Reply To: Star Citizen – Musings
DEV SCHEDULE UPDATE (10/26/2018)
STATE OF THE ENGINEERING FARCE
In my previous article, I outlined just how much of a disaster the much touted 3.3 patch was turning out to be – even for a pre-Alpha build. And so far, it’s been going precisely as predicted. Aside from the fact that they came up with this pledge (complete with a new dev roadmap format) a year ago to release a major point build at the end of each quarter and which they’ve basically missed every time. In fact, the only time they ever made a deadline, was by cutting stuff out of the build, and moving it into the next build. Yay! We made the deadline!
And they did precisely the same thing months ago in which they had 3.3 and then 3.4. Next thing we know (as I wrote here), mere weeks ahead of the annual whale milking that is CitizenCon, they split 3.3 into two parts (3.3 and 3.3.5). And just like that, they were totally going to be releasing (to live) 3.3 during CitizenCon. Then they made an entire spectacle out of 3.3.5 (which btw didn’t even have the much-touted Object Container Streaming enabled) during CitizenCon.
Of course that live release didn’t happen. What did happen was that after a brief stint in Evocati (see: special snowflakes) test, they released it to the Public Test Universe (PTU) for any brave backer to download and “test”. FYI, the dev schedule goes like this: Dev -> Internal Test -> Evocati -> PTU -> Live
I could go on and on about the numerous reports of how completely broken everything is in 3.3, but I’m not going to bother with that. What I want backers to bear in mind is that for over 18 months CIG has been touting two pieces of technology (Object Container Streaming + Network Bind Culling) which they claimed would solve many of their performance issues. I have written numerous articles explaining why that’s patently false, and that even as more content is added, this “tech” will only serve to be a bandaid over a gushing wound. While there are some extolling the virtues of the “increased” performance in 3.3, it’s not only the usual disingenuous nonsense that is the pattern of deceit surrounding this project, but also the extents that “believers” will go. Let’s face it, if you’re standing around on a server in an unpopular location taking screen shots or just pissing about, it’s easy to see a +5 fps gain due to assets not previously loaded. Let me explain in layman terms as best as I can.
Games with lots of content tend to “demand load” them as-needed. What the player isn’t seeing or interacting with, doesn’t need to get loaded or rendered. It’s not rocket science – and every engine tends to do this automatically, while giving the devs some control over it. In the loading of content, you can do it either in whole or by streaming it as-needed. Here is an example of how it works in Unity3D, and how it works in UnrealEngine. This technique can be used for any type of content (level, textures, meshes etc). The challenge is that when it comes to determining what does get loaded and rendered, bad design tends to yield unexpected results. In the simplest form, there is no point in loading parts of the world the player can’t see or interact with, let alone any assets associated with it. If the player is at a location with a view distance of 20km, loading anything beyond that view distance is a waste of resources and leads to major performance and memory issues. With Star Citizen, not only does it load unneeded assets regardless of where the player is, but the problem is further compounded by the high fidelity (model meshes & textures) of the assets in question; thus resulting in huge memory usage as well as performance issues. Both of which have plagued the project since day one, and which got exponentially worse with the Dec 2017 release of 3.0 (which introduced surface tech in the form of moons).
MINING ADVENTURES IN 3.3
If you were getting 15 fps in some area in 3.2, then you’re now getting 20 fps in 3.3, that’s not only a superficial gain, but it also highlights the problem of diminishing returns. Any additional content that gets added is not only going to reduce any perceived performance gains, but also will behave as if no gains previously existed if you went from 15 fps to 20 fps and back to 18 fps as things progress and new content is added. Amid all that, have you been to Levski lately? Especially when there are more than 8 clients there? Did you notice the single digit fps? So what good is a performance gain if it’s not consistent where gameplay matters. It won’t matter if you’re not intending to play the game, but instead just taking screen shots to show off your chariot and a broken and unattainable dream.
It’s all a farce; but hey as they say, beauty is in the eye of the beholder. My guess is that at some point, they’re going to have to either disable OCS and release 3.3 live without it, or just keep kicking that 3.3 can down the road. There is no way they are going to release it live in the current state, let alone release the 3.3.5 branch when they have so many performance and crash issues in 3.3. Funnier still, with 3.3 still in the PTU, one month past it’s Q3 release, according to CIG, they are totally going to release 3.3, 3.3.5 and 3.4 in Q4/2018. I can’t wait, because it’s almost as if they rushed out 3.3 for CitizenCon hype.
WELCOME TO DA GRIND, PEON
IN-GAME SHIP BUYING AND DA GRIND
We didn’t see this one coming because it wasn’t even in the dev schedule – at all. During CitizenCon, it was announced that ship buying was coming, along with REC ship rentals in Arena Commander. The latter was already in the schedule; but then out of the Blue a few days ago, UEC ship buying was enabled. So far the response from the tribe has been epic and hilarious.
Here’s where it gets interesting. CIG makes money from selling ships at hundreds and thousands of Dollars. In fact, funding chart inaccuracies aside, they made almost $4m from CitizenCon spearheaded by ship sales, one (Valkyrie) of which is in 3.3 (released during CitizenCon of course), and the other (Kraken) being a JPEG concept which, like dozens before it, won’t be in the game for quite some time. As they have already pre-sold all interesting ships to the two hundred or so whales who keep giving them money, that’s why they resort to adding new ships (and requisite scope creep) such as the Kraken in order to bring in new cash money. So they have no incentive to allow lucrative in-game ship buying, thus undercutting their primary revenue stream. Hence the ridiculous pricing. The cash (pledge) money and UEC (in-game currency which can be bought for cash) spend are a revenue source, unlike earning ships by grinding for in-game UEC.
3.3 IS LOOKING REAL GOOD!
Now comes the other side of the hilarity. There are currently only three methods of earning UEC in the game. Wait, it’s funny, trust me on this.
- Doing combat engagements against NPC units
- Cargo runs in which you pickup a box and deliver it somewhere. This requires a ship (prices start at around $125) which can collect cargo. The starter ship which comes in the $45 base package won’t work
- Mining on a moon or asteroid. This requires a ship with mining capabilities. The game has three mining ships, the cheapest and only one (implemented for 3.0 in Dec 2017 when surface moons was released) is the Prospector which costs $155
None of the above is going to average more than 4000 UEC per hour of gameplay. With #1 you’re going to die – a lot because combat sucks, the game crashes a lot etc. In #2, if you don’t crash, or die in transit, or the mission doesn’t complete, the most you’re going to make is about 2000-3000 on average. In #3, not only is it tedious enough that watching paint dry has more appeal, but in addition to the crashes, or possible death from other players, you have to deal with a rubbish method of “refining” your minerals which you then have to sell. And the yield is so hilariously low that mining for a living is just one of those things you do when you can’t find a freshly painted wall to stare at.
Most of the whales who have spent hundreds of Dollars on ships, are also collectively freaking out that in-game ship buying with UEC is finally in. What are they pissed about? Well these chuckleheads who have been clamoring about how their money is in support of the game, that it’s totally not P2W etc, are suddenly now faced with the reality that the ships they spent so much money on, and which made them feel special, can now be gained by lesser peons who don’t have to pay anything. So, get this, they’re saying that the UEC amounts are too low. I wish I was joking.
THE CITIZENCON BOUNTY WAS PLENTIFUL THIS YEAR
CLOUD IMPERIUM GAMES (UK) FINANCIALS
The UK parent of the group of companies, has finally filed it’s financials. You read my coverage of the financials for RSI and for F42. Also note that CIG is the parent for the group of companies, and it acts as the US and UK publisher (funding source) for F42 (US, UK, GER) which are making the Star Citizen and Squadron 42 games. As of YE 2017, all companies in the group were insolvent, when taking into account that they had less than $1M in the bank, against a significant debt load, and over $2M in monthly operating costs.
As usual, the latest filing is fraught with errors, omissions, and curiosities. Aside from the EU entities using up 64% of annual revenue (obtained from the highly suspicious public funding chart which is shockingly consistent year on year) for 2017, from these numbers it’s easy to see how the world-wide monthly burn rate for the project is probably around the $3m per month mark where the UK/GER group alone burned through almost $17M (!) of total revenue.
There is currently no financial insight to the US side of things where there are studios in CA and TX; but development notoriously costs way more in the US than it does in the EU. On page 21 (under “Bridge Crew” section) of this Oct 2017 article below, Chris Roberts points out that dev/labor costs in the US are double that of the EU. And he’s not entirely wrong because that’s how it has been traditionally; which is why most US companies outsource if/when they can.
“We try to be quite smart about development costs, so we do a lot in the UK and two-thirds of our developers are in Europe. It’s far more cost effective. Over here you can have two developers for the price of one in the US. In the places where there’s game development in the US, the price of living is really high. We’re up in Manchester and it’s a lot cheaper to live there than in LA. The average salaries in the industry are less for that reason.”
It’s not just the cost of living, though: “We get basically 25 per cent of the UK cost back from the government. And that allows us to hire more people. We wouldn’t have as big an office in the UK if that deal wasn’t there. I think that was a very good move for the government to do that, because now we have around 250 in the UK, by far our biggest group of people” – Chris Roberts, Oct 17 2017
Also, Erin Roberts (brother of Chris Roberts) one of the highest paid game devs in the Manchester area, and who previously sold his F42-UK (studio built with backer money) shares back to the group, increased his pension plan contribution in 2017. In fact, he doubled it. They’re not even hiding the audacity anymore.
With 2017 being a pivotal year, we won’t know more until around this time in 2019 when the 2018 financials are filed; but my money is on the continued drop in operation commitments in the US, with primary focus being in the UK/GER. I wouldn’t be surprised at all if they continued to gut both US studios by 2019, as they have been doing since 2017.
So once again a UK accountant has done the break-down of the filings. I will just quote it below, with minor edits and corrections for clarity and accuracy.
EMPLOYEES
Going to start with something straight forward that is easy to interpret but always controversial. No layoffs. CIG dramatically increased their employee count yet again. These figures must include employees based in Germany and any other countries hidden around the world, but not the USA.
Reminder:
- Period ended 31 Dec 2014 – Average 52 employees
- Year ended 31 Dec 2015 – Averaged 132 employees
- Year ended 31 Dec 2016 – Averaged 221 employees
- Year ended 31 Dec 2017 – Averaged 318 employees
Total salaries rose and the average rate of £43k per employee (including employer’s National Insurance) is broadly in line with previous years. There is however a hilariously obvious and simple mistake in the notes of the accounts. You really don’t need to be a German auditor to spot this one and yet…
Here’s the origin of the very basic error in the accounts but I cannot explain how it was missed.
2016 FILING. NOTE THE 2015 AMOUNT
2017 FILING. NOTE THE 2016 AMOUNT
THE COUTTS LOAN
Since these accounts cover the year ended 31 Dec 2017 and the Coutts loan was taken out around the middle of 2017, the balance outstanding at the year end was £1.54m. There is a little note in the accounts about it which does not entirely make sense but it’s there.
BEYOND THE UK GROUP
One has to bear in the mind the very obvious and simple mistakes that are often littered in the CIG accounts but I think something fundamental has changed with the Byzantine Empire.
Up until December 2016 the US group acted as a sort of de facto parent company of the UK group. They held all the money and booked expenses, did some sort of expense recharge with the UK group. I always held reservations about how this was supposed to work with regard to actual accounting practices. I think during 2017 a number of things have happened…
- Cloud Imperium Rights LTD incorporated on 29 Aug 2017
- CIG UK LTD (The UK parent) increased its investment in unlisted companies from £440k to £462k a precise increase of £22,205. Presumably CIR LTD above
- The UK group is no longer declaring its income to be related to RSI Corp (et al) in the USA
- The note to the accounts which analyses turnover by geographical market, which I pointed out was clearly incorrect, has now changed
CONCLUSION
I would have to assume that the story now is that we are to believe that any cash given to CIG in pledges or subs etc is directly split. Any money given by persons in the USA is given directly to the US group of companies and used there. Any money given by the rest of the world is funneled to the UK/rest of the world group of companies.
It’s a subtle difference but it makes the UK group appear to have actual customers and turnover, rather than handouts from a related company in the USA. It would have been a basic requirement for the Coutts loan and make more sense from a tax credit angle. It gives a plausible reason for the creation of yet another company, though there is no suggestion that the new company is active yet.
It would also mean for example that the US corporate structure would no longer show numbers such as say, $34m income, $17m US expenses, $17m UK expenses and would instead show say $17m income, $17m expenses. Which might be of some interest if your US companies were fighting a lawsuit.
This whole thing is a bit of a tax dodge as well as some shenanigans. Previously one company (US), was booking all the income and receiving most of the cash. Then there is a myriad of “expenses recharged” and such. Which is not normal accountancy language or behavior. Normally it’s just revenue and expenditure. These companies were seemingly listing it as revenue and expenditure but simultaneously making notes about calling it recharged expenses.
This causes some issues when you are looking to take out a loan. If one of the companies at the bottom of the pyramid wants to take out a loan and they only have one customer and that customer is a related party, it raises many questions. Why doesn’t the parent company/related party at the top of the pyramid take out the loan?
Furthermore, there are hoops to jump through when receiving tax credits. In order to qualify for the UK video games tax credit, the company in receipt of those credits (Foundry 42 Ltd) is supposed to be solely in charge of production and distribution of the qualifying video game (Squadron 42? Star Citizen?). Obviously the UK government has repeatedly bent the rules and been flexible to allow it but there’s a certain level of questioning, that the old Roberts empire could not really even come close to fulfilling.
If anything can really be taken away from all of the accounts as a whole, it is the complete lack of consistency. Every year they seem to make up something new. We’ve had a different source of funds from different US companies (CIG Corp and RSI corp). We’ve had a different destination of funds from the US companies (CIG LTD and RSI LTD). We’ve had different methods for calculating how much funds should be transferred, including one year having to restate the previous financial statements to recalculate. Now we have a different method of funds from the US. Transfers covering recharged expenses vs presumably a revenue sharing agreement.
3.3 PTU GAMEPLAY SESSION
FOUR YEARS AGO
As Nov 2018 rolls around, this is a reminder that back in Oct 2012 Chris Roberts started a $2M crowd-funding campaign to build two games, Star Citizen (multiplayer) and Squadron 42 (single-player w/ co-op play). He later moved the campaign to Kickstarter where he raised $2.1M from myself and 34,396 other gamers.
The games were supposedly going to be released by Nov 2014 – something that Chris Roberts in lots of interviews stated was totally going to happen because if the development took longer than two years….
“Really it is all about constant iteration from launch. The whole idea is to be constantly updating. It isn’t like the old days where you had to have everything and the kitchen sink in at launch because you weren’t going to come back to it for awhile. We’re already one year in – another two years puts us at 3 total which is ideal. Any more and things would begin to get stale.” – Chris Roberts, Oct 19 2012
Having discovered a lucrative way to monetize the otherwise dreaded scope creep, by Nov 2014 they had raised $65M without shipping either of the games promised. As of this writing, they have raised $198.7M, even as the scope creep continues unabated. Not a single game is anywhere near complete.
By all accounts, the entire project is an unmitigated disaster that continues being financially propped up by a number of whales who have the opinion that throwing good money after bad is a totally sound decision, even as each release serves to highlight the glaring truth that the project is doomed to fail and they can’t possibly build, let alone deliver what was promised.
Aside from that, they are facing a devastating lawsuit which if they lose (as most of us believe they most certainly will), will completely destroy the company, and backer money along with it.
“My brother is a big of Star Citizen. As an non-Game Designer he thinks they are doing a lot of pioneer work for game development, testing more things and taking more risks than other devs thanks to the crowd-funding. Would you say this is in some way true or is it just that Cloud Imperium Games is more open about their development process?” – Askagamedev response
FURTHER READING
Previously: CitizenCon 2018
How I got involved in this farce