Author: yossariansmith

The Problem with Planet Screens

The Problem with Planet Screens

I hope you like obsessional analyses of minute details of space 4X design, because that’s what I’ve got for you today. Specifically, I want to talk about something I think seriously undermines space strategy games’ ability to let you enter into the fantasy that there’s a whole universe behind your computer screen: the mechanics and interface of their planets.

Planets, in that majority of space strategy games that I’m talking about, are treated like cities in a Civ game. They build stuff, you can upgrade them with a variety of buildings, and they have a bunch of citizens on that you can assign to different jobs. They might on top of that have some internal geography – a more-or-less-stylised map of the planet and the things on it. They’re displayed in a special screen or window of some kind – the interface and the mechanics are intimately entwined, especially in the more geographical cases. Let’s look at some examples before I get into the things that bug me about them.


This slideshow requires JavaScript.

This is the Stellaris planet screen. As you can see, there’s a grid map with pops, buildings, and tile blockers on it, a government screen that gives you a planet overview and lets you set the governor and turn ‘edict’ modifiers on and off, and an army screen that shows you how many soldiers you have there and lets you build more, and also lets you see how invasions are going if that happens.¬† There’s also a spaceport tab, but I couldn’t find a picture of that and then I got distracted by Reddit, so you’ll just have to imagine a list of spaceships you can build and some icons for spaceport modules. Of the four tabs, the map is the one you’ll spend most time on – that’s where you build buildings and assign pops to run them. The whole thing takes up about one-third to one-half of the screen, depending on your resolution and whether the side menu is open, which it is whenever you’re doing more than just looking at the planet.


This slideshow requires JavaScript.

This is the Galactic Civilizations III planet screen. It lets you build stuff on tiles, as in Stellaris, as well as managing the planet’s budget allocations. Shipyards are separate entities, fed by planets, so they have their own interface elsewhere. There’s also a planetary invasion mechanic that’s a bit more interactive than Stellaris’, although I could never pay attention to the game long enough to try it out. It’s worth noting that in both Stellaris and GalCiv buildings give out adjacency bonuses, providing for some decent planet-optimising gameplay.

star ruler 2 planet screen.jpg

This little baby is the Star Ruler 2 planet screen. SR2 has this fantastic mechanic whereby your planets all connect up into a sort of supply pyramid, with small-ass planets at the bottom feeding their resources to level up mid-level planets, which in turn service top-level planets. There are about five levels in all. That’s what’s going on with the weird tree thing in the bottom left. The rest of the screen is pretty standard – stuff you can build on the right, a map in the middle, a build queue in the top right, and some lists of outputs. The notable thing about the map is that stuff gets built on it spontaneously. It doesn’t matter that much, though, because you can build other things on top of the spontaneous stuff and it’ll just crop up somewhere else.


This slideshow requires JavaScript.

This, finally, is the Endless Space system/planet screen. The main management unit in Endless Space is the whole system rather than the world itself, which as I recall gets to have about one specialisation and a bunch of pops on it, rather than anything more detailed. So the planets act more or less like tiles on which you can stack varying numbers of pops to produce different mixtures of resources. There’s also a build menu down at the bottom there, not shown on the screenshots. It looks much the same as the one in Endless Legend – nice little rectangular icons of the things you can build in the system. Amplitude’s UI is always pretty slick.

So what’s the fundamental problem with all these?

One of the things a long, slow strategy game has to do is evoke a sense of place, as I’ve argued before. The problem with space is, there’s not a lot there – the only places worth the name are generally planets. So the planet interface and mechanics have to do the bulk of the work here: they have to make the planets feel like real places. And these fairly canonical examples I’ve given aren’t very good at it.

It’s a hard problem because of all the other things your planets have to do at the same time. They have to be manageable without grinding hours of tedious micro, which puts a cap on the complexity of the mechanics with which the player directly interacts – as your empire grows, the things your planets do have to aggregate fairly nicely so you can, one way or another, manage them in groups rather than individually. They have to be and stay distinct enough from one another that you remember which one’s which. They have to fulfil the basic functions required of a management unit, i.e. producing resources and building stuff. What’s a game to do?

I can’t begin to give a ‘complete solution’ to this, not least because there are certainly a whole bunch of ways it could be done. What I will do is point out a few specific failings in the examples I’ve given, and suggest possible approaches to solving them.

The first and biggest thing that jumps out at me is that none of the maps feel like they show places. The worlds they show you look like industrial estates – no cities, no countryside, just a bunch of facilities sprawling indiscriminately over the face of the planet. Now industrial sprawl isn’t necessarily a bad thing – sometimes you want to have industrial-sprawl planets – but it can’t be the only developed landscape your planet system does.

What I would do to create differentiation is kind of simple: stacking multiple buildings on the same tile. Some buildings, like factories, entertainment venues, and administrative hubs, should benefit from being stacked; others, like farms, ought to suffer from it. Some buildings that stack well ought to benefit from the proximity of others that don’t. What this does is lead the player to build ‘tall’ cities set among hinterlands of ‘short’ farms, parks, opencast mines, and so on – creating a differentiated landscape that feels more like a world that might be lived in.

What this does do, however, is make the interface more complicated. Having to show three dimensions of developments rather than just two is tricky. One potential way around it would be to eliminate one of the other dimensions: show a slice through the planet as a loop of tiles on which buildings may be stacked, something like Kingdom. The advantage of this is that it emphasises the planet’s relationship to the space around it – you can see the outside and the inside, the poles, and have the potential to do things like have tidally-locked planets with hot and cold sides, or show one side of the world blasted by supernovas or other space hazards. The disadvantage, obviously, is that it’s much less map-like, and could come off pretty cartoony, which might not mesh well with the feel you want. YMMV on that one.

Another option is to take a Civ VI-like approach, and upscale the tiles to show multiple buildings next to one another on the same one, with miscellaneous urbanisation-type greebling cropping up around the more-developed ones. A third option, for a game like Stellaris that has pops, would be to simplify the whole thing, and, rather than stacking buildings, just have some buildings accept more workers than others. Even that detail would give the developed landscape a bit of texture, and create the sense that the planet has cities.

So that’s problem number 1. Problem number 2 is that the number of different planet archetypes tends to be limited, especially when your game has terraforming in it. Typically, what you get is this: you have a number of basic planet types – ice, lava, terrestrial, etc. – and then, when you develop them, you get a handful of different roles for a planet – manufacturing, farming, and science being the main ones. The number of planet roles is more-or-less constrained by the overarching mechanics of the game: you can’t have that many different resources to collect, for example, because the game would get unwieldy. They also don’t tend to combine in a super interesting way with the base types: either the types don’t really matter, in which case they stop being interesting, or they make planets more suited to one role than another, in which case your type-and-use combinations tend to collapse into one. This problem gets even worse when you have terraforming in the game, because you (usually) end up bulldozing all your planet types into your species’ favourite variety. What can we do about this?

One approach that could work would be to add a whole bunch of niche planet roles that provide bonuses that affect every planet in your empire (or at least the ones nearby), scale up as your empire grows, or only become necessary if your empire grows over a certain size – prison planets that reduce unrest, tourist hotspots that produce more money the bigger your empire gets, military training worlds that improve all your defence forces, regional capitals required to expand your borders, and so on. These world roles would be pointless, and probably unavailable, in the early game, but would become worthwhile as your empire gets bigger. For every basic role your planets can play, you get at least one buff-planet role to match, plus others for mechanics too minor to dedicate basic planets to.

There are a couple of ways you could push your players into specialising planets like this in a game where your planets have tiles you build stuff on. Firstly, you might railroad them into it by having specialised planetary-capital buildings that unlock other buildings that reinforce the planet’s role. This makes planet specialisations very explicit, but has the downside that the player might feel their planets are merely cookie-cut from pre-existing moulds. Another way to do it would be to have buildings that feed into these roles buff one another, so the more temple complexes (for example) you have on the same planet, the more effective each one is; a third would be to have them negatively affect non-matching buildings – people might not like living next to prison camps – and a fourth might be to have certain specialised buildings require a number of other buildings be on the planet before they can be built. Probably the best approach is to combine all of these in whatever way feels most natural. It’s good to be systematic, but taking a variety of approaches to encouraging specialisation helps individuate different roles further – and if you want to evoke the real world, you need to have things be a little eclectic.

The third and final problem I want to look at is related to but distinct from the narrow-roles issue. I’m talking about the way planets’ types and unique features cease to matter. In Stellaris, for example, planets’ terrain is conceived of as a set of ‘blockers’ you progressively remove. Star Ruler II has maps, but they’re effectively featureless, and get obscured by urban sprawl anyway. GalCiv III and Stellaris have tiles with special resources on them, but those tiles are otherwise characterless. Stellaris does do some cool things with primitive pops, but those are few and far between, and the landscapes themselves still do very little, and rapidly vanish underneath your unstoppable wave of development.

What I would do to solve this again involves adding some content. Essentially, you need a handful of different basic terrain types plus some rarer, weirder features for each planet type you have – and then you need to come up with unique improvements for those terrain types. Your basic ‘plains’ or whatever might get the lion’s share of available improvements, with the more unusual terrain types only allowing for two or three different options, but there have to be interesting decisions for each one. A few improvements might be at home in multiple terrain types; a lot of improvements might be mechanically very similar but thematically different depending on the kind of terrain they can be built on. The point is, as with the planet roles, that development of a world has to increase its uniqueness, not decrease it.

That’s all I’ve got for today, folks – join me next time for some stuff about character mechanics and how they might be used to make a space-strategy world feel alive, probably, unless I get distracted by something else and write about that instead.

Let’s Fix: Stellaris’ Sectors

Let’s Fix: Stellaris’ Sectors

Whenever I play a 4X/grand strategy game, I almost always come across a dozen things that make me think, ‘Oh, that could be so much better with just a couple of changes…’ These posts are the records of what I would do with those things if the opportunity came my way.

Stellaris‘ sector system is a brave attempt to address a perennial 4X problem: micromanagement hell. In many of these games, after you’ve conquered a dozen or so planets, play slows right down: the bigger your galactic empire, the longer you have to futz around to optimise it all. By using resource income penalties to force you to bag up your planets into these larger entities, sectors, Stellaris stops you doing this, thereby avoiding the infinite-sprawl problem.

Unfortunately, it does this in an un-fun and frankly annoying way. Sectors come off as characterless spreadsheet entries, with none of the interest available in planet management – they can’t even be selected on the map, you have to open a menu to find them. Hand a planet over to a sector, and you more or less never see your carefully-customised world again. Your management of the sector is limited to choosing one of about five ‘focus’ options and deciding which planets to put in it (using an atrocious interface that forces you to pay to correct mistakes rather than waiting for you to select your changes and then asking you to confirm them); beyond that, it just sits there and does its own thing, running on not-very-efficient autopilot. The governor makes no demands of you. There might be an independence movement, occasionally, but in the main it just sits there processing resources and building suboptimal buildings you never see anyway.

The fundamental problem here, it seems to me, is that Paradox have not decided what sectors really are. Are they larger management units, like planets, or are they semi-autonomous vassal states? They don’t have the depth of control to be the former, and they don’t have the political liveliness to be the latter; the Imperial Routine mod adds sector factions with leaders and agendas, which makes things more interesting, but the base game does very little with them.

What I would want to do to fix sectors (above and beyond allowing you to select them directly from the map) would be to separate out those two functions and add much more depth to each. The basic sector should be a larger management unit that does fundamentally the same things, and is fun for the same reasons, as a planet. There can then be a selection of reasons – pro-devolution factions, for example – why you might also want to grant the sector a varying measure of self-government; and when you do so, that ought to give rise to interesting political interactions between you, the sector, and your neighbours. To really get into that, though, would involve looking at Stellaris’ political and character system as a whole, which is a bit far-reaching for this post, so I’ll stick to looking at the first function. How do we make sectors work better as management units?

The key to this is to look at all the things the player can do with planets, and have sectors mirror those capabilities. So what does a player do with a planet?

  1. Look at it. The planet screen is (kinda) visually pleasing, with little pictures showing the player the landscape. (To be honest, the Stellaris planet screen is ugly as sin, and only marginally better than the one from Star Ruler II (a brilliantly innovative indie game that sadly ended up feeling lifeless, due not least to its barely-there art direction), but there are still little people and bits of scenery, which is much better than nothing. I might look at better ways to do planet screens in a future post.)
  2. Build buildings on tiles and assign pops to them. The management unit has a bunch of slots and the player decides what to put in them, with an eye to good placement.
  3. Build spaceships. The management unit has a single, central management interface for doing this.
  4. Set edicts and planetary policies. The player spends Influence to shape the development of the planet in various ways.
  5. Manage troops. The management unit has its own defensive ground army and the player can build, customise, and assign offensive armies from its interface.
  6. Produce resources for and consume resources from elsewhere in the empire.

All of these things could quite easily be done on the sector level:

  1. The interface – currently a dry sub-list of planets within the territory manager menu, which doesn’t give you any clue where they are – needs to a) open when you click the large, prominent sector name, and b) show you the planets in their locations on the map. Ogi Schneider and I had a hell of a time trying to make this work in Alliance of the Sacred Suns, but we had more information to communicate. Stellaris can probably fit everything into tiles only slightly bigger than those on the planet interface, and allow for a bit of scrolling within the display, which might be projected onto the main map to save space. This is not an exhaustive specification, obviously, but the main required elements would include each planet’s image, name, population, specialisation (see the next point), resource surpluses/deficits, and happiness; the tiles would be more or less elastically tethered to their stars on the map. Each star might also have a smaller tag attached to it showing the inputs and outputs of the mining and research stations in the system. There would also be a similar map for troop movements, substituting information about troop numbers, types, and rally points for the default mode’s political-economic information. You’d also need a build menu for each of those modes, a population breakdown by species (either in a corner somewhere or under a different tab), a place to display the governor and active edicts (again, probably under a different tab), and a shipyard interface (mostly the same as the planetary one). The idea would be to follow the pattern of the planet interface wherever it makes sense, to reduce the amount of new interface the player has to learn.
  2. Rather than building particular buildings on particular tiles, the sector interface ought to allow the player to choose planetary specialisations: agri-worlds, research worlds in various specialisations, diplomatic nexi, administrative centres, military and civilian industrial worlds, various kinds of mixed-use world, and so forth. The AI would then develop those worlds in the appropriate direction. Planet-level and intra-sector migration policies might be set, in lieu of reassigning pops on individual tiles. Mining and research stations and shipyards could be built on planets in the sector directly from the sector interface.
  3. To have a single sector-wide shipbuilding interface would be simplicity itself. Although different spaceports can be specialised to have different strengths and weaknesses, just allowing each build order in the queue to be toggled between ‘get produced as fast as possible’ and ‘wait for this (/one of these) specific shipyard(s) to be available’ would handle that problem. The sector AI could add things to and reorder the build queue, strictly on the understanding that it can’t increase the wait time for anything the player has ordered built; the player might set aside shipyards exclusively for either their or the sector AI’s use.
  4. Replicating the planetary edicts and policies mechanics for sectors should not be a problem – sector-wide edicts and policies can work in exactly the same way. It might make sense to make them more cost-efficient than planet-level ones, given they’re more of a blunt instrument.
  5. Troop management is a more interesting problem, because rather than the troops all sitting on the same planet, they may be raised on a bunch of different planets, with different bonuses to skill, etc. The case of defensive armies is relatively easy – the player can simply have a dial on each world that says ‘Have this much strength of defensive armies’, perhaps with lists of checkboxes for which upgrades to give them (an even mix of all the boxes checked). Assault armies, in their various exciting guises, will be a little more complicated: they should have a muster point or points (between which they’ll distribute themselves evenly) in the sector, and the player must have the option to choose which of the variously-buffed versions to produce. This might be a little fiddly; unfortunately, the whole army interface is already kind of fiddly. Anyway, so long as it’s all done through a central interface, it’s less of a pain than going through and doing it planet-by-planet.
  6. Finally, resource imports and exports should be a) possible and b) managed through the sector interface. Right now, all you can do is bung them some energy or minerals, or spend Influence to drain their coffers. This is better than nothing, but still not great. Instead, their day-to-day surplus or deficit of each resource, including the special ones and excluding capital costs, should be visible, and the player should simply set how much of each is to be imported or exported by the sector to the main empire, with the option of saying ‘As much as you need to cover your maintenance’ for each resource. High imports and low exports would make the sector’s people and governor happier, while the reverse ought to do the reverse; perhaps Influence could be spent to perk them up for a while, to reproduce the resource exchange that takes place in the game as it is currently. This is a more complicated and hands-on approach than is taken by the planet interface, but it’s necessary in order for the player to be able to keep tight control of their stockpiles: they need to know what’s going to be going out and what’s going to be coming in, rather than being blindsided by some idiot AI building a bunch of expensive-to-run boondoggles without their knowledge. As a priority (because it’s a point of high visibility for the player) the AI ought to make use of imports assigned to its sector – if I send it some Beltharian stone, it ought to have the good sense to build a reactor with it – and address its own deficits if it can, within the strictures of the planets’ assigned roles.

The end result of these changes would be to make sector management engaging in the same way planetary management is – keeping the amount of micromanagement the player has to do at a constant, pleasant level throughout the mid- and late game, whilst still allowing them to have a feel for the shape and character of their empire.

Join me next time for, uh, more kibitzing about Stellaris, most likely. There are things I’d like to see done with characters and procedural storytelling, and a rejected planet screen design concept from Alliance… that I still really like, so you’ve got that to look forward to. Until then, keep it real.

Some published articles

Some published articles

Meet the superfans still playing Populous: The Beginning, at RockPaperShotgun

An interview with Limit Theory’s Josh Parnell, at

A look at the community around 1990 racer Stunts, at RPS

A deep dive on spaceship design in games, TV and movies, at

The strange, retro world of Vocaloid fan games, at RPS

An investigation of how The Legend of Zelda: Breath of the Wild will affect the open-world game market, at