Phil Mason Does Not Understand Space

Phil Mason, a chemist who makes science education videos under the pseudonym ‘Thunderf00t’, has recently been producing some videos on the topic of space that contain very large amounts of misinformation, and misunderstanding of the topic. He has a special hostility for Elon Musk’s space plans – perhaps enhanced by the clicks a video mentioning Musk or SpaceX can gather. I’ve held off on this for a while, as I don’t want to be very negative, but I have seen enough people cite him as an authoritative source that I feel some debunking is in order. Given that he himself is somewhat aggressive in his purported takedowns of others ideas, very keen to laugh at what he views as stupidity, and plastering the word “BUSTED” over his video thumbnails, this seems like fair game.

Eppur Si Muove

Before moving on to his attacks on SpaceX, I want to give a demonstration of Mason’s credibility in this field. He and his supporters make much of him being a scientist by training to add weight to his takedowns. However, it is easy to demonstrate that outside his chosen field, he really is not authoritative at all.

In one video 3 years ago, he goes after a news segment about a proposal for a skyscraper hanging from a tether in space, attached to an asteroid. Let me just say that this is a silly idea; the architect who proposed it appears to have lifted the concept directly from the pages of Neal Stephenson’s hard sci-fi novel Seveneves, and then stripped it of the context in the novel that made it make sense. Most of the criticism Mason directs at this is on the mark, although given that this was a throwaway story about an art project designed to get an architect some publicity I am not sure it needed to be taken apart with such anger. Did anyone actually believe this was seriously going to be built, and needed to be talked out of it?

But in the course of valid criticisms, Mason makes an absolutely astonishing physics blunder. In discussing the practicality of bringing an asteroid to Earth to act as a counterweight, he attempts to calculate the delta-V required to move an object from Ceres to Earth. Leaving aside the fact that there are asteroids much closer to Earth, and the fact Ceres has a fairly eccentric orbit, lets look at how he does this. Here is a transcript from about 4:35 into the video:

“I mean lets just take a look at the big numbers. If you’re in the asteroid belt – and just so we’re clear the asteroid belt is the region between the orbit of Mars and Jupiter. So lets just take the largest object in the asteroid belt, that’s Ceres. To escape from the Sun, from the orbit of Ceres, you need about 25km/s. If you’re on the Earth, you need about 42km/s to escape from the Sun. So, this being a state function and all, If you want to get from the Earth to the asteroid belt, you need a change of velocity – a delta V – of about 17km/s”

This is a fundamental mistake. The solar system described only by its escape velocity at each radius is not a state function. The phrase is meant to claim that the property in question completely and uniquely describes each point in the system, and this is nonsense. At any point in the solar system, there is an escape velocity – but specific objects also have their own velocity and thus kinetic energy that has to be considered a part of the system. The fact that his calculation involves velocities is a red herring; escape velocity is the situation where kinetic energy equals gravitational potential energy and so really the values given here for Ceres and Earth are descriptions of the Suns gravity well, not the velocity of any celestial bodies. You certainly can’t just subtract one from the other and come up with a delta V for a transfer orbit, because you are moving from one orbit to another, not one fixed position to another. What this calculation actually gives you is the velocity required for an object floating in the solar system but not orbiting to push radially outwards to a greater distance.

So Mason literally forgot that planets orbit the Sun.

The correct way to do this calculation is to work out an elliptical orbit that contacts the two bodies in question, and then use the vis viva equation to work out how fast an object on this orbit will be moving when it is close to the two planets. Then comparing this to the velocities at which these planets orbit the Sun, you get the delta-V for the arrival and departure burns. Here is the calculation briefly:

These values don’t seem to different to Masons, do they? The velocity required to maintain a circular orbit is only small than escape velocity by a factor of the square root of 2 (~1.4) so most velocities in these kinds of calculations occupy a similar range. The real difference comes when we take into account that both the origin and destination are moving. Earth orbits at around 30km/s whereas Ceres orbits about 17km/s. This gives a total required delta-V of around 10km/s.

Note that this doesn’t deal with the matter of leaving the gravity wells of Earth or Ceres – but neither does Masons calculation. The point is not that he made a mistake – due to the simplifications used, my calculation is not even entirely correct – its that he made a deep conceptual error and didn’t notice. He waded into a field outside his expertise and made a fool of himself.

It may seem uncharitable to pick apart this error – but Mason has done the exact same to others. When another YouTuber made a mistake of not understanding what delta-T meant in terms of thermal expansion, he made a huge deal of mocking this, and even dug out his targets PhD these to try and discredit him. In thinking that have two velocities allowed him to calculate a delta-V in the sense it is meant in orbital dynamics, Mason made a very similar type of error – and frankly I have treated him a lot gentler than he has treated those he has attacked.

Trying to Dunk on SpaceX

It is quite evident that Mason does not like Elon Musk. He attacks the man at every opportunity, and claims every business he is in is fraudulent. His ill-informed comments about electric cars, which can be seen as ridiculous by anybody who has driven one, will have to wait for another day. Here I am simply going to discuss his relentless attacks on SpaceX.

In a somewhat confused video from 2019 attacking Mars One (a now defunct scam attempt to colonise Mars) alongside a TED speaker, and throwing in some excessive inferences from the BioSphere 2 experiment, Mason seems quite keen to link it all to SpaceX.

He is, perhaps intentionally, fudging a few dates here; he is confusing the timescale for a settlement on Mars (which Mars One made ridiculous claims about which he rightly dismisses) with the timescale for first boots on the ground according to SpaceX, which they currently aim to be in 2027. That an ambitious date to be sure, but not inherently ridiculous. Basically, because some not-very-credible claims that invoked SpaceX were made by other groups, Mason wants too claim that SpaceX themselves are not credible.

In this case, he grudgingly admits SpaceX has reduced launch costs but underestimates by how much (more on this later). However, he then goes on to show the most beat up returned booster he can find – I think its one that aborted a return to launch site and ditched in the sea, but could be wrong – and then repeatedly plays the video of a failed abort system test of a Dragon capsule. Given how this problem was resolved to NASAs satisfaction and Dragon has since carried two crews safely to the ISS, this segment has not aged well. He claims that “reusability at the cost of reliability is a false economy“ without any concrete evidence that reliability is low, and states that the Dragon capsule is “Just a clone of the Apollo capsule”. Aside from have better technology on board, carrying more people, being safer and reusable I guess that is true, although snarky and irrelevant. The main problem is what he does next.

Immediately after this somewhat unfair snark about Dragon, he immediately switches to news coverage of the SpaceX proposal for Mars exploration – which critically, does not use Dragon. Mars One showed an early version of Dragon on the Martian surface – but remember that SpaceX had no association with them at all. Much of the remainder of the video is like this; oscillating between claims SpaceX make for their upcoming Starship launch vehicle with a straw man version of their current capabilities.

In one segment, Mason states that Saturn V and Falcon Heavy are “essentially the same rockets” but the latter being half the price. An estimate of the per launch price of the Saturn V is $1.34 billion (adjusting for inflation from this article) for 140 tonnes to LEO (~$10,000/kg) whereas the Falcon Heavy in its most economical configuration (expend the core, reuse boosters) can lift about 50 tonnes to LEO for about $100m (~$2000/kg). So its 5 times cheaper, not 2 times cheaper – and that is the price SpaceX charge, not the cost to them. Starship promises to be cheaper still, but at this point Mason is pointedly ignoring that. This won’t be the last time he gets SpaceX prices wrong.

He then claims that projected ticket prices to Mars ($200,000 per person) are ridiculous because they would require a 10,000 fold reduction in cost and that this would be like promising to get the cost of a car down from $20,000 to $2. This is a very poor analogy – cars are expensive because they must be manufactured. The goal of Starship is to achieve 100% reusability – so nothing needs to be manufactured per flight. There are grounds to doubt that SpaceX can get the per flight costs low enough to pull off their plan – but noting that large manufactured goods don’t tend to drop in price 10,000 fold in a short period of time does not do that, as Starship will not require per flight manufacturing. Given this, what is offered here is just an evidence free appeal to ridicule.

More recently a video in January claimed that Mason has “BUSTED” using Starship as a point-to-point transport system on Earth.

There is a lot of snark that doesn’t contribute to the argument much. The superficial resemblance of SpaceX boosters landing to the DC-X landing in the 1990s is meaningless, especially given that a Falcon 9 booster has to deal with harsh reentry conditions, guide itself to a moving ship at sea, relight its engines, and perform a ‘suicide burn’ due to its high thrust-to-weight ratio. This is really not relevant though – some old stuff looks a bit like new stuff, so what?

The first substantive point Mason makes at around the 11 minute mark. He points to the use of pressure-fed rocket engines in the Apollo Lunar Module and Service Module, and how SpaceX rockets all use turbopump engines instead – claiming that this is why “nobody has ever really cracked launch failures”. This may be news to United Launch Alliance, a company founded in 2006, averaging 5-10 launches a year since then, who have never had a launch failure. Nor has SpaceX with the latest iteration of its Falcon 9 rocket (Block 5) although they have had some booster recovery failures. Mason declares that, contrary to these facts, failure rates linger at 1-5% because he appears to be aggregating all rockets in this stat, and to be blunt Chinese rockets are not as reliable as American ones. Attributing the supposedly inevitable failure rate to turbopumps, he then goes on to make an astonishing claim

“A single point failure on the booster stage and the entire rocket and its crew are lost. This just wasn’t true with Apollo”

This is shown with a shot of the proposed SpaceX Superheavy booster with its large array of engines. Seemingly, Mason believes that a turbopump failure in one of these engines entails a loss of the rocket. This is just not true, and has been shown to be untrue with flown rockets. The Saturn V rocket that launched Apollo 6 suffered two engine failures on ascent and one in orbit, and the vehicle was not destroyed. The one that launched Apollo 13 suffered an engine failure on its second stage, and the vehicle was not destroyed – the famous accident happened much later and was unrelated to the Saturn V rocket. During the STS-51-F mission in 1985, the space shuttle Challenger lost an engine due to a turbopump failure and was able to continue on to perform its mission through an ‘abort to orbit’. Last year a SpaceX Falcon 9 lost an engine on ascent and still completed its mission, although the booster was lost on recovery. SpaceX have demonstrated planned engine shutdowns of Raptors in flight on their Starship prototypes. They have also had two engines fail destructively on relight during SN8 and SN9 test flights – and their neighbouring engines continued firing until the test articles crashed due to insufficient thrust. Engine-out capability is a well established technology, and so painting a single point failure in a turbopump as dooming a spacecraft flies in the face of reality.

Mason attempts a cost comparison between an Airbus A380 and Starship, basically assuming that the cost is proportional to the fuel cost – not too unreasonable. The problem is, that this leaves Starship only 4 times more expensive than the Airbus, so he has to start fudging the numbers. He takes a fragment of an interview with Gwynne Shotwell where she says the first Starships will carry a hundred passengers, and then takes this to mean that Starship will only ever carry 100 passengers in a point-to-point configuration, which nobody has ever claimed. The size of the pressurised volume clearly means it can take more, likely a comparable number to the A380. Musk claims a $2 million launch cost is possible, which would be about $1 million of propellant at current prices and another $1 million ground operations. If 500 passengers can be fit in it (a low estimate) then each ticket would cost $4,000 one way at that price. Looking on flight websites now, a flight from London to Sydney economy class would cost me around $1,500, for a 35 hour trip with two stop offs. Paying just over double to do this in less than an hour would be a very attractive proposition to a lot of people. Of course, there may be grounds for claiming that this launch price cannot be reached – but Mason produces no such argument here.

Lots of the sources in this video are misrepresented by Mason. He tries to insinuate, based on an off the cuff remark by Shotwell about doing business trips to Riyadh and being home for dinner, that the only market for Starship point-to-point is business meetings which could be better done over Zoom. Later, he takes the below clip of Tim Dodd (The Everyday Astronaut) talking about abort systems and interprets it as an argument against Starship:

“So I guess the question should be, would I ride on Starship without an abort system. For now the answer is no. I think we should see at least a few dozen flights without crew first. We should find the limits and the boundaries, you know, maybe have some failures or two, and only once we’ve seen Starship fly 10+ times reliably without any failures would I consider getting on one.”

It should be clear to most that he just wants to see it tested rigorously (as SpaceX intend to do) before it takes humans. Mason then uses Dodd’s stats for the Space Shuttle and Soyuz rockets, combined with the spurious 100 passenger-per-flight number, to claim that on average one person on every Starship flight would die. This is like using the safety record of a DC-3 to make claims about how risky it is to fly on an Airbus A320neo. Much of the video is devoted to projecting failure modes of other spacecraft onto Starship, even if they make no sense. It won’t ever have a foam strike on its heatshield as it is not side mounted, and it won’t have a side booster recontact as it doesn’t have side boosters – these being the causes of the only manned spacecraft/rocket losses in the past 30 years. A proper risk model of Starship can’t be built in this manner. It needs to be done by someone who actually understands rocketry and statistics.

The Cost of Reuse

This crusade against SpaceX has culminated in his latest two videos, which are proving popular online with Musk haters.

There has already been a great debunking video by Shazmosushi showing why Mason is wrong about SpaceX prices in the first of thse two, using proper research on the matter.

I wanted to add to this some mistakes that were not covered in this video though, and there are plenty to go around.

The central claim of the video is that Falcon 9 is not significantly cheaper than the Space Shuttle and that SpaceX are just conning everyone with low launch costs. He starts of by claiming that the $50 million price for a reused Falcon 9 versus $62 million for a new one is a saving of “just slightly above 10%” when it is closer to 20%. This isn’t a big mistake – he makes those later – but its a taste of the kind of sloppiness he displays here.

There are a few segments of what Mason clearly considers his ‘greatest hits’ against Elon Musk, but those relate to electric vehicles and ‘hyperloop’, and are a bit out of the scope of what I want to discuss here. He only gets into actually trying to make a case against SpaceX at about 6 minutes in; so you can skip to that part (likely he is padding out his videos to service the YouTube algorithm, as many creators are forced to do.) At this point he introduces an article citing a 20 fold reduction in launch costs from Shuttle to Falcon 9, by Harry Jones of NASA Ames Research Centre, and then proceed to (in his mind) disprove this claim. As you might have gathered by now, he does nothing of the sort.

As Shazmosushi has covered the errors in Masons prices for Falcon 9 well enough, I shall look at what he did with the Space Shuttle. Mason uses multiple figures for the per launch cost, but mostly tries to use $450 million. This is, as far as I can tell, sourced from a NASA Q&A and is the marginal cost of a launch, not including fixed costs. An actual analysis published in Nature found that the average per mission cost between 1991-2010 was $1 billion in 2011 dollars

This is generous, as it excludes the costly development and ramp-up period, which added in would bring the per flight cost to around $1.5 billion. The figure above is useful as it shows both launch rate and cost, and where they are equal the cost for that year is exactly $1bn. Cheaper years in the mid 90s are balanced out by more expensive years and groundings, which can’t be excluded from cost calculations. Were Falcon 9 to be grounded, SpaceX would have to claw back the cost of that in future flights, so this is fair. This leads to SpaceX being 13 times cheaper in expendable mode, if you choose to ignore the large early costs of the Shuttle. This isn’t quite as good as the 20x improvement claimed in the original paper – but it has been arrived at by sweeping huge Shuttle costs under the carpet, choosing the most expensive Falcon 9 flight, and ignoring the difference between the costs of the Shuttle and the price of a Falcon 9. SpaceX is presumably making a profit on these vehicles, given their valuation as a company.

Mason’s main case here depend entire on throwing out kg to LEO as a metric, and replacing it with cargo and crew deliveries to the ISS. A Falcon 9 which in expendable mode can deliver 22 tonnes to LEO can only deliver around 6 tonnes to the station. However, Mason entirely neglects to do the same for the Space Shuttle, leaving it with its full, low inclination LEO payload for comparison. When a launch vehicle leaves the pad, it starts with a velocity due east imparted to it by the rotation of the Earth. The closer the launch azimuth is to east, the better this can help boost the spacecraft to orbit. Its not much, but due to the rocket equation it does add up to a significant payload reduction for higher inclinations. The ISS is at an inclination of 52 degrees, making it easier for the Russians to reach from their spaceport in Kazakhstan but harder for US spacecraft to reach from Florida. Mason appears to have been grabbing payload figures from Wikipedia, which are probably accurate enough for this discussion, but seems to have not noticed that directly below the figure he uses as its payload in all cases, 27,500kg , is the actual figure for payload to the ISS: 16,050kg

It gets even worse for the Shuttle though. This figure, about 60% of the payload to a lower inclination, is for the total amount that can be carried by the Shuttle. But the Shuttle is not a truck – you can’t just pile up crates in the back and fly to the station. To take pressurised cargo to the ISS, it had to use a Multi-Purpose Logistics Module (MPLM).

This was a cylindrical module carried in the payload bay, that was attached to the station during the mission using the Shuttle’s robotic arm. The average payload in one of these was just over 10,000kg, or 36% of the payload Mason is claiming for the Shuttle’s payload to ISS here. Having put his thumb on the scales several times, he then gets a figure of $20,000/kg for both Falcon 9 and Shuttle using the $450 million marginal cost and $60,000/kg for the Shuttle using the actual program costs.

Lets do this calculation honestly; each Shuttle delivered 10,000kg for $1bn and each Falcon 9 delivered 6,000kg for $133 million based on the total cost of the resupply contract from NASA. The figures are from 2011 and 2009 respectively so inflation is not really a factor. This gives $100,000/kg for the Shuttle and $22,000/kg for Falcon 9. At delivering cargo to ISS, Falcon 9 is 5 times cheaper in price, even before factoring in the profit SpaceX will be making. Next there is carrying humans. Mason assumes a crew of 7 on the Shuttle and divides his lower cost of Shuttle flights by this to get $65 million per seat as a cost. SpaceX charges NASA $55 million seat as a price. Using the actual cost of $1 billion, the Shuttle seat price is $148 million.

Now the Shuttle carried humans and cargo – which for space station missions is important, and does lead to the Shuttle being undervalued in some metrics. So a complete comparison for an ISS mission would be to take the cost of sending 10,000kg using Falcon 9, and add the cost of 7 seats, and then compare with the actual cost of a Shuttle launch. You would then get a cost to do exactly what the Shuttle did using Falcon 9 rockets of $605 million. SpaceX, with a far lower development cost, can reduce the price of doing what the Shuttle is absolutely best at by 40%. If SpaceX had any significant competition, they could surely get it lower – after all, where do you think the money for Starship development comes from? They are obviously making a profit.

Does this analysis support Mason’s point? Not really. The reason the Shuttle ended up just servicing the ISS (with the odd Hubble repair mission) is that it wasn’t good for much else. Hauling satellites to orbit is the main business of rockets and cost/kg to orbit, not to pressurised cargo and crew to ISS, is the key vital metric. This is why the paper Mason claims to refute used cost/kg to orbit and came up with a 20x saving. Whilst it is true that this metric does disadvantage the Shuttle – this metric is precisely what it aimed for in its original design. Even using the very specific metric of combined cargo and crew to ISS, Falcon still wins – albeit not by as much.

In the 1980s the Shuttle was driven from the commercial launch market by Ariane rockets, to the point where the US government didn’t want to subsidise the Shuttle to keep it competitive and more or less exited the market. Now the current incarnation of Ariane is itself being driven out of the same market by Falcon 9, as demonstrated in Shazmosushi’s video. SpaceX have also squeezed the Russian Proton pretty much out of business – and for legacy space, that was a very cheap rocket. Is Mason going to argue that the Shuttle was cost competitive with Proton?

On the subject of Falcon 9 reuse, Mason presents a spreadsheet he claims shows it can’t be economical. For a moment I will put aside that some of the models he presents here show breakeven points which some currently flying boosters have already passed (the most experienced on is currently on 8 flights and a couple have made 5 flights). I want to focus on a bit of numerical slight of hand during his presentation.

Screenshot from the video

The two inputs he uses are rocket refurbishment cost as a fraction of initial rocket cost, and reusable payload as a fraction of expendable payload. He acknowledges elsewhere that the correct value for the latter is 0.7, not 0.5, but continues to use that value anyway, and only varies the cost for reuse. Even then, as in the screenshot above, he can’t help but produce a vindication of SpaceX without realising it. But that isn’t the worst part. Here is a transcript of when he is discussing it at about the 19 minute mark:

“With the SpaceX numbers it’s not that you only take about 50 percent to low Earth orbit it’s nearer 70 percent but you can then add in the fact that the second stage is never reused… you… the minimum amount of cost for a re-launch is 20 percent…”

Whats happening on screen at the start of this transcript is his cursor is hovering near the top of the payloads to orbit column with the top ‘0.5’ selected. Then when he says ‘but’ he moves the cursor away, the spreadsheet fades to an old news story about proposed second stage reuse, and hes changed the subject. When he returns to the spreadsheet he doesn’t mention that column again. It seemed immediately obvious to me he didn’t want to put the real number into his model, and I had intended to remake the spreadsheet to show what would’ve happened if he did. A Twitter user named Parker beat me to the punch however:

As you can see, the correct value for the payload penalty of reuse changes things dramatically. Even with the unsympathetic value for refurbishment Mason used initially, breakeven occurs after 3 flights. That he knew this would explain the jarring cutaway he used to avoid typing that number into his spreadsheet.

In last bit of dishonesty in this video, he goes after one military contract and trips over his feet trying to argue that SpaceX prices are subsidised by the US government at about 24:25.

“SpaceX claims that their rocket launches are about $60 million a piece whilst they’re charging the military $316 million for a launch – but don’t worry, Gwynne Shotwell – you remember her – she’s got a great explanation as to why the price went up by a factor of six… … they have to account for a quarter of a billion dollars extra costs… Gwynne Shotwell ‘insisted that the company’s launch prices are not going up. SpaceX is however charging the government for the cost of an extended payload fairing’ – seriously, an extended payload fairing”

He then goes on to Google the cost of a SpaceX payload fairing ($6 million), ignores the possibility this extended fairing might need some R&D, and declares this a rip off. Except… he cut off Shotwell’s statement mid sentence. He showed a screenshot of the article, and zoomed in on the phrase “extended payload fairing” but on the way in the viewer can see the rest of the quote “…upgrades to the company’s West Coast launch pad at Vandenberg Air Force in California, and a vertical integration facility required for NRO missions”.

Mason seems to have noticed this, as in the next segment (which I suspect was edited in later when he realised what he had done) he offers dismissive statements about the cost of these two extra items:

“So what do we got – the cost of the extended payload fairing, upgrades to the company’s West Coast launch pad – the company’s West Coast launch pad? OK – um, and the vertical integration facility required for the NRO missions. Which I assume basically means they want a crane to load stuff on to the rocket.”

He didn’t even bother to Google it. A vertical integration facility is not a crane. Its a massive tall building with a lot of equipment in it. Here is a picture of ULA’s one. It could quite easily cost millions of dollars to build one, which is probably why the US government agreed to pay.

United Launch Alliance’s vertical integration facility. Not a crane.

Beyond the many errors though, there is a big elephant in the room here. The reaction of the launch market to SpaceX itself is very telling; commercial satellite operators have flocked to the company, and competitors such as ULA and ArianeSpace have announced new launch vehicles and their own reusability projects. Have all these companies been bamboozled by Elon Musk, and only Genius YouTube Skeptic Phil Mason can see The Truth? This is the domain of conspiracy theories.

Doubling Down

As I was writing this blogpost, Mason uploaded a second part to his recent video attacking SpaceX. It didn’t get much better.

He begins with a sarcastic ‘apology’ for not including inflation in his previous video’s figures, throwing snark at his critics for pointing out that he got something wrong, because he doesn’t believe it impacts the model. Mason clearly has a double standard for this sort of thing; in his world, it seems if someone else makes a mistake, its because they are a moron – whereas if he makes a mistake, the person who points it out is a moron because it doesn’t matter. Its a bit like an extreme version of the Fundamental Attribution Error.

Mason takes as a slam dunk against any criticism one statement from a NASA payload specialist, and a calculation using the artificially low cost of the Shuttle discussed above, claiming that Falcon 9 cargo to ISS is expensive. He ignores all other professional opinion on this because he has found the one guy who says he is right. He also presents this quote, inaccurately, as proving ‘costs to orbit went up’ which would not be true even if the calculation he cites were accurate, due to the difference between ISS delivered cargo and payload to orbit pointed out above.

The main body of this video is attempted to portray SpaceX abandoning propulsive landing for Dragon capsules as a failure, or a lie, or evidence of incompetence. In 2017 the company stated it was being abandoned because it was expensive to qualify for NASA astronauts, and because Musk no longer considered it the right approach to landing on Mars – Starship being well into its design phase at this stage. Perhaps they are lying and they couldn’t actually do it; but that seems very unlikely given that this same company has propulsively landed several iterations of Grasshopper and Falcon 9, and three Starship prototypes. Mason does not provide any evidence of deceit; he simply chuckles at how they said they were going to do something and then didn’t end up doing it, as if this in itself is evidence.

Then in defence of the fact that he completely missed the distinction between price for the customer and cost for the manufacturer, he went off on a tangent claiming that people wouldn’t want to put expensive probes on cheap rockets. He deflects a criticism with an entirely meaningless tangent rather than answer it. Theres a lot of memes and clips from his other videos attacking Elon Musk – lots of fluff to pad it out for the algorithms. He quibbles over the meaning of the term ‘human rated’ to cover for the fact he got it wrong in the first video, and the next substantive point is made at around 19 minutes, when he plasters a big red FAIL over Musks prediction (was this a prediction or a declaration of intent?) of sending Dragon capsules to Mars next to his own prediction from 2016 that reuse wouldn’t break even – which he clearly considers a success despite it being shown wrong by his own model in the last video (when you put honest numbers into it).

There is a lot less of substance in this video, but the main thing is Mason’s attempt to claim reuse can’t get costs down due to ground costs. He begins at around 19:20

“I couldn’t find any decent breakdown of launch costs but I did find this one from United Launch Alliance and it feels to be in the right ballpark…”

and then proceeds to demonstrate why you shouldn’t use your feelings to do analysis

“…give or take one third of the costs go to running the launch facility and the infrastructure and the personnel and all that sort of thing. Well if this is true then your SpaceX launch looks something like this. One third of it can’t be reduced because that’s what you’ve got to run anyway, and Musk says his second stage which isn’t reusable is about 20 percent of the cost of the rocket or thereabouts. Even if all you have to do is roll up and add more fuel to your rockets you can only reduce your costs by about another 50 percent.”

The problem here is that he once again doesn’t get the difference between fixed and marginal costs. Keeping the power on in those facilities costs the same no matter how many rockets you launch. Maybe the payroll goes up a bit with higher launch cadence, but it certainly won’t scale linearly. You pay engineers by the month, not by the launch. So the notion that ground costs are fixed is nonsense; the reuse costs (e.g. drone ships) that he tacks on without even giving specific numbers for also don’t scale linearly with launch cadence. Regardless of what feels in the right ballpark to Mason, these costs can change. They can also vary between companies, and ULA struggles to compete on price with SpaceX even for their expendable rocket, which is why they are phasing out all their rockets in favour of a new design (Vulcan).

It is also worth noting that his source is for a ULA is a military contract, which involves a lot of extra costs. Including those related to the vertical integration facility, which I cannot stress enough, is not a crane.

In what Mason seems to think is a final knockout blow, he plays footage of the 90s DC-X vehicle doing a vertical hop alongside the recent crash of Starship SN9, claiming they are essentially the same. Ironically, he actually undermines his own claim here without realising it. The clip of DC-X is long enough to show something quite important; its engines are firing the entire flight. This gives it a much shorter time and altitude limit than the Starship prototypes, and also critically meant that they didn’t have to figure out how to relight an engine during flight – trying to do this was the cause of both Starship crashes. Rocket powered landers are not new – what is new for Starship is being able to turn off the engines, fly in the ‘belly flop’ position’, and then flip and relight the engines. But Mason didn’t notice any of that, he just crudely reasoned by analogy.

Don’t Fool Yourself

Mason got into YouTube very early, with videos debunking creationists going back to 2007. His content really reflects that era – edgy, confrontational and frankly of low production quality. It feels the platform has mostly matured past this, leaving his channel as something of an anachronism. His goal in these “BUSTED” videos is to make someone else look stupid, so he and his audience can feel smart, which is also quite characteristic of the internet of the 2000s. Honestly, I enjoyed ‘pwning’ creationists back then too. But then I grew up.

In his race to get to the dopamine hit of making someone else look stupid, Mason is not careful in his own work. He focuses in tightly on anything he feels proves his point at the exclusion of everything else. He utterly strawmans other people’s positions and dismisses criticism out of hand. He does sloppy calculations that are biased towards the outcome he wants. Given that he is apparently a working scientist, he presumably can do better work than this when he wants to – but in these videos he behaves with all the objectivity of a addict chasing his next fix.

If he sees this and responds – I don’t especially want him to, as I have little desire to engage with him – then no doubt he will label me a “Musk Fan” and find some reason, however contrived, to call me dumb. This is just how he reacts to criticism historically. He very likely won’t admit any fault, and certainly won’t consider that him being a “Musk Hater” is clouding his own judgement.

Being mean and snarky doesn’t make you correct. It doesn’t directly make you wrong either – but the personality traits that lead you to do these things are also likely to sabotage your own reasoning. As Richard Feynman once said, “The first principle is that you must not fool yourself — and you are the easiest person to fool.” – and Phil Mason seems to have become very good at fooling himself.