Star Citizen – Year Four

[Dragunov]

Hi Derek,

 

I am also a backer of Star Citizen but luckily I only invested USD 65. I am also very suspicious about the whole project.

 

What I dont understand is why you always repeat that the development started 4 years ago and that 500+ people are working on it. From what I know that is not true. Maybe you should be more precise with these statements.

 

Even if the prototype was in development for one year starting 2011, it doesnt make much sense to add it to the whole development time. And the whole year 2013, when less than 50 people were working on it, shouldnt be added either. I would say Star Citizen is in development for 2,5 years (middle of 2013 till end of 2015) with a staff of 150 people.

 

Best regards.

[dsmart]

Sorry no, that’s not how that works. The prototype (proof-of-concept) stages all took time, money etc. So yeah, it’s 4 yrs. You don’t just ignore that, throw out the time, effort, money – just because you can.

That never – ever – happens in reality. That’s like saying when you’re about to build a house, the time taken to find the land, get the permits, draw up the architectural plans etc, should all be discarded because they are of no relevance to when they started the construction.

Chris himself has gone on the record as saying that “development” started in 2011. His words. Nobody made that up.

Also, here is a statement from Ben Lesnick back in Dec 2014.

“One important note: the footage you see in the trailer is in-engine, not pre-rendered! Chris put together a live Star Citizen demo for the announcement at GDC Next. Chris created the Squadron 42 demo by himself over the course of a year or so, with support from his old friends where needed (especially for the art, as the demo needed three ship exterior models, internals for the Bengal and one cockpit.)

This was not a game, or even something that would become a game – it was a proof of concept to show investors (and then the public) that the CryEngine could be adopted for the game we wanted to make. There was no privately funded game development period with different investors. We had investors lined up to support the project IF we could hit the crowd funding goals… but not before. There were no development studios or producers or schedules or network engineers or 3D artists or anything of that nature working on Star Citizen before the end of the campaign in November, 2012.

As for the overall speed of development today – we’ll always do our best to show you how it’s going. To be honest, I came in thinking exactly the same thing: how are we going to do these ten thousand things and finish what we’ve promised? Martin Galway explained it to me best: making a game is a lot like a roller coaster… it feels like you’re going up and up and up forever, seeming to have a minimal return. But that’s because the first half of the experience is building the systems that let the ride down happen. There’s a point in the development cycle where we hit the top and suddenly you see how the production team has plotted all the pieces to start coming together. I think we’re near the top of the coaster now – for the first time, I can look at the pieces and absolutely understand how Star Citizen happens.”

So, it’s either 4 or 3 years, depending on who is counting. And if they paid themselves for the work prior to Nov 2012 as has been rumored, then it’s all the more relevant to count that 1 year.

And from the original 2012 Kickstarter page:

We are aiming for a AAA game experience. But depending on the funding levels reached, we may have to limit the experience for the initially released game version. Nonetheless, Chris Roberts and his teams have shown consistently that they are able to develop epic story-based games. Even with our very limited self-funding we have been able to do already a lot of work which is why we can show you not just concept art and a cinematic trailer, but an extensive demo of actual game play. So, we are confident that even with limited means we will be able to deliver an amazing experience.

Oct 19th, 2012 Interview with Chris Roberts

“We’re already one year in – another two years puts us at 3 total which is ideal“

And the 500 head count is a number that Chris himself has already quoted numerous times. Many have come and gone, and right now that count, across four locations, is around 300 people. Just because there are less people now, doesn’t negate the work that those who have since left, had done.

 

[dsmart]

Some more thoughts on the Persistent Universe and whether or not the game is in fact an MMO. Read more about how I think they’ve partitioned the world to fit within their zones + instancing architecture.

First of all, right now, there is nothing persistent about 2.0. Nothing.

These are the facts:

Chris flat out said it wasn’t an MMO, right off the start.

2012 Kickstarter FAQ: Is Star Citizen an MMO?

No! Star Citizen will take the best of all possible worlds, ranging from a permanent, persistent world similar to those found in MMOs to an offline, single player campaign like those found in the Wing Commander series. The game will include the option for private servers, like Freelancer, and will offer plenty of opportunities for players who are interested in modding the content. Unlike many games, none of these aspects is an afterthought: they all combine to form the core of the Star Citizen experience.

2012 (pre-Kickstarter) RSI website pitch. Navy wings are made of Gold

Star Citizen brings the visceral action of piloting interstellar craft through combat and exploration to a new generation of gamers at a level of fidelity never before seen. At its core Star Citizen is a destination, not a one-off story. It’s a complete universe where any number of adventures can take place, allowing players to decide their own game experience. Pick up jobs as a smuggler, pirate, merchant, bounty hunter, or enlist as a pilot, protecting the borders from outside threats. I’ve always wanted to create one cohesive universe that encompasses everything that made Wing Commander and Privateer / Freelancer special. A huge sandbox with a complex and deep lore allowing players to explore or play in whatever capacity they wish.

…

Playable offline or online, co-op with friends, you sign up for a tour of duty with the UEE fleet, manning the front lines, protecting settlements from Vanduul warbands.

…

If you distinguish yourself in combat, you might be invited to join the legendary 42nd Squadron. Much like the French Foreign Legion of old, they can always be found in the toughest areas of operation and always snatch victory from the jaws of defeat, regardless of the odds.

…

Upon completion of your tour you’ll re-enter the persistent Star Citizen universe with some credits in your pocket and Citizenship to help you make your way. But in the universe of Star Citizen when one conflict ends, another is just around the corner. You’ll have opportunity to spend more time with your squadron mates as additional Campaigns are released as part of the content update plan.

RSI Forum comment Nov 2014 (<— ironic)

We have chosen Google Compute for our initial cloud implementation as we think its the best combination of power, price and flexibility. We are attempting to build a dynamic server system where local nodes can be spun up to handle the hi-fidelity server “instances” in areas that would help reduce the ping for people that are matched together.

Arena Commander is our test bed for this. When you join a multiplayer match you are currently connected to a game server by the matchmaking service. This server eventually will spin up on demand in an appropriate location to the people that the match maker has put together.

In the PU as you travel around a Star System (or jump from one to another) every time you come out of “warp” (or jump) you’ll be handed off to one of these server instances that will be spun up on demand taking into account where the people that have been contextually matched together are playing from. As we’re first prototyping / building on Google Compute this will naturally happen where there are Google Compute data centers. With some extra work we can fold other Linux Server Cloud providers into the matchmaking and server management. But it doesn’t make sense to do this before we’ve even finished the base system on Google Compute. Right now we spin up a fixed number of servers in the Google NA data center for the current multiplayer. One of the ongoing engineering tasks is to make this dynamic based on demand and then at different data centers around the world. Once this happens we would be ready to expand it to other cloud server providers if need be.

Then it became an MMO. Just like that. Why? Probably because the experienced engineers made it clear to Chris (who has no knowledge or experience building MMO games – ever) that his bullshit “dream” couldn’t be built without the architecture of an MMO.

About the game – Experience a first person universe

More than a space combat sim, more than a first person shooter and more than an MMO: Star Citizen is the First Person Universe that will allow for unlimited gameplay.

Then there was this: CHRIS ROBERTS ON MULTIPLAYER, SINGLE PLAYER AND INSTANCING

One thing I don’t like about most MMO structures is the fragmentation of the player base between these “shards”. If you had joined much later than a friend of yours, there may not be room on his world instance anymore and you have to join another parallel one and so cannot play together. This is one of the nice things about the Eve Online design – everyone plays in the same universe.

Despite the above message, it’s precisely what the heck is going on right now in 2.0 and there is no persistence in that PU. None.

And everything you read in that above missive, is pure and utter bullshit. They haven’t built any of it, have no hopes (or the tech) of building it.

And until then, my guess is that it’s going to remain a sharded, instanced world, bereft of anything remotely cohesive, massive, let alone persistent. In other words, a glorified instancing game in which you and a group of friends (12-16 atm) exist in your own pocket universe and doing basic stuff and none of what has been promised in a persistent world.

Dev explanation of instancing from back in July 2015

I think some of the confusion comes with the vocabulary, traditional instances are “copies” of an area which a small subset of players inhabit – reducing things like network lag etc. However this does also limit the player experience as not everyone is truly a part of the same world at the same moment. In our case we would like achieve the highest concurrency we can while still making the game playable. To that end we are approaching instances both in this traditional way as well as working on some newer ideas. If we look at it traditionally, an instance of an “area” like planetary orbit will have a maximum number of player slots, the logic of the Generic Instance Manager (GIM) will handle things like assuring that friends/parties as well as adversaries, factions, skill levels, and other attributes are all taken into account to place people into these instances. This will work in conjunction with a “Galaxy” server which simulates the universe to determine the likelihood of “encounters” or the population currently in orbit of a planet and stages the instances.

To speak to the newer ideas, and your question of “can I see through other instances?”, if you think about an instance as described above than I would say you would not “see” into another instance, it Is essentially a copy of the same area of space (parallel world of sorts) – however this does not mean that an instance is limited to one actual “server”. So for example (these are just example numbers not hard facts) if a server can support 20 people and another 10 people (through GIM or “Galaxy” logic) need to be added to that instance; a 2nd server would be spun up thus allowing a current total of 30 people in that instance shared across 2 servers (a missile object fired by any person in that instance could be seen and interacted with by any other person). This could theoretically be scaled continually, however there will most likely be rendering and network limitations at some point, let’s say that number was 100 players. So you would have 5 “servers” capable of supporting 20 players each make up a single 100 person “instance” – players can seamlessly move from server to server behind the scenes and be unaware.

This is not that same as changing instances though, which could be another set of 5 servers with another 100 people in a copy of that same area. In that specific moment the 100 players in Instance A will not see the 100 players in Instance B, though imagine 2 ships in Instance A and 1 ship from Instance B fly away from that planet towards the same open space, at some point those 3 ships will be “moved” out of their planetary orbit instances and perhaps have an encounter in a Battle Instance that was initiated around them, so you really can’t think of the different “instances” as fixed, they are not like realms, where only a subset of players exist. They are quite fluid and constantly being updated and refined by the GIM and Galaxy services.

On the slightly more technical side, depending on the contents of an instance it could be that it will be scaled differently than another instance, 10 Carriers in a single instance may cap itself to 70 players across 5 servers instead of the 100 in my above example. There are a vast array of metrics that go into determining these things, from both in-game stats as well as hardware wise like cpu and mem limitations. New technologies like containerization and the zoning system will help us spin up servers faster and allow for the seamless backend mentioned.

It is important to note that much of this is in development by many different teams and the final application and usage could vary from this roadmap. While not directly addressing each point you ask hopefully this gives you some insight into the instance system.

And this mixed message is part of the on-going engineering problems they’ve had and is not unlike all the other development missteps whereby stuff gets done, ripped out, redone etc. Rinse. Repeat.

I am going to say it again. There is no way on this God’s Earth that they can build the game they pitched and promised. Not as a sharded instance, nor as a persistent world MMO.

And note that here we are in Nov 2015 and in 2.0 they still can’t get more than 12 clients in any “instance”. For a game that’s supposedly launching in 2016, there is still no evidence of anything persistent about the world. Let alone half the shit they promised.

Thing is, MMOs can be instanced while still be persistent.

For those of you who can’t grasp the concept, here is a well-written and very relevant primers:

What is an MMO? What Defines An MMO?

What MMOS learn from non-MMOS

[Peter Balish]

I was curious to see if anyone had any more information on the ‘Z fighting’ or jittering mentioned earlier in the month. I am very curious about the tech behind the large world and how it was pulled together, and was curious if anything else turned up that corroborated the 32*2 thoughts, or possible some other kind of work around. I was watching a video earlier today: https://www.youtube.com/watch?v=VyWijRL9VfA of the most recent patch (c? d?) and I can’t tell if they have done something clever with the shaders and shadowmaps to mask the issues, or if I just am not noticing the problems.

[dsmart]

Graphics issues are easy to fix; so I’m sure that if you are no longer seeing those anomalies, that they are actively fixing them. Though I don’t see any such notation in the 2.0 patch notes thus far. Could be they don’t deem them important enough to list.

[Peter Balish]

Thanks for the response! I initially had the same thought as in your other post that the glitches were a result of issues with the large world conversion, or that they were a ‘sign’ they had fudged some of the implementation, and therefore it would be a tough thing to mask with graphical fixes – especially the jittering of the player character which seemed to get progressively worse as the ‘time in game’ went on.

I will keep watching videos (or streams are probably best to catch it as it happens) – but do you think it should get worse based on distance from a ‘game object’ like a station or planet?  By my understanding, they could have set planet/station locations, and then the player location would be relative to that.  If they have made a full conversion to 64 bit, I hope an industry mag or blog is able to get a close look at how they did it!

[dsmart]

I don’t believe they’ve done it. And there won’t be anything ground-breaking about it for any mag to write about since some of us have already done it years ago.

I just watched a stream which showed the same problems with jittering (not to be confused with z-fighting) issues and similar precision related issues.

[Von Neely]

I played EvE for several years as well.  Your monologue about Goon strategy is valid, and I don’t see it contradicting anything that I said earlier: Their primary weapon is vast numbers and the organization that allows for actual teamwork even with such vast numbers in play.  I am not contesting this – I said the same thing myself, so I’m unsure what you meant to prove by repeating it more verbosely.

Here is the only question/argument I could find in your reply: “That’s what a dedicated playerbase gets you, how can you see the future and know for a fact CIG money men can’t copy some parts of this for future funding?”

That is a good question.  If CIG was smart they would certainly rip-off a lot of EvE’s ideas (no shame in that, it’s how the Roman Empire got built, after all).  Yet I don’t see them doing that because CR’s ego/”grand vision” won’t allow for external influences.  The classic “it’s not a good idea unless I came up with it” fallacy.

Furthermore, at this point it would not matter if they did start stealing ideas from EvE because, as I stated before, they’ve already crossed the point of no return.  The “lifetime ship insurance” automatically creates a infinite power gap between those whom have it and those who do not.  To use a metaphor, imagine a single hornet.  Now imagine that this hornet is immortal – if struck down, it will rise again like “Wolverine” from the “X-Men.”  That hornet decides to attack, let’s say, a gorilla.  It stings the gorilla.  The gorilla squashes it.  Hornet comes back, stings the gorilla again.  Gorilla squashes the hornet again.  Hornet comes back, stings the gorilla again.  Repeat this as many times as needed until the gorilla is dead.  The hornet then flies away triumphant.

“Lifetime ship insurance” is a game breaker and it’s already in the game.  What’s worse is that it’s something that is bought with real money.  This is not only bad because it’s obvious pay-to-win, it’s horrifyingly bad because they can’t ever take it back.  If they ever realized/accepted that this was a bad idea & tried to cancel it, what would happen?  At the very least they’d have to refund all the money that they made from selling that particular item (see the entire rest of this blog for why that won’t happen).  At the worst they’d have to refund the money and vex those same backers enough that they backed out completely – which would lead to more refund requests and more bad publicity from the burned whales.

CIG has trapped themselves pretty bad with that.  Yes, they might be able to survive pay-to-win ships if they adopted EvE’s “no refunds upon death” policy, but then whales would stop buying them.  Yes, there are lunatics in EvE who will pay $1000+ of real cash to get into a Titan.  They have whales of their own, but they’re not entirely funded by whales.  Meaning if all the people who sell PLEX for shiny new ships were to suddenly quit the game itself would survive.  CCP are not dependent on PLEX/whales/legitimized RMT.  CIG, on the other hand, are entirely dependent on the sale of spaceship pictures.  To risk losing sales for any reason – even good, logical reasons like “not breaking the entire game” – isn’t an option for CIG even if they wanted to.

And they don’t want to.

[Author]

Posts