For a long time I've been thinking Deeeeeep Thoughts about construction games. I've written about a dozen essays that I never published, and here's another one!
I think there are several categories of games we're just overlooking completely. They require a completely different approach to their gameplay, pacing, and monetization, which is why so few get made. The category I find myself most interested in is the construction games: games where you build stuff.
Obviously, there are construction games out there in the wild. But I think that, for the most part, they are not very good at being construction games. Nearly all of them are "face this challenge" games. You have to do X or survive Y, build yourself a setup that can do that.
Even Kerbal, my flat out favorite construction game, is like that. Kerbal is a bit more interesting, though, because the player can define the challenge however they want before building something and going on the mission. This is one key to the puzzle.
Another construction game that's quite interesting is Minecraft. Not survival mode (usually), but creative mode. Vast cities, castles, artwork - players expressing themselves via bricks and blocks. Their creations are rarely "functional" in any particular in-world sense, because there are only a few functions that Minecraft's universe actually cares about. But they are very functional out of world: players view these creations, play around in them, and enjoy them. It lies somewhere between creating art and creating levels, exactly where depends on the creation itself.
That's the other key to the puzzle.
The heart of this situation lies in players ability to both express what they want to express and explore new things to express.
If you give players something that lets them express what they want to express, you've given them a tool rather than a game. For example, Twine.
What I'm focused on instead is the idea that a game and its environment can push players to not just express themselves freely, but also explore new concepts and frameworks that give them a scaffold to continue to do so.
That's the strength of a construction game.
See, few players have something interesting they want to say really badly. There's always a percentage, sure, but it's very low. Rather than simply providing a tool for that tiny fraction of a percent, a proper construction game can lure the rest of the players into wanting to say something.
You can see this with Space Engineer and the bundle of recent space games that let you build starships. Players have to build a starship. It's so freeform that they can build more or less any ship they want. Unfortunately, the majority of starships built in these games are very poor, just functional frameworks with a few dull attempts to build something vaguely hully. The majority of the rest are clones of starships that exist elsewhere - the Enterprise, for example. Neither of these kinds of constructions are very expressive.
The problem is that the construction is too desynched from the game. The construction has turned into a tool rather than a game. The players who have built their functional frames with halfassed ship stuff stapled on are people that have chosen to not use your tool. The people who have built replica ships have chosen to use your tool to express something, but it's nothing particularly personal or interesting. There will be the small percentage of users that use your tool deeply, as mentioned before. And some of the replicas may verge on that - for example, a room-by-room, deck-by-deck reconstruction of the Enterprise.
Now let's compare that to the way Kerbal lures players into rocket construction.
If you look at various screenshots and videos of Kerbal, you can see that even among high-level players there are very, very different schematics being used. The launch platform is usually one of three specific concepts, but the actual payloads are so wildly varied that sometimes it's hard to imagine these people are playing the same game. While some of these launches are clones of existing spacecraft, that's usually reserved for parts mods rather than in-game construction.
Instead, the reason for this high level of variation is because of the high level of variation in mission objectives and procedures. A player that decides to land on the mun has a massive variety of possible steps. For example, do they return from the mun? Do they stop at several locations to gather science? Do they drop rovers or science pods? Do they have an orbiter they re-dock with? Is it crewed or drone or half-and-half?
Even if two people have the exact same mission parameters, they can still end up with very different ships because of their personal approach to the physics of the situation. The lander - is it attached to the top of the rocket, or in the middle somewhere for a skyhook-style drop? Is it affixed to the main launch platform via a central column or via a radial attachment out at the edges of a wider base? Does it have enough RCS and SAS to land with a small footprint, or has a slightly easier-to-land approach been taken by giving it a wide, flexible landing base with plenty of legs? Is the crew capsule on the bottom so the crew can just hop out directly onto the dirt, or is it on the top and there's a bunch of ladders? How have you laid out your solar panels?
There are specific mission objectives, and all the parts attached to the lander play a role in that. Their exact position and interaction with neighbors also plays a role: just attaching another rocket engine won't help any. You need to place it so that it makes sense physically.
This gives players less freedom than if they just select modules and place them wherever they like (like in Minecraft or much of Starship Engineer), but that's not a bad thing. Within the constraints lies a secret, and that secret is this:
Nearly everyone feels proud about their rocket design.
Not every rocket design, obviously. But the big missions? Yeah. People think "I built that cool freaking thing and it worked except for this part where it exploded, and it was cool."
Sure, people might feel that way about a Star Trek Online ship they've bought for real cash, or a WoW character with a rare armor set... but not nearly so deeply or often. And the reason people feel proud is the same in all these situations: they feel proud because they have found a good solution for the challenges they face.
The difference in the amount of pride can be explained by simply looking at how many aspects the player has control over. In Kerbal, you control the mission objectives, the mission structure, the piloting, and the construction. In STO you just control which prefab mission to go on, which ship chassis you own, and which slots have which gear in them. Moreover, rather than having several complex options to choose from, you normally just strive to use whatever has the highest stats. There's way less freedom.
But people do feel proud in games like STO and WoW. Very proud. The thing is, they generally don't feel hugely proud of their avatar: they feel hugely proud of their performance.
Oh, did you solo a team boss? Did you go on a 50-hero raid that worked out? Did you bring your team back from the edge of death with perfectly timed heals? Those are things that make you glow with pride.
... Look at them.
They involve choosing your own objective, choosing your own path to reach it, playing the moment-to-moment game through the mission, and making sure your character and team are properly set up to accomplish it.
All of this leads me to believe that we have a deep, deep well of creative power lying untapped. However, this is just the surface, just the marker we can look for to see that it exists. There is a lot of power hidden deeper in this concept, where today's games rarely go.
And that would be the part where you choose your own mission objectives.
See, most games are starved for mission objectives. Even Kerbal has about six, and you just point them at various planets and combine them into chains. But people often launch mission objectives that have no in-game reason to exist.
Starbases in Kerbal never had any reason to exist until recently. Despite that, people created them. Vast complexes that the astronauts could float around, even though it accomplishes absolutely nothing in-game. It is a self-made mission objective that the game cannot even see. The game cannot judge how well you accomplished your mission, because it is impossible for the game to even understand what you are trying to accomplish.
Kerbal's strength is that all of its few mission objectives are "physics judged". Kerbal doesn't say "you want to land on the Mun, you need this and this". Instead it says "this is how landing works, in terms of physics. There is a mun over there." Then it lets you try whatever you want and judges you by simply executing physics. This allows the players to judge how well things went for themselves, and also to combine complex objectives in chains and parallel missions.
Minecraft is the same way in creative mode, except that the game physics play a much smaller role. The players judge their own creations and the creations of their friends according to what they think the player was trying to accomplish. This cool castle looks neat. That giant picture of Mario is neat. Mission objectives: cleared!
All told, I think there is a lot of value in having physics-bounded objectives like those in Kerbal... but I think Kerbal doesn't do it very well. I think Kerbal does it very clumsily... and yet even that faint, wafting scent of physics-bounded self-directed missions is enough to make the game insanely good.
The question is whether you can build a game where that is at the forefront.
A game where players choose their own (usually physics-bound) mission objectives, chain and combine them in complex ways, and accomplish them using constructions where the elements are combined in relational ways? Ideally, we could even lean on mods to add more possible in-world phsyics-bound objectives, like Kerbal's communications mods.
To me one of the biggest missing elements are people. Not the concept of people, but individuals. I think there's a limit to how much people will care about something that isn't alive in some sense. Even in Minecraft, many of the constructions sort of pretend there are people around, even though there aren't any. In Kerbal, there is always a lot of chatter about making the astronauts more important, more interesting.
Just think of the Sims. It was very popular even though it was an awful game. The people in it never made even the slightest bit of sense, never behaved even vaguely like people... but because they had human faces and could be customized, players would happily feel that their house had people living in it.
... Maybe I'll publish this essay, even though it's a mess.