So the UK group filed their 2017 financial accounting. It’s as hilarious as it is eye-popping.
Star Citizen just sent out the latest newsletter. Chris Roberts wrote a lengthy response to the P2W fiasco currently playing out.
I don’t know what ‘fiasco’ you’re talking about. There’s no ‘fiasco’.
Some people are worried that they will be disadvantaged when the game starts for ‘real’ compared to players that have stockpiled ships or UEC. This has been a debate on the forums since the project started, but this is not a concern for me as I know what the game will be and I know how we’re designing it.
It’s all right there, this isn’t anything to be concerned about. They are currently designing the game to make it not be a factor at all. In a few years they will really have the design down and they will share it with everyone and get constructive feedback before they start implementing the game.
50 client ground battle.
BEFORE (actual footage)
AFTER (lots of splicing, cutting of loading screens, jank, crashes etc)
Chris just had another hour of verbal diarrhea in which he basically threw John Pritchett (he left back in June) of the much maligned IFCS (flight dynamics) under the bus.
My thread:
KEY PARTS
https://www.twitch.tv/videos/292690918?t=19m51s // what is perfect flight model
https://www.twitch.tv/videos/292690918?t=23m58s // new version of IFCS
https://www.twitch.tv/videos/292690918?t=26m54s // current system is too complex
https://www.twitch.tv/videos/292690918?t=27m51s // refactoring physics different rigid body simulation
https://www.twitch.tv/videos/292690918?t=51m28s // OCS explanation
https://www.twitch.tv/videos/292690918?t=42m17s // P2W question
That second video (the edited one) was cool as hell.
Not trying to be snarky, but why are those animations so bad? They’re herky jerky and glitch all over the place.
Genuinely asking. I’m assuming nothing is “revolutionary” in getting human walking animations to at least be somewhat smooth in 2018. These ones look far worse than Unity asset flip products, so what gives?
I suspect two things. One, there maybe a lot of netcode interference with the 50 man play. Server updates are probably not optimal. They probably are not letting the client do the reporting of positions/hits and stuff. (cheat prevent).
Second and less likely for demo purposes, the machine the game is based on is struggling due to the load.
There’s two fundamental reasons:
The network layer absolutely can’t handle all the information they broadcast all game state updates to all clients for everything that happens on the current server. As the number of clients on the server barely doing things increases framerates plummet. When something actually happens on the server everything freezes or crashes. They have been claiming that the fixes for this are in the next patch for a couple years at this point. Maybe 3? (years, not patches). It is claimed that they will be next patch again for Q3 which they have already pushed back but their fake external roadmap doesn’t show it looking very complete.
Second, they have just ludicrously high poly count models for everything. Which is why a lot of the static screenshots look great, but also leads to performance issues.
There’s various other problems (way over complicated/broken flight model that takes too long to process, etc etc)
Thanks for the info, guys!
Well, they have 64 players at a time in Battlefield 1 and nobody’s teleporting all over the place and the game looks just as nice. Is it something to do with the engine?
Its not that simple of a question to answer, but it is related partially to the tech. @Lantz points out their models are super high rez. Another is the implementation as to how outcomes are determined (on the client or on the server). Many FPS games allow the client to report whether or not they actually hit something (which exposes itself to cheats as I mentioned). Another thing is BF1 etc have been out for a long time and their code is much better proven. Can’t say the same for SC code.
More problems in the netcode is in that 50 player battle there may be attempts to synchronize player commands, so that slows performance even greater, as then you are waiting for the weakest link to confirm something. Who knows what they have under the hood.And there is a lot going on in that video.
You do have to be somewhat forgiving in this regard (“Alpha/Beta disclaimers” here). But I agree people would settle for lower rez and better performance.
In this day and age and for the level of this project, and for what they are talking, 50 players should not be a challenge. Over a decade ago I participated in a LAN party series that one of our big events was “Rocket Area 50v50” where 100 players would play Q2 Rocket Arena. Closed network, but the server was probably on par with my phone nowadays. There was some lag in the big battles, but really the quake engine netcode was incredible. Then with Q3 they began limiting the max player.
Think Derek has mentioned they did not go with experience for their netcode people. That would be reinventing the wheel in this day and age. Sometimes that’s a good thing for support reasons, but on the scale they want maybe not.
I think Derek’s main focus of showing that video was that the previous one showed them waiting at loading screens for literally hours. Not the player experience to enjoy.
EDIT: I will add these things are not easy, but with their budget they should be doable.
Another thing to consider is that they, as I understand it, basically have baseline Cryengine networking code give or take. It’s not built around large environments like they have. And they have basically hacked around at the engine to support it to some extent. But fundamentally they are just running cryengine fps game server instances.
But the crazy thing is that even if they got this working (after a few years of clearly not making progress) it wouldn’t actually be what they need. They are selling the dream of large capital ships quite literally for thousands of dollars a ship and a single game instances for people to play in. Nothing about how they plan to make that work is known. They wave their hands around and spew nonsensical mesh words around.
Even if they complete the next year of scheduled features on time (they won’t), they still would have none of that stuff even started.
They did add highly detailed destructible watermelons to one of the Star Marine levels.
Then they had to remove them because destroying them was crashing clients.
This actually happened. Says a lot about their priorities and ability to make an actual game.
Oh please send a link if able
Star Marine
- Removed watermelons from Echo 11 due their destruction causing a client crash.
https://mobile.twitter.com/shrimpsykate/status/1024231138733293569
Even that second edited video convinces me they have no idea what they’re doing (as if I needed further convincing). So that’s the kind of game Star Citizen is supposed to become?
I watch that video and I can certainly appreciate a cool enough edited battle scene, but there’s nothing in there I see and think, “Ooh, I want to be part of that”. Waiting on deck and saluting when your transport arrives? And then having to float over to your transport and sit in it? Running around in the open trading laser blasts with other players down my gun’s iron site? While one other guy gets to drive around in a moon buggy and one other guy gets to hover overhead in a ship and shoot at the moon buggy?
It’s as if they have no idea how much meticulous game design goes into a Battlefield game, how carefully engineered the experience is. The guys making Star Citizen seem to think, “Well, we throw 50 people out there and a game happens.”
-Tom
Can I pick the seeds out of the watermellon?
Depends. How much are you willing to pay?