "Cost per launch", why wouldn't it be? But... I don't really think we can do that on this article yet. We don't have a solid number on how many times SLS will launch total so (unless I'm wrong about that) I don't think we can reasonably divide the development cost (from way before first launch) into an unknown number of launches. So I think the "Development cost" / "Cost per launch" (two figures) should stay on this infobox, but I do still think that "Cost per launch" has just got to include manufacturing labor and suppliers in order to be accurate to reality. At the same hand, we definitely need to do (better) justice to this discussion on the actual article in a Budget section like what Space Shuttle has, in my opinion. Also I don't think it's a simple question when there's a long term controversy / debate on how to characterize the cost of the space shuttle :) (edit: I said more on this here). Leijurv (talk) 18:40, 20 July 2020 (UTC)
 * Does that ~$209 billion include any costs for developing the Crawler-Transporter? Launch Complex 39? Eggsaladsandwich (talk) 20:21, 20 July 2020 (UTC)
 * I don't know, but taking a very quick glance it seems like they were built for Apollo then repurposed. Off the top of my head I'd say it probably just includes the refurbishment costs? Leijurv (talk) 23:01, 20 July 2020 (UTC)
 * I think many of the cost differences 1.5 to 1.6 billion are based upon the constant dollars date. 2020 dollars would make it closer to 1.6 than 1.56 in 2010.  There is also the difference of 132 missions that didn't fail vs 134 the left the pad.  The shuttle refurb time was usually 6 months or longer so often the number of missions were determined loosely by shuttle availability.  Particularly toward the end of the program and in the post challenger years.  In either case, you hit upon the key question. "We don't have a solid number on how many times SLS will launch total so (unless I'm wrong about that) I don't think we can reasonably divide the development cost (from way before first launch) into an unknown number of launches."  We don't really know how many there will be at the low end but we know what is possible at the high end.  At the low end, we might see two launches before the program is cancelled.  One if the first mission fails is also a possibility though.  With a change in administrations and the possibility of competing commercial vehicles beating it to launch, a lot could happen.  At the high end there are missions slated until 2030 with the first launch at the end of next year.  Nasa has stated repeatedly that they don't believe that more than one can be made per year but Boeing has stated that they plan on making 2 per year by 2024.  This gives us a range of 1 to 17 launches by the end of 2030.  A more realistic lower boundary is probably 9 missions though as cancellation is unlikely and human rating another vehicle within the next 5 years is also unlikely.  In either case, it is almost certain to be hopelessly obsolete by the end of 2030 even with continuing development.  The total program costs are around 20 billion so far and if the budget continues at 2.5 billion then we are at 45 billion by the end of 2030.  No matter what we do at this point, we can't have a launch cost simply because it has never launched and won't for another year and a half.  Is it not simpler and more honest to just agree on a realistic projected cost range and list it is as "projected costs"?  With 17 launches we are at 2.64 billion per launch and 9 launches gets us 5 billion per launch out to 2030.  FWIW though, the cost per pound to LEO at 5 billion per launch is about on par with the space shuttle which sort of makes sense if you let a Kerbal rearrange parts. SandowTheHeretic (talk) 13:23, 21 July 2020 (UTC)
 * Oh boy. Lol. So let's say we take Space Shuttle as a guide for this (which I think is perfectly reasonable). The high end of the "Cost per launch" range there is the program cost divided over the number of launches. Inclusive of all development. I didn't really process this, but wow. Seems like you're saying realistically that this could get up to $5 billion? When I look at the nearly $20 billion development cost and the low number of launches, that doesn't seem too unreasonable. For the lower end, we can also take guidance from the Space Shuttle article: even though their marginal number (one more launch in the same year) was $252 million, the lower range is $576 million. Clearly, the precedent is that the theoretical marginal figure for one more launch in the same time period is NOT to be used in the infobox. What should the lower range be here? Probably the launch cost excluding development, which is $2 billion? So maybe "Cost per launch" should be $2B to $5B? That's quite high but Space Shuttle is probably the best precedent there is so... Perhaps ? Leijurv (talk) 17:01, 21 July 2020 (UTC)
 * If you look (here) You can see where the low number comes from. The last year of the mission had almost all of the support costs removed and they fired off three of them.  There was a joke during the active years that the first shuttle of the year cost 5 billion and the rest were free...  Trying to figure out the real costs to tax payers of any government program is nothing new.  Most of the actual costs the last year of launches come from previous years work.  Ultimately it simply isn't representative to use that year for the costs any more than it is to use the two years after challenger in which there were 2 launches and 15 billion was spent. The only number that is fair is the average across the program. In SLS we need to make realistic estimates of what the future will bring which is obviously harder.  Personally, I don't believe that we will see a launch in 2021 so 2022 making a range of more like 8 to 15 launches more realistic. I don't think 5 billion is over the top for the upper boundary. SandowTheHeretic (talk) 02:00, 22 July 2020 (UTC)
 * Sorry, I must be missing it, but I don't see the $576M figure there?
 * Also, I'd replace we need to make realistic estimates with "we need to report the realistic estimates that were made by reliable sources" :) Leijurv (talk) 02:56, 22 July 2020 (UTC)
 * I was mostly referring the budget/launch chart per year and pointing out that dividing the missions flown in the last year by that years budget gets you in the ~$300 million range. And you are right, we should of course use referenced research whenever possible. However we should not treat basic transparent math as something that we are allergic to.  Wiki policy is to not use your own research for articles which is a sensible thing to apply if we were pouring over nasa budget reports ourselves. Something like "(20+25)/9=5" is not original research any more than concluding that "fire is hot" would be. Currently there are 11 planned SLS launches. Artemis 1-10 and the Europa Clipper. So our transparent research free math becomes something like (20+25)/11= $4.09 billion. SandowTheHeretic (talk) 11:48, 23 July 2020 (UTC)
 * Always important to keep this, this, and this in mind. I believe our total budget figure is a summation of budget figures for example... why are we performing arithmetic summation on values that were subject to inflation over time, instead of the column that is inflation adjusted to the current year? I don't think we should sum up the raw numbers at all, for what it's worth. Leijurv (talk) 17:02, 23 July 2020 (UTC)


 * This question is very broad. What are all of the proposed alternatives and what are the reasons for them? I see a proposal to split it into several figures as well as a proposal to broaden the range? &lt; Atom ( Anomalies ) 23:46, 1 August 2020 (UTC)

"Conclusion" of the RfC for the SLS Launch cost
Hi everyone, After months of debates I think a clear consensus has emerged. I went trough the different opinions given in the last 9 months (since November 2020) and tried to compile them in a manner useful for the dilemma at hand :

"1) Keeping the $500 million launch cost figure :

For : Jadebenn

Against : Leijurv, Maomem, 174.251.160.98, SLSgal, Fcrary, Soumya-8974, N2e, Timlograsso, Zegfred, Skytie, TheSkalman, 2001:56a:7797:e800:acd7:1c51:cba5:6365, RundownPear, NguyenVy1993, Beaucouplusneutre, Sun Creator, SandowTheHeretic

Unclear : SkywalkerP"

"2) Separating Marginal Launch Cost ($900 million) and Launch Cost ($2 billion) - Votes for eliminating the marginal cost completely will be considered as a vote for separating them :

For : Leijurv, Maomem, 174.251.160.98, SLSgal, Fcrary, N2e, Timlograsso, Zegfred, Skytie, TheSkalman, 2001:56a:7797:e800:acd7:1c51:cba5:6365, RundownPear, NguyenVy1993, Beaucouplusneutre, Sun Creator, SandowTheHeretic

Against : SkywalkerP, Jadebenn

Unclear : Soumya-8974"

If you spot a mistake or you have been inadvertently miscategorized please say so and I will correct accordingly.

If you want your username added please state the proposition (1 or 2) and you position (for or against).

If a clear consensus persists I will make the changes in 24h.

Please do not debate here!!!! This is specifically for the tally. Any debate should take place in the sections above! Thanks! - Moamem (talk) 01:52, 14 July 2020 (UTC)

, that's not how RfCs work. WP:NOTAVOTE. You're also mis-ascribing editors' positions.

Can you please actually let discussion continue? You do something disruptive every time an editor vaguely supports your position, and it completely throws off the debate. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 02:35, 14 July 2020 (UTC)
 *  I am beyond furious at your unacceptable behavior! You do not have the right to delete my comments from the talk page! As I said on your talk page this is not a reason to delete my comment. Here are the very specific circumstance under which you can delete other's comments : Editing others' comments. Not liking them ain't one of them! This is unacceptable! - Moamem (talk) 03:11, 14 July 2020 (UTC)
 * If you have any type of issue with any of my comments, we can discuss it and I am open to editing MY OWN comment! But DO NOT edit my comments yourself! - Moamem (talk) 03:25, 14 July 2020 (UTC)
 * I'm not so sure... clicking that link WP:NOTAVOTE I do see There is no absolute prohibition on polling, and there are often objections if a poll is summarily closed or deleted on sight using a claim that they are forbidden. Editors who feel that a poll is inappropriate under the circumstances may instead note that further commentary is needed, encourage the discussion to migrate back to a free-form conversation, or open a related discussion. and Polls may be helpful in coming to a consensus and in evaluating when a consensus exists, but consensus can change over time. I see no problem with summarizing who said support of what. I also do not see any summation/count or implication that a majority wins. Leijurv (talk) 03:19, 14 July 2020 (UTC)
 * Yes, polling is fine it's just not the end all be all. And even if it wasn't that's still not a reason to edit my comments :Editing others' comments! I'm so furious right now! - Moamem (talk) 03:25, 14 July 2020 (UTC)

Requested for closure at Administrators%27_noticeboard/Requests_for_closure. --Soumya-8974 talk contribs subpages 10:02, 19 July 2020 (UTC) Requested for closure at Administrators%27_noticeboard/Requests_for_closure. --Soumya-8974 talk contribs subpages 10:01, 19 July 2020 (UTC)
 * Hi, I think I made my positions pretty clear on this matter, Ill just sum up my main arguents :
 * 1) $500 launch cost is as N2e put it "utter fiction" :
 * It's the initial target launch cost for SLS back in 2011, was unachievable then but now we KNOW it's not even in the realm of possibilities.
 * It's more or less the cost of the RS-25 engines on the first stage ($400 million excluding development cost) and we have those for the next decade worth of launches.
 * NASA has stopped quoting this figure since at least 2018. It's an outdated figure (wasn't realistic to begin with)
 * The sources that cite this numbers is a proposal for a mission. Their clear objective is to lower the cost of their proposed mission to get funded! In the source the section about cost and methodology is basically redacted. This is not a neutral source but a partisan one!


 * 2) Marginal cost Vs Real cost :
 * Marginal cost IS NOT the cost! For me it seems self evident. If it doesn't for you please read the definition on Wikipedia : Marginal_cost
 * No other rocket uses Marginal cost as its launch cost. NONE! Saturn V, Space Shuttle, Falcon 9, Atlas 5, Delta 4... None of them. This is for me the most compelling argument : consistency.
 * No one think marginal cost when they read cost : this is an argument about understandably and confusion to the reader. When people read Cost no one expect it to exclude fixed cost for a reason I still don't understand.
 * With SLS NASA gets a pass on a lot of "creative accounting" : Separating the cost of the ground infrastructure from that of the launch system, Not capturing the full cost of the system in its lifetime estimates (see OIG report), Totally ignoring the development cost in its launch cost calculation... But trying to totally erase Fixed cost? That's a new frontier.


 * Given the overwhelming consensus and the reasonableness of the requests (I means since SLS is planned to be launched less than once a year well into the 2030's, what's the point of even mentioning Marginal cost? Or why not include dev cost estimates over the planned mission (12 for now) like STS does?) this minimal compromise seems to be the least we can do. - Moamem (talk) 16:07, 20 July 2020 (UTC)
 * It is pretty clear that the 500M figure is not the current truth of the matter. I'd also like to add that the source supporting $500M is scheduled for 2035, and when taken on balance with the other sources of estimated future figures (e.g. the administrator's estimate of $800M to $900M several launches in the future), it is quite clearly an aspirational estimate of what SLS might cost then.
 * The RS-25 engines are a reasonable "sanity check" that does demonstrate that $500M is unlikely, I don't think we can use them for much more than that. Perhaps it could be mentioned in the article body though.
 * The source is indeed not neutral. I argue above that this is a NASA-internal figure regarding what is essentially "inter-team billing".
 * I don't think we can have a marginal cost anymore. See my argument here. I think the better way to convey the information is to add a new field for "Launches per year". So we would have "Development cost", "Cost per launch", "Lanuches per year", and "Program cost per year".
 * I agree that "development cost" vs "fixed cost" deserves consideration. See here. I agree with the argument for consistency with other articles. In that discussion about the space shuttle, it turns out that NASA did a full analysis of the marginal cost of adding one more launch to the schedule in any given year, and it came out to $252M. Yet, the lower range of the cost per launch figure on Space Shuttle is $576 million. For consistency, I don't think we can use this hypothetical figure of "what if we could build one more rocket in a year, with no increased labor costs, no increase in workers". It isn't realistic and won't happen, it's a thought experiment. SLS is taking years to construct, and the best estimates are 6 to 12 months per rocket after all development is complete and everyone is up to speed. It isn't realistic to assume they will be able to +50% or +100% their pace with no increased labor or supplier costs. I don't know why we're entertaining this idea, and we certainly shouldn't base our infobox figure on it. The NASA spokesperson quoted in Ars Technica was clearly trying to damage control and proffered this idea to minimize the painfully high cost number, by essentially saying "we might be able to improve our processes in the future so we can make more rockets at the same labor cost" (is my interpretation, see here for the source). Leijurv (talk) 19:05, 20 July 2020 (UTC)
 * Sure, I do agree that Marginal cost is absolutely not needed in the infobox, like any other rocket on Wikipedia. This was a proposition to "appease" those who wanted to cite this number. If anything marginal cost of SLS might be a totally irrelevant metric since unlike Saturn V or the Space Shuttle, SLS is planed to fly less than once a year for the next decade and even those rockets don't cite it!
 * The best solution would be to have the launch cost as the exact quote from the OMB report : Over $2 billion per launch once development is complete
 * I don't know what you exactly mean by "Launches per year"? The number of launches per year?
 * For the "Program cost per year" it should somehow include the ground system cost like every other rocket or at least mention that it is excluded. - Moamem (talk) 00:57, 21 July 2020 (UTC)
 * I don't know what you exactly mean by "Launches per year"? The number of launches per year? Yes. :)
 * The idea being: if we hypothetically wrote something like "1 launch per year, $2B per launch, $2.5B program cost per year" the reader can realize that $0.5B is going to something other than launch on an ongoing basis e.g. development of the next booster engines or some such. I don't think it's a 100% necessity but I think it could clear up what's going where. Open to having my mind changed though.
 * I agree, it should say $2 billion, after development or something equivalent.
 * I saw no mention of ground systems. If launch support is not part of the program then where is it billed? Leijurv (talk) 01:51, 21 July 2020 (UTC)
 * I don't know about this "Launches per year", it's kind of an unknown. I prefer not to speculate here.
 * For "Program cost per year", we should use the current year or average, maybe both IMO.
 * For the "launch cost", what's the issue with whole quote? Too long?
 * You did not know about the Ground systems? Welcome to SLS creative accounting. Basically NASA is counting SLS ground infrastructure as a whole different project, almost $4.5 billions as of 2018. So basically every SLS $ figure you've ever seen does not include infrastructure cost like Mobile Launch towers and processing bays... That's true for the launch cost, ground infrastructure is not captured in that figure. I went through NASA FYI and made a table some time ago, basically $500 million a year, a billion in 2018, I'll update the Ground system wiki page with budget when I have the time. You can see the details of this here : https://en.wikipedia.org/wiki/Exploration_Ground_Systems - Moamem (talk) 05:20, 21 July 2020 (UTC)
 * That's a reasonable position.
 * Sure.
 * Yes, I think it's too long, and we could say the same thing in way fewer words (4 instead of 9).
 * That is so cute and creative! Looking at that article, what on earth does this mean? Unlike previous work focusing on a single kind of launch vehicle, such as the Saturn V or space shuttle, EGS is preparing the infrastructure to support several different kinds of spacecraft and rockets that are in development, including NASA's Space Launch System (SLS) rocket and Orion spacecraft for Artemis 1. uh.... I thought Orion, Artemis 1, and SLS are all going to be one launch? So this is.... still a single kind of launch vehicle, no? Leijurv (talk) 05:27, 21 July 2020 (UTC)
 * Yeah.... So the payload isn't included in the launch cost so the capsule and service module aren't part of the 5 billion per launch. If you are talking about a standard payload then it is entirely reasonable to exclude the payload cost from the launch cost.  If your payload is a few hundred pounds of meat then perhaps including the service module at $1 billion and the Orion capsule at $800 million are also reasonable.  Nasa also moved the costs of the boosters out of the SLS budget and got called out on it by the OIG.  The platform costs getting shifted away isn't quite as blatantly dishonest but not by much.  If all we ever launch are 9 crewed missions then the SLS Launch cost is probably north of 7 billion each. SandowTheHeretic (talk) 02:39, 22 July 2020 (UTC)
 * I was mostly talking about how the wiki article itself is calling SLS + Orion more than one kind of launch vehicle, I don't quite get it.
 * Nasa also moved the costs of the boosters out of the SLS budget and got called out on it by the OIG. Got a RS on that? Might be worth putting in the article. Leijurv (talk) 02:52, 22 July 2020 (UTC)
 * “NASA's estimates show the SLS program grew to costs of $8.75 billion (only 25% more than the original cost estimate from the beginning of the SLS program, or baseline), but the agency had removed $889 million in costs related to engines and solid rocket boosters, "because SLS Program officials determined those activities were not directly tied to Artemis 1," the rocket's first test flight, the inspector general said in the report. Because NASA did not "rebaseline" the program to account for the lower cost, this "masked" the growth of SLS, the report added.” This was an article about the 3/10/20 oig report. I looked at the original but couldn’t quickly find it. https://www.space.com/nasa-sls-megarocket-cost-schedule-oig-report-2020.html SandowTheHeretic (talk) 15:11, 22 July 2020 (UTC)
 * Nice! Probably worth putting in the Criticism section of the actual article? Could probably just go for it tbh. Leijurv (talk) 17:07, 22 July 2020 (UTC)
 * Is this the original report? Leijurv (talk) 17:09, 22 July 2020 (UTC)
 * So while I do agree that less is better, in that case the 5 extra words are quite important :
 * "Over" means that the $2 billion is the MINIMUM estimated price tag, and doing away with it might confuse people into thinking it's the actual estimated price tag.
 * "Once development is complete" means that unlike every other rocket on Wikipedia, SLS Launch cost does not include dev costs. It's an important info since those cost are over $20 billions and climbing.
 * As for the "creative accounting" while one could argue that SLS could potentially launch something other than Orion, it is clear that removing the SLS ground systems is a laughable attempt at trying to lower SLS price tag! - Moamem (talk) 02:59, 22 July 2020 (UTC)
 * would still convey that this doesn't include dev costs, no? Leijurv (talk) 03:04, 22 July 2020 (UTC)
 * I don't see an issue with it. I still prefer the other language since it has the added benefit of being a direct quote. But this is fine - Moamem (talk) 03:14, 22 July 2020 (UTC)

Hey guys, I am not one to defend SLS "creative accounting, but can we please focus on the estimates at hand? I think that calculating End Of Life costs of SLS is a fascinating topic I don't think that this is the right place to have this discussion.

I urge you to settle on a consensus language that has few speculation and is clear and precise.

I reiterate my position Launch cost : Over $2 billion per launch once development is complete, and either separating or eliminating the Marginal cost (since no other use this figure).

Can we at least agree on that?

Thanks - Moamem (talk) 03:12, 22 July 2020 (UTC)


 * Yeah, you are correct. Let’s get it done. “Over $2 billion after development” is the shortest sensible wording. Marginal costs should be removed. SandowTheHeretic (talk) 15:11, 22 July 2020 (UTC)
 * Yep! Leijurv (talk) 16:39, 22 July 2020 (UTC)

Hi everyone, After months of debates I think a clear consensus has emerged. I went trough the different opinions given in the last 9 months (since November 2020) and tried to compile them in a manner useful for the dilemma at hand :

"1) Keeping the $500 million launch cost figure :

For : Jadebenn

Against : Leijurv, Maomem, 174.251.160.98, SLSgal, Fcrary, N2e, Timlograsso, Zegfred, Skytie, TheSkalman, 2001:56a:7797:e800:acd7:1c51:cba5:6365, RundownPear, NguyenVy1993, Beaucouplusneutre, Sun Creator, SandowTheHeretic

Unclear : Soumya-8974, SkywalkerP"

"2) Separating Marginal Launch Cost ($900 million) and Launch Cost ($2 billion) - Votes for eliminating the marginal cost completely will be considered as a vote for separating them :

For : Leijurv, Maomem, 174.251.160.98, SLSgal, Fcrary, N2e, Timlograsso, Zegfred, Skytie, TheSkalman, 2001:56a:7797:e800:acd7:1c51:cba5:6365, RundownPear, NguyenVy1993, Beaucouplusneutre, Sun Creator, SandowTheHeretic

Against : SkywalkerP, Jadebenn

Unclear : Soumya-8974"

If you spot a mistake or you have been inadvertently miscategorized please say so and I will correct accordingly.

If you want your username added please state the proposition (1 or 2) and you position (for or against).

If a clear consensus persists I will make the changes in 24h.

Please do not debate here!!!! This is specifically for the tally. Any debate should take place in the sections above! Thanks! - Moamem (talk) 01:52, 14 July 2020 (UTC)


 * This is now the third ping notification from this. I believe that one was sufficient. Every time you re-add it, it sends another notification ping. Perhaps just link to users instead of pinging again. Leijurv (talk) 03:22, 14 July 2020 (UTC)
 * Can you give me an example I use this syntax : [ [ User: Leijurv | Leijurv ] ] - Moamem (talk) 03:27, 14 July 2020 (UTC)
 * I don't quite know how it works, I just know that I got a ping each of the three times you added this. Perhaps you could simply write out just: Leijurv. With no linking. Leijurv (talk) 03:30, 14 July 2020 (UTC)
 * Me too. However, I am currently unclear with the inclusion of $500 million (aka proposition #1 unclear). --Soumya-8974 talk contribs subpages 04:37, 14 July 2020 (UTC)
 * Oh sorry! When you ironically (I guess?) said "Therefore, we should use "$500 million–$4 billion (est.)" on the infobox, right?". I thought you were advocating removing the $500 million figure. Corrected! - Moamem (talk) 05:58, 14 July 2020 (UTC)

Coverage of Cost in the Article Itself
As N2e said, ... the biggest problem is that infoboxes should, per wiki guideline I believe, restate info that is already well explicated, and well sourced, in the prose in the article body. It seems that this article, as it currently stands today, is trying to do all the heavy lifting in the Infobox itself ... each statement says no more than the source supports, with adequate context, without undue emphasis on any one of the many cost estimates that exist, maintaining neutral point of view. Then, the infobox can simply summarize the range of launch costs that is supported in the article prose.

While the discussion on the cost stated in the infobox is well intentioned, the scope is simply too limited. The truth is that this article's discussion of cost is nebulous and confusing to follow. The article has references to costs for the future in History sections, there are figures that are pulled from unrelated proposals, and the price that is referenced throughout the article doesn't even include the cost of assembly and launch. These are glaring errors that don't follow WP:UNDUE, even creeping into WP:NPOV territory. I think sorting out the article could greatly inform what the infobox says, as a summary / culmination of the article itself. I'd like to start a discussion on this, and I'll kick it off with a lot of complaining!

Let's jump into a concrete line-by-line of these issues:

First, the price estimates of the SLS are splintered throughout  and. These costs are explicitly mentioned to be future costs, and yet their placement betrays that by implying that the price estimates for a craft that is yet to launch have been firmly established. This issue is further compounded by the conflation of the  and   sections, where cost estimates are sprinkled indiscriminately throughout both. Take, for example, the two following points:
 * In 2013, the Space Review estimated the cost per launch at US$5 billion, depending on the rate of launches
 * On 1 May 2020, NASA awarded a contract extension to Aerojet Rocketdyne to manufacture 18 additional RS-25 engines with associated services ...

Why is the former in the, and the latter in the  ?

Now the next critical issue is one of basic source veracity: The $500M figure pulled in as the lower bound for the cost per launch is, in the best of faith interpretations, aspirational. This unicorn figure was sourced from a concept study report for a project that is slated for 2035, and explicitly suggested by NASA headquarters — The very same entity who has even today explicitly refused to make any conjectures as to what the real price of an SLS launch might be. In summary, the $500M figure was given to a mere concept study on a project to be launched in 15 years by the very agency that refuses to give a number, and this number just happens to be the lowest number cited anywhere in this article. Let's stop citing everything and start citing with relevancy in mind.

Then we have contradictory sources with no critical analysis on either of them, served as though they are of completely equal stature. In the  section, the reader is confronted with In May 2019, NASA's Office of Audits ... A letter from the White House to the Senate Appropriations Committee revealed. We need to do legwork here and assert to the reader that these two sources are not even, because they're not.

Claims internal to NASA standing on their own should be handled with extreme caution. When even internal figures are hopeful and unrealistic and they have refused to offer any official statement regarding the cost of the SLS, then any further claims are dubious at best. And so, to level that their claims should be considered with the same level of trust given to a report with the highest possible level of criticism written by the White House OMB, which measures the quality of agency programs, policies, and procedures to see if they comply with the president's policies and coordinates inter-agency policy initiatives? At that point, to maintain that these sources are equivalent in their trustworthiness begins to enter WP:FRINGE.

And still in that very same section, we have absurd and irrelevant cost comparisons to a completely different space program that is entirely unrelated to the SLS program. Nestled between the offending "false source equivalence" highlighted above, it's stated that By comparison, a Saturn V launch cost roughly $1.23 billion in 2016 dollars. Not only is this completely unprompted and undue in this section, but it's flat out misleading and effectively wrong. It's the Space Shuttle program, not Saturn V that is being replaced by the SLS. The article is abundantly clear to this effect, stating so in the section lead: SLS is intended to replace the retired Space Shuttle as NASA's flagship vehicle. Given this, why are we comparing the cost of the SLS to that of the rocket that flew from 1967 to 1973. We should looking at the Space Shuttle which flew from 1981 to 2011. This should absolutely be changed to By comparison, a Space Shuttle launch cost $576 million so that we actually make an apples to apples comparison of the cost of launch (where both costs are excluding the cost of development).

With the critical issues above in mind, this paragraph is unacceptable in its current state. It juxtaposes NASA's unreasonably low estimate with the cost of a rocket that has no direct connection to the SLS and last flew 47 years ago, and creates a bad false equivalency of sources. These history sections are a deeply insufficient smattering of wildly varying cost estimates or guesses from numerous different programs and departments, all with no connecting thread, no glue, and no commentary on the sources.

Of the seven sources cited in the infobox for cost per launch, NONE are cited or explained in the article.


 * This is the major meta problem. I think you've summarized it well.  Just a bunch of random costs/prices, mixing the two sometimes, cited in infoboxes is sure to lead to Infobox wars, as it has.  When the war heated up, and got to more issues about editor behavior rather than issues about improving the article, I just quietly backed away to let the dust and debris from the editor conflict settle.  N2e (talk) 10:54, 5 August 2020 (UTC)

Taking our first infobox source, why don't we fully state in the article something like The NASA Administrator shared in December 2019 that he hopes to reduce costs over time for SLS by negotiating with suppliers, and believes they will end up getting to $800M to $900M per launch, but can't say for certain yet.? There's nothing wrong with putting statements in context and giving fuller quotes. I'm sure everyone would agree that we can't just say Bridenstine said SLS will cost $800M to $900M and hide the rest in a ref or note. Doesn't mean we can't say it at all, with proper context given to the reader!

Why is there no discussion in the article about Europa Clipper? The controversy over its launch vehicle ended up exposing the painfully high figure (which we do actually state A letter from the White House to the Senate Appropriations Committee revealed that the SLS's cost per launch is estimated at blah blah), but why don't we give more context about this being from Europa? How is it possible that this letter is not cited in Criticism? the use of a commercial launch vehicle would provide over $1.5 billion in cost savings. The Administration urges the Congress to provide NASA the flexibility called for by the NASA Inspector General and consistent with the FY 2020 Budget request. To reiterate, this could very well be the highest level possible criticism, coming directly from the White House Office of Management and Budget, which, again, measures the quality of agency programs, policies, and procedures to see if they comply with the president's policies and coordinates inter-agency policy initiatives. But this quote is not discussed. We only pull out the $2B figure from earlier. What??!?!?!?

Costs to assemble, integrate, prepare and launch the SLS and its payloads such as Orion (funded under the NASA Ground Operations Project,[102] currently about US$400 million[6] per year) Sorry?? What?? Almost can't believe what I'm seeing here. Why do we drop this bombshell in a unassuming bullet point mixed with silly and obvious things, after "SLS costs don't include previous rockets" and before "SLS costs don't include its payload". How can we possibly justify sliding in like that the idea that the SLS program cost we use throughout the article doesn't include the cost of assembling, preparing, and launching SLS??????

There are no current NASA estimates for the average costs per flight of SLS, nor for the SLS program recurring yearly costs once operational. In 2016, the projected annual cost for Orion, SLS, and ground systems was US$2 billion or less.[107] NASA associate administrator William H. Gerstenmaier has said that per flight cost estimates will not be provided by NASA.[108] I'm just thinking "yes, and?". NASA doesn't want to give estimates. So? Is Wikipedia bound by some mysterious NASA NDA? We can absolutely provide estimates of per-flight cost, and we should provide more, from more secondary sources. If NASA won't provide estimates, and leaves us with only independent or external sources (even critical ones), then so be it?? Obviously, we should keep this passage in the article for sure, but we shouldn't use it as a shield to hide behind and say we don't need to provide external coverage of this because NASA doesn't want to.

One last thing: I want to be clear that I am not pulling WP:DTTC or anything. I think we can truly explain each of these figures in the article text in a manner that explains why we arrived at whatever figure ends up in the infobox. And we can probably do so without stating any of the sources are wrong, but rather outdated, or internal, or not inclusive, or too inclusive, or anything. I see this as closer to WP:CONFLICTING than controversy.

When several reliable sources conflict, some of those sources can sometimes be demonstrated to be factually erroneous, thus resolving the conflict. If NASA announced that a given program will spend X million dollars on a SLS launch, at some point we have to realize that the figure is absurd. Say X was a hundred. Would we put that in the infobox? Say X was ten. Or one. Sometimes, when sources conflict, we do have to judge them against each other in a manner that does take into account what they're saying and how realistic it is. In my eyes, when considering a rocket that has cost $20B already and will cost $2.5B per year to launch once-ish per year, it's perfectly reasonable to believe $2B per launch. That same rocket spending $400M on the rocket engines, per launch? I don't know much about rocket engines, but that sounds about right. That same rocket costing $500M per launch? It just seems impossible, ONLY because of the preponderance of other reliable sources that allow us to judge that this simply doesn't add up. This isn't OR or SYNTH, it's just WP:EDITDISC. If WP:SYNTH prohibited this, there would be no possible solution to WP:CONFLICTING!

As WP:EDITDISC says, It is not original research to make judgement calls on what content to include or not include, how to frame an issue or claim, or what claims and subjects are suitable for Wikipedia. We are not here to robotically compile facts and citations according to a strict set of rules, we are here to create and edit an encyclopedia. This task requires the application of judgement and discretion in order to create a neutral and readable encyclopedia. The policy on original research is sometimes misconstrued as a blanket prohibition on any application of judgement or critical thinking by editors. The intent of that policy was never to turn editing into an unthinking task, and our articles into mere compilations of published data.

Basically, I fully agree with N2e that we're putting the cart before the horse by nailing down the infobox on its own set of sources without even using or explaining (let alone contextualizing) those sources in the article itself. Leijurv (talk) 06:48, 22 July 2020 (UTC)


 * You are of course correct. We need to revise the entire article for coherency.  The info box is a critical point of contention because it is what most people will see. Go google "cost of sls launch".  Right... that is why.  Frankly it is hard to read as much on this project/launch vehicle without seeing gaping holes and conflicts.  It is hard to move forward with that while an edit war rages.  If Jadebenn has backed off his absurd position (and I haven't seen anything from him in a few days) then I think we can move forward.  We have a critical mass of people here that have each read more than any rational person would on a subject they don't make money off of.  If we want to do an overhaul, I think we can make short work of it with a little division of labor. SandowTheHeretic (talk) 17:32, 22 July 2020 (UTC)


 * I don't think Jadebenn is gone, just taking a break (see here).
 * Yes, there are certainly issues here. Concretely, what do we think should be rearranged? I ranted above at length, but I probably missed some things. There's definitely a balance to be struck between an unannotated and unexplained infodump list of numbers from sources all across the board throughout the past and future, and hand holding the reader exactly through what to think and feel. Currently we're too far towards the former I fear. Leijurv (talk) 18:50, 22 July 2020 (UTC)
 * By the way: I'm going to sit on this until someone who doesn't completely agree with me weighs in :) I believe that will make the article better than if I just go in and change everything. But if that doesn't happen I'll go for it in a week or two, per WP:SILENCE. Leijurv (talk) 20:15, 24 July 2020 (UTC)
 * I'll try and write up a response to this soon. Just letting you know that I've seen it. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 17:59, 29 July 2020 (UTC)


 * I very much agree with the overall approach suggested by Leijury. That approach ought to be to explicate the costs, and cost estimates, and prices quoted to third party payload builders to get a launch (which may be priced well below cost) and do so all with WP:NPOV, avoiding WP:UNDUE, and ensuring all prose is well sourced.  N2e (talk) 10:54, 5 August 2020 (UTC)


 * As I've been gone from the discussion for quite some time (and the length has gotten quite long), can you clarify what you mean? – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 07:34, 9 August 2020 (UTC)


 * Okay, I don't know what happened there. I agree addressing things in the article text should come first. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 08:00, 9 August 2020 (UTC)


 * I've read your argument (sorry for taking so freaking long) and I really do not understand it. The core of your argument is that the lower figure is impossible. Now why would the OIG and several NASA missions be citing impossible figures? You're unilaterally declaring a source incorrect and using misinterpreted figures to that end. For instance: RS-25s don't cost $100M a pop, the cost of a decade-long development, production, testing, staff, and integration contract (that also produces new engines) just happens to arrive at that figure when you divide everything by the output. The fundamental aspect is that NASA does both development and production at the programmatic level. This is reflected in their contracts. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 07:38, 9 August 2020 (UTC)
 * I do fundamentally agree the article text could use improvement, though. The issue is that there's clearly a disagreement on how these costs are defined, and I'm not sure that can be fixed by putting them into the main text. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 08:00, 9 August 2020 (UTC)
 * Sorry for spamming your notifs like this (it's very late where I am and I'm tired), but one final point I'd like to make is the non-linearity of yearly program costs when it comes to contracts. For example, if a multi-billion SLS block buy contract were signed tomorrow for SLS cores through the 2030s, some of those yearly program funds will be going towards paying for that every year (because NASA doesn't have the budget flexibility to pay for multi-year contracts lump-sum). This would lead to a situation where the cost of each of those SLS cores is effectively counted more than once in our infobox figures. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 08:06, 9 August 2020 (UTC)
 * First, I'd like to say that I strongly disagree with merging this section into the RfC. It twists the topic of the RfC into a general discussion about the article, and muddies the threading. Specifically, the main line of what we were talking about was here (I believe that's where the meat of it is), which used to be (more or less) around the end of the RfC, but now this is just a intermediary section heading. I'd very much like to put the sections back how they were.
 * Nevertheless, I'll also respond here.
 * I've read your argument I'm noticing (both a few weeks ago and now) a bit of yanking out one point that's easiest to argue against, while ignoring the rest. It makes me feel as if only one portion is being read and the rest ignored. This comes up as I just link to past things I've said that were never replied to.
 * I really do not understand it I'll try and explain further.
 * The core of your argument is that the lower figure is impossible. No, not really at all. I've mostly been calling into question everything about the $500M source other than the specific number, except in this one case. Because, one part of the point I'm trying to make down here in this section was that editorial discretion is Something That Exists, and to that end, I used an example that sometimes when sources contradict one of them is just plain wrong, or saying something completely different. To demonstrate this, I asked rhetorically if a program budgeted $1 million for an SLS launch, does that mean SLS costs that much? The answer is clearly no.
 * I hope I can fix this miscommunication by very clearly stating what I think of that $500M figure. It appears to me, that you believe that SLS costs $500M to launch and a NASA program, while writing a concept study report for a hypothetical space telescope, was advised that their budget proposal should allocate $500M for an SLS launch in 2035 are interchangeable statements. They are not, at all. Can you help me understand why I see such a huge disconnect between those concepts and you don't?
 * I'll start by listing three immediate things: 1. This is a primary, internal, non-independent, conflicting source. This team has every incentive to downplay the budget. I don't know why this could be considered a RS, it's just an internal report? It isn't independent of the subject at all. We could absolutely use it in the article, so long as we tell it like it is: something like A concept study report for a 2035 space telescope was advised by NASA HQ to budget $500M for an SLS launch 2. We have no idea how this number was arrived at. We don't know if the figure is an actual present-day value, or a guess of what SLS might cost in the future aspirationally. It could be a offhand order-of-magnitude guess, like the cited Bridenstine source for the 800M to 900M idea. 3. Last but not least! On top of all that! As I've said several times now, If the "SLS team" doesn't bill the "Europa team" for labor, does that mean the labor is free? Nope! It still occurred, and NASA, the parent organization, is still footing the bill. To make this fully concrete: even if this came to pass exactly as this concept study report imagines, and the Origins space telescope pays $500M of its budget to SLS, I don't think it follows that the SLS cost of launch was $500M. This is for two reasons: these are internal politicized numbers, and no part of NASA is operating at a profit in intent or in practice. I covered this here. This is just accounting magic; no money is changing hands; it's all being cut from the same NASA budget pie. And as we know, NASA loves to shift around this pie inventively and creatively. The OIG report was super interesting this year, see this source. I'm also interested in what you think of the sheer irony that the cost to, and I quote, assemble, integrate, prepare and launch the SLS (emphasis mine) is excluded from all our cost figures. Cost of launch not including cost of launch. Does this bug you?
 * (and regarding that source, the part I'm talking about is this paragraph: NASA's estimates show the SLS program grew to costs of $8.75 billion (only 25% more than the original cost estimate from the beginning of the SLS program, or baseline), but the agency had removed $889 million in costs related to engines and solid rocket boosters, "because SLS Program officials determined those activities were not directly tied to Artemis 1," the rocket's first test flight, the inspector general said in the report. Because NASA did not "rebaseline" the program to account for the lower cost, this "masked" the growth of SLS, the report added.)
 * Now why would the OIG and several NASA missions be citing impossible figures? Strawman.
 * I don't see several missions, or an OIG. I certainly didn't mention any, just the one with the $500M figure. The phrase "seems impossible" was used in the sense of: seems impossible in comparison to the other sources that I'm also looking at. I invite you to reread. Please remember the distinction between WP:CONFLICTING and WP:SYNTH. I don't believe or accuse the Origins team of citing impossible figures. They are, more likely than not, just stating future figures. Could SLS get down to $500M per launch by 2035? I could see it happening, if volume increases drastically and contractor relations are completely overhauled. The part that seems impossible is applying the number to the present day/year/decade.
 * The other cited source for $650M is also for 2035 by the way, and it gives even less indication of how the number was arrived at (instead of "we asked HQ and they said X" we only get a "we assumed X" - this is another internal report, why would we consider this assumption reliable?)
 * You're unilaterally declaring a source incorrect and using misinterpreted figures to that end. It isn't incorrect. I completely believe what the source is saying! It's completely correct and believable that the Origins space telescope was advised by NASA HQ that their 2035 launch deserves $500M of budget in their project concept study report. No argument there. The part that I'm declaring incorrect is where you make the leap from there to the idea that is what SLS costs. That's just not what the source is saying at all. I also like the unilaterally while the sentence very clearly stated that it isn't unilateral, and is based on the other sources: It just seems impossible, ONLY because of the preponderance of other reliable sources that allow us to judge that this simply doesn't add up.
 * I'm also curious what you think I'm misinterpreting, just that rocket engine stuff? Let's talk about that!
 * For instance: RS-25s don't cost $100M a pop, the cost of a decade-long development, production, testing, staff, and integration contract (that also produces new engines) just happens to arrive at that figure when you divide everything by the output. All right! That sounds really complicated. (I'm just using this source for the rest of what I'm about to say). So, the RS-25 engines have cost a lot to develop, they have to restart production lines, lots of fixed costs and such. But they already did this in 2015! This is incredibly convenient and this allows us to actually take a look at the marginal add-on cost of adding some more engines. This should give us an accurate figure for cost per engine going forward! So, in 2015 there was a contract to make six more RS-25 engines for ~$1.7 billion. That's very expensive per engine. Makes sense though, because they said Aerojet Rocketdyne has restarted the production lines, established a supplier base and is building engines using advanced techniques that reduce both the cost and time for manufacturing each engine. Then, five years later, NASA asked for even more engines. This is a classic marginal cost! The original contract for the six isn't done yet, this is purely add-on work for making new ones. The source says: This includes labor to build and test the engines, produce tooling and support SLS flights powered by the engines. So what's the cost, and for how many engines? On 1 May 2020, NASA awarded a contract extension to Aerojet Rocketdyne to manufacture 18 additional RS-25 engines with associated services for US$1.79 billion. Oof.... That looks a lot like $100M each, on the margin, doesn't it?
 * The idea that you're arguing against would be if I took the $3.5B total cost and divided it by the 24 engine contract and arrived at $146M. The Ars Technica article did do that and arrived at that number, which might be where the confusion came from, but I did not state the 146*4=584, I used the correct figure (times four engines per core stage), and said: That same rocket spending $400M on the rocket engines, per launch?
 * does both development and production at the programmatic level This doesn't mean we get to throw up our hands and say that we can't extract meaningful numbers. It also doesn't mean that we get to throw up our hands and say that weird internal budgeting stuff is the best we can do. See what I said here and here.
 * I'm not sure that can be fixed by putting them into the main text It absolutely can! I really do think we need a down-to-earth and clear coverage paragraph in the article about cost. We have it already basically, it's just fragmented into sentences scattered throghout the entire article :)
 * ...a situation where the cost of each of those SLS cores is effectively counted more than once in our infobox figures... Given that SLS hasn't launched yet, this may not be apparent, but after Artemis 1 it should be quite clear to the reader that "Program cost to date" will certainly include the cost of the rockets built to date. I don't see a problem with the cost per launch being included in Program cost to date, it seems completely natural to me. What is the alternative? Do we subtract cost per launch out of program cost...? I don't see any problem with what you describe in that paragraph. Leijurv (talk) 18:59, 9 August 2020 (UTC)
 * I've read your argument I'm noticing (both a few weeks ago and now) a bit of yanking out one point that's easiest to argue against, while ignoring the rest. It makes me feel as if only one portion is being read and the rest ignored. This comes up as I just link to past things I've said that were never replied to.
 * Fair. It is difficult, however, to respond to every point you raise. I tend to single out the argument I find most pertinent if only because the alternative is much more laborious. I will try and be better about it this time.
 * No, not really at all. I've mostly been calling into question everything about the $500M source other than the specific number, except in this one case. Because, one part of the point I'm trying to make down here in this section was that editorial discretion is Something That Exists, and to that end, I used an example that sometimes when sources contradict one of them is just plain wrong, or saying something completely different. To demonstrate this, I asked rhetorically if a program budgeted $1 million for an SLS launch, does that mean SLS costs that much? The answer is clearly no.
 * ...I do not quite understand. Are you arguing the figure is just plain wrong and NASA is not budgeting correctly?
 * I hope I can fix this miscommunication by very clearly stating what I think of that $500M figure. It appears to me, that you believe that SLS costs $500M to launch and a NASA program, while writing a concept study report for a hypothetical space telescope, was advised that their budget proposal should allocate $500M for an SLS launch in 2035 are interchangeable statements. They are not, at all. Can you help me understand why I see such a huge disconnect between those concepts and you don't?
 * Not really. The source for the $2B is an OMB policy document. Technically, no-one really 'knows' what an SLS launch costs yet, because it hasn't. They're all estimates.
 * I'll start by listing three immediate things: 1. This is a primary, internal, non-independent, conflicting source. This team has every incentive to downplay the budget. I don't know why this could be considered a RS, it's just an internal report? It isn't independent of the subject at all. We could absolutely use it in the article, so long as we tell it like it is: something like A concept study report for a 2035 space telescope was advised by NASA HQ to budget $500M for an SLS launch 2. We have no idea how this number was arrived at. We don't know if the figure is an actual present-day value, or a guess of what SLS might cost in the future aspirationally. It could be a offhand order-of-magnitude guess, like the cited Bridenstine source for the 800M to 900M idea. 3. Last but not least! On top of all that! As I've said several times now, If the "SLS team" doesn't bill the "Europa team" for labor, does that mean the labor is free? Nope! It still occurred, and NASA, the parent organization, is still footing the bill. To make this fully concrete: even if this came to pass exactly as this concept study report imagines, and the Origins space telescope pays $500M of its budget to SLS, I don't think it follows that the SLS cost of launch was $500M. This is for two reasons: these are internal politicized numbers, and no part of NASA is operating at a profit in intent or in practice. I covered this here. This is just accounting magic; no money is changing hands; it's all being cut from the same NASA budget pie. And as we know, NASA loves to shift around this pie inventively and creatively. The OIG report was super interesting this year, see this source. I'm also interested in what you think of the sheer irony that the cost to, and I quote, assemble, integrate, prepare and launch the SLS (emphasis mine) is excluded from all our cost figures. Cost of launch not including cost of launch. Does this bug you?
 * Okay, so you are arguing that internal NASA programs are not being costed fairly. That clears some things up. I have some issues with the other parts of your argument, though. For one, a White House budget proposal isn't exactly an "independent document" either. Again, explicitly policy. For the other, the OIG stated in their Europa Clipper report that the cost of an SLS launch is $876M. Why does that not meet the qualifications, but the $2B figure does? Even if you want to drop the $500M as being "too speculative," the OIG has no motive to distort the cost.
 * That sounds really complicated. (I'm just using this source for the rest of what I'm about to say). So, the RS-25 engines have cost a lot to develop, they have to restart production lines, lots of fixed costs and such. But they already did this in 2015! This is incredibly convenient and this allows us to actually take a look at the marginal add-on cost of adding some more engines. This should give us an accurate figure for cost per engine going forward! So, in 2015 there was a contract to make six more RS-25 engines for ~$1.7 billion. That's very expensive per engine. Makes sense though, because they said Aerojet Rocketdyne has restarted the production lines, established a supplier base and is building engines using advanced techniques that reduce both the cost and time for manufacturing each engine. Then, five years later, NASA asked for even more engines. This is a classic marginal cost! The original contract for the six isn't done yet, this is purely add-on work for making new ones. The source says: This includes labor to build and test the engines, produce tooling and support SLS flights powered by the engines. So what's the cost, and for how many engines? On 1 May 2020, NASA awarded a contract extension to Aerojet Rocketdyne to manufacture 18 additional RS-25 engines with associated services for US$1.79 billion. Oof.... That looks a lot like $100M each, on the margin, doesn't it?
 * No, NASA's doing long-term design changes. As you mentioned earlier, they like to spread the pie. So the RS-25Es come first, then the RS-25Fs, etc. The contract extension is not just for engines. It's not even just for engines and the personnel and facilities required to make them (if it 'were', I'd be more inclined to agree with you). It's not even just for the engines, personnel, facilities required to make them, SEI, and human-rating testing campaigns. It's for all that and any future developments and design changes as part of the R&D effort to simplify and streamline RS-25 production. Program cost? Absolutely. Launch cost? No.
 * Given that SLS hasn't launched yet, this may not be apparent, but after Artemis 1 it should be quite clear to the reader that "Program cost to date" will certainly include the cost of the rockets built to date. I don't see a problem with the cost per launch being included in Program cost to date, it seems completely natural to me. What is the alternative? Do we subtract cost per launch out of program cost...? I don't see any problem with what you describe in that paragraph.
 * ...I'll concede this last one. Serves me right for arguing at 4 am.
 * One last thing,, Sandow did an improper manual close of the RfC. That's why I reverted your edit. If it had been the bot, I wouldn't have. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 20:57, 9 August 2020 (UTC)
 * Okay, nevermind. Seems Legobot did a sweep a couple days back and whacked away the RfC, so wasn't acting in bad faith by adding the close template. Sorry, bud. – Jadebenn (talk &middot;&#32;contribs &middot;&#32;subpages) 21:08, 9 August 2020 (UTC)
 * BTW I'd still be interested in a reply to the main discussion thread here.
 * Are you arguing the figure is just plain wrong and NASA is not budgeting correctly? Those two ideas aren't the same thing. As I say later on, the figure could be correct. That could be how much the Origins team needs to give up from their budget. That doesn't mean that their launch costed that much. For example, my point about Exploration Ground Services. If NASA shatters the cost of launch into a million different places and does some inventive budgeting (e.g. the boosters thing I mentioned from OIG), that doesn't mean we have to go along with it and present the cost of launch number that they'd like us to. Basically, just It isn't incorrect. I completely believe what the source is saying! It's completely correct and believable that the Origins space telescope was advised by NASA HQ that their 2035 launch deserves $500M of budget in their project concept study report. No argument there. The part that I'm declaring incorrect is where you make the leap from there to the idea that is what SLS costs.
 * And again, it was a rhetorical question. But maybe you disagree with my rhetorical answer...? If NASA came out tomorrow and said that Europa Clipper gets to launch on SLS but only has to give up $1M of their budget to do so (thus making SLS a much cheaper option for them than a commercial provider), is the cost of launch for SLS now $1M?
 * The source for the $2B is an OMB policy document. Technically, no-one really 'knows' what an SLS launch costs yet, because it hasn't. They're all estimates. That's right, they're all estimates. But some are better than others. Making an estimate that cost of launch will be equal to a budget sleight of hand, based on an internal rebudgeting for 2035, is pretty much objectively worse than a present-day estimate from an external agency which measures the quality of agency programs, policies, and procedures to see if they comply with the president's policies and coordinates inter-agency policy initiatives, because that external estimate is more all-inclusive and resilient to said budget trickery. I'm almost at a point where I'd say that any launch cost that excludes Exploration Ground Services is inaccurate. Cost of launch not including cost of launch is absurd. We don't have to concede to their budget trickery.
 * Okay, so you are arguing that internal NASA programs are not being costed fairly. That clears some things up. This should not even be an argument, it's a fact! NASA is doing some really weird stuff with their budget and estimates. See here. It's gotten so bad that the OIG is calling them out: Seek more transparency in current, future and overall cost and schedule estimates for SLS and other human spaceflight programs by reviewing NASA program management policies and the Human Exploration and Operations Mission Directorate. and Create a cost-accounting model that would separate each SLS "deliverable" to make it easier to track costs and performance and award fees. and NASA's cost tracking does not show how much the delays are affecting the program's baseline, the report added. At the end of fiscal 2020, NASA will have spent more than $17 billion on SLS, which includes $6 billion "not tracked or reported as part of the ABC," the report stated. and The study examined how well SLS is meeting its cost and scheduling goals, including how NASA is tracking and reporting these goals and how contracts are being managed. The authors of the report found that NASA "continues to struggle managing SLS program costs and schedule," struggles that "can be attributed to challenges with program management, technical issues and contractor performance." and NASA's estimates show the SLS program grew to costs of $8.75 billion (only 25% more than the original cost estimate from the beginning of the SLS program, or baseline), but the agency had removed $889 million in costs related to engines and solid rocket boosters, "because SLS Program officials determined those activities were not directly tied to Artemis 1," the rocket's first test flight, the inspector general said in the report. Because NASA did not "rebaseline" the program to account for the lower cost, this "masked" the growth of SLS, the report added. Also see this.
 * For one, a White House budget proposal isn't exactly an "independent document" either. It is independent of NASA as it helps guide what NASA's entire budget should be. It is external, and measures the quality of agency programs, policies, and procedures to see if they comply with the president's policies and coordinates inter-agency policy initiatives.
 * The OIG stated in their Europa Clipper report that the cost of an SLS launch is $876M. I covered this in the very first post I made on this talk page, diving into all the sources... here A NASA Office of Inspector General (OIG) report in May played down the cost savings between SLS and commercial alternatives largely because it offered a much lower cost estimate for the SLS. That report estimated an SLS launch at $876 million, versus $450 million for commercial alternatives. However, in a follow-up letter in August, OIG said NASA could save up to $1 billion by launching Europa Clipper on a commercial vehicle versus the SLS, adding that NASA needed to decide how to launch the mission in the next few months in order to procure a launch vehicle in time to support a potential 2023 launch. I don't think we can say SLS costs $876M if the cost savings from switching away would come out to over $1B, unless a commercial launch would cost negative money. Or, the better explanation (in my eyes), is that $876M was how much SLS wanted of the Europa budget, but the figure doesn't include everything else relating to the cost of launch that would still be spent throughout NASA. And the figure appears to have been superseded by later analysis, both internal (OIG $1B) and external (OMB saying $1.5B cost savings).
 * The contract extension is not just for engines. It's not even just for engines and the personnel and facilities required to make them (if it 'were', I'd be more inclined to agree with you). This doesn't line up with what the source says, at all? The same reliable engines that launched more than 100 space shuttle missions have been modified to be even more powerful and The follow-on contract to produce 18 engines is valued at $1.79 billion. This includes labor to build and test the engines, produce tooling and support SLS flights powered by the engines. This modifies the initial contract awarded in November 2015 to recertify and produce six new RS-25 engines and brings the total contract value to almost $3.5 billion with a period of performance through Sept. 30, 2029, and a total of 24 engines to support as many as six additional SLS flights. This seems to directly contradict what you're saying, the source literally says labor to build and test the engines, produce tooling and support SLS flights powered by the engines.
 * The bot did remove the RfC template though... Diff here. Removing expired RFC template. (strikethrough added because edit conflict) Leijurv (talk) 21:42, 9 August 2020 (UTC)


 * I have decided not to participate in this conversation since you, Shadow and other are doing such a good job, but I had to intervene for this one : The reason for $876 is very simple, it's marginal cost, and everybody else is talking about normal, everyday, dumb cost! NASA just decided to start quoting marginal cost as cost without ever telling the public. But be sure all those under a billion prices are Marginal costs. It's just missing over a billion in fixed costs... It's as simple as that! - Moamem (talk) 23:17, 9 August 2020 (UTC)
 * Ah yes. The whole "marginal" idea. Yet, the hypothetical of "what is the cost increase if we built and launched one more" is just not the same as "what is the cost increase if one more of these was built and launched, with no need to hire additional labor, in the exact same timeframe (no extra time taken)", so I'm a little hesitant to describe cost numbers as purely "marginal". Leijurv (talk) 04:13, 10 August 2020 (UTC)

It's been almost four weeks since the participants in this discussion on "Coverage of Cost in the Article Itself" had anything to say about it. I just found the details here, and read through it, today.

I continue to believe it is of critical importance that the whole "cost issue" of SLS be explicated in prose, with neutral language, and not saying more than a reliable source can support in the body of the article. Per WP:MOS, the infobox "summarizes key features of the page's subject" and all of that which it summarizes should be explicated in the article prose itself. It should not be left only in the Infobox, despite the good sourcing now and the hard-discussion that many editors participated into get to that infobox-cost consensus.

So where are we on this knotty topic with this particular article on the Space Launch System? Seems like a lot of good summarizing occurred above, but then the discussion just ended without any clear resolution. N2e (talk) 11:43, 5 September 2020 (UTC)
 * By this point, I'm of the opinion that the most neutral thing to do would require actually describing the level of budget trickery that's going on, as background. Primarily what I just said here, pulling from sources like this and this. We should tell it like it is: this is a very expensive and far over budget project, and a LARGE portion of the cost sources we have come from entities under every incentive to downplay the cost. Simple as that. Example 1: The literal cost of launch itself has been farmed out to something more general called NASA Ground Operations / Exploration Ground Services or some such. I think we absolutely need to fold that into any "cost of launch" number that's anywhere near reality, even if it's just written in prose (I'm imagining a sentence literally saying something like: "launching SLS will also cost $400M per year for NASA Ground Operations, which only exist to support SLS and its payloads, but are budgeted separately") Example 2: That stuff with the boosters: NASA's estimates show the SLS program grew to costs of $8.75 billion (only 25% more than the original cost estimate from the beginning of the SLS program, or baseline), but the agency had removed $889 million in costs related to engines and solid rocket boosters, "because SLS Program officials determined those activities were not directly tied to Artemis 1," the rocket's first test flight, the inspector general said in the report. Because NASA did not "rebaseline" the program to account for the lower cost, this "masked" the growth of SLS, the report added. Please correct me if I'm wrong, but I interpret this as follows: NASA budgeted $7 billion for SLS, including the solid rocket boosters (of course it includes the literal rocket engines of the rocket). NASA pulled out the cost of developing those SRBs from the SLS budget and put them somewhere else. Now SLS appears to be cheaper, and much closer to the original budget. But the budget was not adjusted to account for the removal of this very large cost (was not "rebaselined"). The end result is that 12% of the original SLS budget's value has been pulled away to another cost center, without decreasing the budget. That's amazingly shady. Because NASA did not "rebaseline" the program to account for the lower cost, this "masked" the growth of SLS, the report added. Yeah. Example 3: Quoting ridiculously low figures to other internal NASA teams for flights 15+ years in the future, with the end result that those absurd numbers make up a large portion of the cost figures available to the public. Leijurv (talk) 02:12, 12 September 2020 (UTC)