According to the team at Automaton, an “innovation-focused” developer based in the United Kingdom, the reason that the battle royale genre feels so cumbersome — with its rubber-banding, awkward animations and unrealistic landscapes — is because those games are simply built the wrong way.
Games like PlayerUnknown’s Battlegrounds, Fortnite and Escape From Tarkov, they say, are prone to be buggy and destined to be sterile because of the limitations baked into their infrastructure. Their upcoming game, whose working title is Project X, aims to rebuild the modern survival shooter from the ground up by putting the game world itself into the cloud. Then, to hear them tell it, a user’s PC is free to simply act as a window into that world.
In reimagining how they structure the relationship between the server and the client, the developers at Automaton believe they can increase the fidelity of the experience dramatically and the player count exponentially. The pitch, as explicated on their website, sounds more than a little far-fetched. Automaton is proposing between 100 and 1,000 players all battling it out on a massive, open-world map kitted out with guns and vehicles.
I was so skeptical that I read their entire website twice just to make sure there was no sign of the word “blockchain” anywhere before I sent them an email. After reaching them by Skype late last week and pressing them long and hard about the project, I’m cautiously optimistic.
What they claim to have done is built a game map bigger than The Elder Scrolls 5: Skyrim and mounted it in Improbable’s SpatialOS cloud platform. They then coupled that world with an end-user client built from Crytek’s CryEngine, leaning on its real-time rendering technology to drive the experience on the player’s PC.
Simply put, they built the servers to do only what they’re good at, which is handling lots of data and moving parts. Then, they built the client to do only what it’s good at, which is pushing pixels, polygons and lighting effects out of the monitor and into the player’s eyeballs. By doing so, said Automaton’s CEO James Thompson, they’ve managed to free up a lot of computational overhead to make dramatic improvements to the quality of the game world.
When they say that Project X will be “higher fidelity” than its competitors, they don’t mean simply in terms of graphical enhancements. They mean in the depth of the immersion. Reactive wildlife. Player footprints. Dynamic fire. Persistent destruction. Foliage displacement. The kinds of things that a smart hunter can use to stalk their prey.
They’re proposing a game that looks like What Remains Of Edith Finch, one where you can track another player by the subtle path their passage has carved through a thicket.
“One example would be that wildlife will react to players,” said Lawrence Barnett. “It will react dynamically. So, if you run through a bush you’ll scatter a deer to the right. If there’s another player to the right of you that you’ve yet to have seen and that deer runs into that player, it will then turn around and look really rather startled before finding an alternative route.
“We’re trying to create this foundation that will lead on to empowering the player in many, many different ways. It’s a little bit like an RTS, where players actually end up using the map not just as a map. It means far more than that.”
The first glimpses of what they’ve created with Project X will be shown at the PC Gamer Weekender in London on Feb. 17. Another, more industry-focused presentation will be made at this year’s GDC. But, later this year, Automaton is committed to launching a 100-player version of a battle royale-style game into open beta. That, they say, will be the true test of their model.
From there the goal is a 400-player, squad-based version and in 2019 a 1,000-player open-world sandbox, complete with single-player missions and objectives.
It all sounds, quite frankly, a little too good to be true. But they swear it isn’t.
“It’s actually SpatialOS in combination with CryEngine that makes it work,” said Thompson. “You need a fully real-time graphical simulation to be able to translate that experience to the user. Not only do you need a really vast amount of scalability on the server, but you can’t actually use a lot of those features on Unreal Engine or, in fact, almost any other engine at all.
“Everything [the user sees on their screen] is basically computed on the fly, and what that enables us to do is simulate a massive world on the server. Then, as the player moves through that world, you swap in and swap out the necessary information that they have to see. … There’s no way your computer could handle all this by itself.”
What the team is working on right now, they told Polygon, is closed beta testing of small sections of their massive world map. By refining them discretely, in chunks, they’re hoping to develop areas of the map that foster a more dynamic end-game. Imagine a final round of Battlegrounds that plays out like Call of Duty 4: Modern Warfare’s classic map, Crash, and not just two players hiding behind bushes in a field “trying to see who wiggles their toes first,” as Sam Hills put it.
When they’re finished, the map for Project X won’t have the same feeling of randomness, or the almost nonsensical placement of terrain that was common on Battlegrounds’ first map, Erangel. Instead, it will feel more fluid, like the geotypical terrain common in games like Arma 3’s Tanoa and Ghost Recon Wildlands Bolivia, while providing the same kind of approachable first-person gameplay seen in traditional shooters.
More importantly, the map will organically encourage different playstyles. Where you land after the initial drop will have a huge impact on how you play. To get that kind of density of set-piece areas, the team is starting small and filling the map in as they go.
“There was interesting gameplay in the original DayZ mod for Arma 2,” Hills said, “around how would I interact when I would go into [the main city of Chernogorsk], because suddenly I’m not in all these bushes and stuff anymore. [...] It’s about tailoring how those set pieces are going to work, so that when we build a bigger map and we kind of fill them with it [...] you can drop into somewhere that is an interesting area [that is] built in a logical way by looking at smaller maps and then expanding out.”
On top of that, the team said they’re using other concepts like machine learning and analytics to build a game that is capable of adapting itself to players.
“We want to build a living, breathing world,” said Barnett, “rather than Barren Russia 2.0.”
We’ll know more after Feb. 17. We might even finally know the game’s real name.