Insightful, in-depth and entertaining features on Planetary Annihilation, real-time strategy and related subjects.

Taking Clan Wars from Large to Huge (I)

The road from multiplayer to massively-multiplayer and how to navigate it

With yesterday's War, we are back to 20+ player lobbies (22 for CW7, to be exact, and it would have been 24 if we had found another caster and both Guitarman and I could have played for BRN and CULT respectively, instead of casting), and so the question arises again of how do we deal with mechanical and perfomance issues once the lobbies start exceeding the 25- to 30-players number that seems to be about the most that the fastest servers available today can accommodate with minimal-to-tolerable lag (remember that CW2 had 30 players, and CW3 had 25, so we have considerable experience in games of this size by now, and enough data to draw at least some preliminary conclusions about them).
   First off, let me say that one of my Clan Wars' main goals is to increase this number indefinitely, and I won't hear any objections to this. I am not interested in imposing some arbitrary limit on this number any more than Jon Mavor and Uberent were interested in imposing some arbitrary unit cap to their game (although TONS of players have been asking for one for ages). The LACK of limits of any kind is precisely the ESSENCE of PA, as I have explained at length in the first part of my Videogame Art essay on the game [ > ], so if you are interested in expounding on the joys of arbitrary limits, save your effort and go talk to someone who'll listen, because I sure as hell won't. Even in my Clan League, which is a far more conservative event, the ultimate goal is to be playing 10v10s and 20v20s, etc., and the only reason we're still playing mere 3v3s is that there aren't enough players in most clans at the moment to allow us to use more players, but this is changing fast, as veteran players hear about our events and come back to the game, and new players see how awesome the game can be if played right and jump in, so expect to see progress there too at some point, and "progress" in our events will always mean "size increase", in every aspect: number of players, planets, units — even casters, eventually, since as BRN leader Elite Sardaukar pointed out in an 18-page thread about our events in Uber's forum that has since inexplicably vanished, how is one man expected to be everywhere at once in such huge wars as these? PIPs help, but have their limits, so three-person caster teams will be obligatory, at least for the Wars, a few months from now, let alone for the many years during which we plan to keep playing the game and growing the events and the playerbase.
   So there won't be any discussion on this front: "bigger is better" as far as I and The Cult and Insomnia are concerned, so if you think that "medium-size is better", or worse still that "smaller is better", you are talking to the wrong people. However, once the theoretical question has been settled — as I just settled it — there still remain formidable mechanical and technical challenges to overcome on the way to achieving that size increase that I have just defined as "progress", in my terminology. And I will now talk about how I plan to deal with those.

   To start with, the problems seem to begin at a player number of about 25, and by the time you've passed 30 it seems you have to take some drastic measures to keep the lag in check. The exact number at which these measures should start kicking in is unknown and debatable, but also not very imporant: the important thing is that these measures MUST kick in at some point, to avoid the game becoming unplayable. What's worse, is that my goal of increasing the player number indefinitely clashes with my goals of increasing the unit count and expanding the system size. More players means lag kicks in faster as the unit count rises than if we had fewer players, and it also means being forced to use a smaller system in order to get the players killing each other's units faster, and thereby keeping the lag in check. Indeed, the optimal player number in order to test the maximum number of units the game can field without lagging is two, and the same goes for system size. So we have a set of utterly contradictory goals in our hands, and have to choose between them.
   To cut a long story short, I will ALWAYS choose players above all. I am simply not interested in turning away people — real human beings — who set aside time on their Saturdays to show up in my events, in order to get a few extra units or an extra planet in my games. Players will therefore ALWAYS trump all other considerations in my mind when designing the formats for the events, and especially in the Clan Wars, which is designed precisely in order to attract the biggest number of players. As for why I am making this choice, apart from the obvious reason of preferring more human beings in my events than AI routines or simulated terrain, it's quite simple really: Technology will eventually allow us to add bigger systems and more units, without much trouble. All we have to do is wait for Intel and AMD to do their work, in due time, and it will happen. It is basically inevitable, and there's nothing for us left to do but wait (with the exception of coders like mikey who can improve PA's performance, but that's one goal I can't help with because I am essentially coding illiterate, so all I can do is wish them best of luck and hope they deliver). Building a large playerbase, however, is NOT something inevitable; indeed, what is inevitable is that this larger playerbase will NOT happen unless we work on it, week in week out, to make it happen, and turning people away from events is NOT the way to make it happen. So I won't turn them away, ever (unless they break the event's rules or generally act like pricks of course, in which case I will STILL always leave the door open for them to return if they promise to respect the events, the other players, the organizers, and me personally and what I am doing for the game and the scene). In sum, I am focusing on the players because they are the scene's most precious resource, but for that reason also the most fragile and difficult to cultivate, and so everything that can be done to create the conditions under which this most precious resource of all can grow must be done, and I am convinced that the best thing that can be done is to keep running the Clan Wars with a "come one, come all" policy, as Elite has aptly described it, and no player limits.
   So what do we do then, to keep the lag in check, if moderating the player numbers is out of the question?
   As far as I can see — and feel free to make other suggestions if you have any — the only two options left to us are 1) to divide the players between separate Wars, perhaps on the same day and even the same time, or 2) to carefully craft the system to feature minimal metal so that players can't build many units and flood the server with them, and also force them to start killing each other very quickly, the best way of which to accomplish this last goal would be to feature an appropriately designed system with the "spawn anywhere" option turned on, which results in lots of commanders dying within minutes (as opposed to the normal way of doing things, in which a game can go on for half an hour without any commanders dying, let alone an entire clan).
   And that concludes the first part of my analysis on how I plan to take Clan Wars from large to huge. I've so far given the background to the problem, defined its parameters and examined how they relate to each other, and briefly mentioned the two possible solutions that I plan to examine. So please look forward to the second part of my article going up soon, in which I will exhaustively analyze these two options, and if in the meantime you have any suggestions to make that you think I haven't thought of, or any comments at all on the above, post them in the forum and I'll have a look at them and take them into account in the second part of my analysis. Until then,

See you on the battlefield,


Discuss this article in the Cosmic War forum.

Next Post Previous Post