-
Notifications
You must be signed in to change notification settings - Fork 20
Description
Description
As we're getting closer to release, the issue of hosting servers, seeding and lobbying is worth re-examining.
High Pop Story
Currently, there's budget allocated to the server hosting which should support around twenty instances. Depending on comp/pub distribution, this could host between 200 to 500 players. Playercount scales linearly with cost, so supporting a larger playerbase would be managable, but a little costly.
Relying on community servers to be set up is great in theory, but the ratio of players to server hosts is likely to decrease even further with the release of rebuild, as the older population with more disposable income is proportionally displaced by a younger crowd with less purchasing power.
Low Pop Story
The historical issue with low pop is mainly one of seeding; getting people to start playing without opponents is difficult. Bots do help but aren't a guarantee that people would be willing to stick around.
Seeding empty servers has a couple of problems, it's known to be pretty unenjoyable for the seeder, with a prisoners dilemma type experience where you might feel obligated to stick around even though the server is empty because if someone joins are you're the only person there and you're AFK, the second joiner is very liable to leave before you notice them and get back.
Seeders are also often long term community members with relatively high levels of skill and map knowledge, whereas second joiners are in many cases new players looking to figure out what the game is about. This often leads to a 1v1 scenario where the time invested in seeding is kneecapped by the seeders own skill, as they're quite easily able to 8-0 the new joiners. This often leads to one to two maps of stomping, alternatively no actual gameplay and just a friendly walk around a few maps, which while maybe interesting for the new player, isn't really ripe for long-term engagement.
Proposed Solution
As a potential player during low pop, I'd like to be able to queue without having to commit to a specific server or seeding an empty one.
As a potential player during high pop, I'd like to be able to queue up and start playing, even if dedicated servers are all full.
For this I propose two processes, in reference to the steamworks Lobbying API.
Low pop: A player creates a lobby for playing a game. This lobby fills up with minimum ten players, then they are all directed to one dedicated server.
High pop: A player creates a lobby, a minimum of ten players join, then if no server can be found to fit the lobby, one player is selected to host, leveraging steam networking to prevent issues with upnp and keeping the host and players IPs private.
Lobbies should have selections for comp or pub, gamemode, and region. Listen vs dedicated server selection should be a trinary switch; automatic, dedicated or listen.
During low pop, clicking "find servers" could issue a popup saying "no one is online right now, would you like to start queueing?"
Alternatives Considered
The current setup is okay, UI wise especially for the older demographic who is used to server browsing manually. But the details listed about with regards to the issues of seeding low pop can't be resolved with the current system.
Additional Context
The interface reference can be found here: https://partner.steamgames.com/doc/webapi/ILobbyMatchmakingService
and here: https://steamapi.xpaw.me/#ILobbyMatchmakingService
The steam networking interface reference is available here: https://partner.steamgames.com/doc/api/ISteamNetworking
Note: Lobbies are not necessarily only for queueing in the actual game, it's possible to set up a system where someone connects their steam ID to discord and we're able to let people queue without having the game started. It's also possible to pull the queues currently open and display them in external platforms with a publisher API key.
Metadata
Metadata
Assignees
Labels
Type
Projects
Status