Sovereignty: Emergence is neat | EVE Online

Sovereignty: Emergence is neat

2009-09-14 - 作者 CCP Greyscale

We've been thinking about nullsec for quite a long time. The last big round of changes were made in Revelations II back in summer 2007, and we've been watching the results ever since. Some things worked out pretty well. Some things not so much. Some things have changed in the intervening period. And now here we are, two years on, and we've been given a mandate to re-engineer the dynamics of nullsec. Which is exciting, and challenging, and maybe a little scary. We think this stuff is kind of important, and it's not like there's anyone else in the industry who we could talk to about this stuff even if we wanted to. Nobody else does - has ever really done - what we do here: it's undiscovered country.

So anyway, here we are today. Nullsec is largely the domain of large, 2-3000 member PvP alliances, grouped up into inevitable coalitions and engaged in not-quite-impossibly large wars. Costs are mosty covered at the alliance level by a combination of old money and high-value moon minerals. The latter continue to rise in price due to ever-increasing demand from invention, and the after-effects of last year's exploit-related burp invalidating the calculations used to construct the Alchemy pressure-release valve. Most of the space that's up for grabs is owned by a clone army of ideologically-distinct but functionally-similar alliances, making the entire political landscape depressingly homogeneous. The state of the military art is not much better - sub-capital fleets are wheeled out for cyno-jammer take-downs and then packed away before they can fall victim to multiple doomsdays, leaving huge capital fleets to park themselves in front of a never-ending procession of starbases. And the smaller groups, the newer organizations hoping to gain a foothold in the Great Game, are left begging for crumbs around the edges. Who's going to let security-risk nobodies into their back yard when they'll never be able to compete pay as much as a single dysprosium moon?

We're not convinced that this is the best, most interesting, most dynamic and most emergence-friendly state of being for nullsec, so we're going to make some changes.

Why nullsec is worth working on

Nullsec is cool and different and awesome because of emergence. It's not the most populous area of the game, sure (and more on this shortly), but it provides one of EVE's most compelling and unique experiences. It does this because, by and large, we let you the players call the shots. This does have some impact in empire, but in nullsec the effect is writ large.

By giving players and player organizations tactical and strategic freedom, we allow a situation to arise where each challenge is different from the last, because every time there are different people involved making different decisions which result in different outcomes. You may have seen this effect in trailers such as The Butterfly Effect, and it usually goes by the name "emergence". And it's awesome.

The reason emergence works is that players make decisions. The more decisions that players can make, the more emergence you get, and the more interesting the experience is. Therefore, a primary development goal in nullsec is to enable players to make decisions, which can be boiled down to two directives.

First, try to give players tools. More tools give players more options, which means more decisions. Of course, to have value these decisions need to be meaningful - it's not enough to say "you can paint your shed red or blue" if the color of the shed has no impact on anything else.

Second, try to avoid telling players what to do or how to do it. The current sovereignty system, for example, mechanically prescribes a certain path to conquest, which limits the number of command decisions to be made. Obviously you need some mechanics in order to reach a definitive outcome - which lessens the number of decisions but also makes them more meaningful - but in general, the strategy is "deregulate, deregulate, deregulate".

The other thing

As mentioned above, nullsec isn't the most populated area of the game, and doesn't contain anything like the majority of EVE's characters. This presents both a challenge and an opportunity for this expansion. A challenge, because we obviously have to be careful not to ruin the gameplay for everyone in empire by accident, but an opportunity because we can change the balance here and give more of our players a chance to experience nullsec gameplay.

We're aware that some players just aren't interested in the nullsec experience, and that's fine, but we're also aware that there's a lot of players who'd like to try it out but can't seem to get started - in no small part because of the problems outlined in the first few paragraphs. If we have a really compelling game experience, and we have players that want to try it out but can't, then we're doing something wrong somewhere.

Where we're going with this

Ok, so that's pretty much the top-level view. Let's drill down a bit to some of the big whats and whys.

The first big departure is the actual sovereignty system itself (which is only a small part of the whole picture). Right at the start of the project we asked "why do we even need a sovereignty system?", and the main argument for keeping it has nothing to do with shooting at things. Rather, the biggest reasons for having a mechanical system of ownership are to have something we can use to regulate who can do certain things in a system, and (more importantly) to let people stake out their territory. Being able to say "this is our space, we fought hard for it, and now everyone can see what we achieved" is important to a lot of people.

A system to do this can be fairly lightweight. It needs to handle systems changing hands, of course, but it can afford to be descriptive, rather than prescriptive. Currently we have a prescriptive sovereignty system: you fight over sovereignty explicitly, with the sovereignty mechanics determining who owns the system. A descriptive system says who's in charge, so it only needs to change hands after the dust has settled and one side has emerged triumphant. The actual fighting is deregulated - rather than mechanically telling you what to do (shoot sixty hardened starbases), you just need to do whatever it is you need to do so that at the end of the day the enemy goes away.

Of course, there's one class of thing that can't be left entirely free-form, and they're the things that helped bring about the current sovereignty system in the first place: stations. Outposts and conquerable stations are the river-crossings of EVE - each one lets you project power all around it, and as a result they're pivotal military objectives. Station ping-pong - waking up in the morning and finding that someone in a different timezone had taken your station, and the first thing you had to do was shoot it again to take control of it so you could re-dock - was very silly and we don't want that to come back.

There's no reason that the solution to this has to be the sovereignty system, but it does need to be timezone-proof. There's also no reason that it needs to take two weeks for an outpost to change hands - while comparatively shorter switches give the defender less time to mount a defense, they also make re-conquest easier. The combination of a lighter, descriptive sovereignty system and a separate mechanism for outpost conquest should (we think) lend itself much better to emergent outcomes.

Density and density

Sovereignty and outposts are roughly half the problem. The other half are the two related concepts of resource density and population density, and here everything ties itself into a messy knot that can be unravelled in a fairly elegant way.

Where to start? Resource density in nullsec is too low to support a high player density, which limits the number of people that could theoretically live in nullsec. Moon mineral values mean that there's no need at an alliance level to worry about other resources anyway, which limits the number of people who are actually allowed to live in nullsec. A lack of population or vulnerable resources means smaller fleets have little strategic relevance. Alliances hold vast tracts of space that they have no actual use for, simply because they can, locking out other groups from using it.

These problems are all interlinked, and solving them with a few key changes should bring a lot of good results.

Firstly, let people upgrade their space, and in particular its resource density. By increasing the resource density, you increase the potential population density, and by letting players do it rather than simply seeding more resources, you open up more decisions and more emergence.

Secondly, reduce the amount of income that can be derived from mining moons. In conjuction with the first change, this means that the best way to raise funds for an alliance will once again be to fill your space with as many people as possible, upgrade your space as much as possible and watch the money roll in.

This serves several masters. It gets more people into nullsec - one of our objectives - by making big alliances want more people in their space. It makes it much harder to be a big, rich, military alliance; rather, things should move more back towards the old dichotomy between big rich carebear alliances and smaller, poorer military alliances, because history (both in EVE and in the real world) shows that badass military organizations can't handle crop rotation without going soft and squishy. This dichotomy leads to more interesting conflicts; balanced but non-symmetric wars and political interactions between organizations with wildly differing objectives tend to be more entertaining than fights between largely identical groups. And if some alliances are relying heavily on lots of people working in their space on a regular basis in order to fund all their activities, interesting and strategically meaningful small-fleet combat materializes on its own, without resorting to "here is a structure for twenty ships to shoot": there's lots of soft targets for roaming fleets to harass, and space-holders have a pressing financial incentive to keep the residents of their space safe and fight off any incursions.

Thirdly, charge rent on systems. This allows us to scale the rent based on how well-developed a system is, which means it's less of a no-brainer to upgrade (meaningful decisions!), and also reinforces the idea that the more people are using a system, the more money it'll make you. In conjunction with a few well-placed additional penalties, it also combats alliance sprawl, leaving more space up for grabs and again letting more people experience nullsec.

Obviously, the anti-sprawl mechanics are a bit of a soft limiter, as you can always split up your alliance into multiple "alt alliances" to work around any possible mechanic in this vein. That's ok though, although to explain why needs a short digression on social structures in EVE.

The most stable social structures are almost always corporations, and they're also the ones with the most value for players. Corporations usually survive turmoil, and they represent the strongest set of social bonds. Alliances are fairly stable and represent some additional social value, but often fragment after major defeats. Finally, coalitions of alliances are pretty unstable and rarely last beyond whatever war brought them together. (It's also interesting to note that the number of real people in the average large corporation rarely exceeds Dunbar's Number, and that the average stable military alliance is almost always ~3000 players divided into 6-8 major corporations, but that's not directly relevant.)

Groupings of "alt alliances" fall somewhere between regular alliances and coalitions in terms of stability (and by reducing the number of people in alliance chat to a more manageable number, likely actually increase social utility), so even if alliances attempt to circumvent soft limits by fragmenting themselves, they're decreasing their stability and to some degree at least increasing the number of political entities present in nullsec, both of which lead to more conflict and more interesting emergent behavior. And of course in addition, by adding some non-linear cost scaling, the upkeep system will likely encourage at least some multi-region alliances to consider whether they really need all that additional space or not...

Recap that for me?

We implement the following:

  • A simple, descriptive sovereignty system
  • A separate mechanism for governing outpost conquest
  • A way to increase the resource density of your space (as well as other cool gubbins)
  • A reduction in the value of moon minerals
  • An upkeep system for the space you hold and develop

We get (hopefully!):

  • A more comprehensible, streamlined and robust way of showing who owns a particular system
  • A better conquest experience
  • More organic, meaningful and fun small-fleet combat
  • Less territorial sprawl by major alliances
  • A more diverse and interesting political landscape
  • More opportunities for players to get involved in nullsec
  • More awesome emergent gameplay

If it works out like we're hoping, we think this is a pretty good outcome.

-Greyscale

Postscript _This is actually my third stab at this blog. The first was a 3000-word rehash of some internal documents, which was interesting but too wordy and not informative enough, and the second draft I binned after getting to 1500 and realizing I was still warming up... We even discussed not doing this at all for a bit, but decided it's worth doing what's essentially a theory-dump for three reasons. For one we find this stuff really interesting for its own sake, and figured that a few of you might too, and for another we've found internally that a lot of the things we're doing make no sense until you have the "why" of it explained.

The third reason though is to show that we really have thought about this stuff. Nullsec gameplay is a big deal and a lot of you are rightly worried that there's a huge potential to screw this up badly (I know it keeps me up at night sometimes). We think though that we've got a good handle on the underlying theory for what goes on out there, and that gives us a good basis to move forward on. It should also go some way to explaining why we're being fairly comprehensive here. The current systems in nullsec are a bit like a house that's been built up piecemeal from a single small hut, and while it has a lot of rooms, the layout doesn't make a lot of sense ("why is there a toilet in the middle of the living room?" "well, three years ago..."). Most of the prior discussion we've seen, both internally and externally, has been limiting itself to knocking through a few walls and rearranging the furniture; what we're trying to do here instead is to level the entire building, and then rebuild the foundations and the ground floor according to an actual plan. The resulting structure won't (initially) have as many rooms as the current one, but it's been designed with coherent future extensibility in mind, and more importantly the toilet will actually be in the bathroom this time round._