Expansions

My thoughts on game expansions have changed over the last few years. I used to consider elegance the most important aesthetic virtue, I believed that every concept should be expressed in its simplest possible form (but no simpler, of course). Expansions necessarily add more to a game than is essential, and so were not to my taste. Content-based games are a separate category of thing - finite content is eventually exhausted and must be replaced - but a systemic game can last indefinitely, and so never needs to be expanded.

I particularly disliked expansions for board games. They tend to make the games longer. They drastically increase variance and streakiness by increasing the size of card decks (a larger deck is less reliable even if the proportions of card types are kept constant), destroying the delicate balance of the base game. They usually have more complex rules, making them harder to teach and slower to play, for little (if any) gain.
Carcassonne is a particular offender in my experience; manifold expansions churned out for profit with no respect for what makes the core game good. A lot of the strategy in base Carcassonne revolves around knowing what tiles exist and forcing your opponent into configurations that are impossible to resolve; the expansions weaken this by making all configurations possible - you can still try to play the odds, but the strategy largely reverts to maximising your own score rather than spoiling your opponent's position.

But I've learnt that this isn't the whole picture. Many expansions are poor, but this doesn't invalidate the entire concept. Elegance is important, but it can be worth sacrificing some to gain in other areas. Expansions can be done right if you're careful. Add things because they're worth adding, not just for the sake of having more stuff. Understand what's good about it in the first place and preserve that, enrich it, don't break it.

Race for the Galaxy is a perfect example of a game that didn't need to be expanded - it starts with a beautiful system, finely balanced, stable under hundreds of plays - but which grows even deeper with the expansions. This is the game that really sold me on the concept. Designer Tom Lehmann has written about the thought that went into the expansions. Powers that played well but were complex to learn were withheld for later expansions, allowing players to master the depths of the core game system first. The power level of strategies in the base set was skewed slightly towards those that were more difficult to grasp, to encourage players to persist with them - and then balanced back to the centre in the expansions. Close attention was paid to the problem of increasing deck size - expansions add more "explore" powers allowing players to select from more cards, and the third expansion adds a "search" action which allows each player to trawl the deck for a card type of their choice once in a game.

In my own work, I'm finding expansion-type design delightful. Inventing new games from scratch is more important to me overall, but there's a playfulness to building on top of an existing structure that's very satisfying. And it's a lot easier to get into when motivation is at low ebb.
What makes a good expansion depends entirely on the game, but there are some general questions I've been asking myself. Where can new content fit into this? What regions of the design space are so far unexplored? What parts of this system are static that we could make variable, give the players control over? What unexpected corners of the rules can we interact with? What things aren't connected; can we join them up somehow?


I've been messing with Glitch Tank again in the last couple of days. Similar thoughts to when I patched it before. Should have a patch up soon with some bug-fixes and a little bit of new stuff. It's a very tight game, it doesn't have a lot of room to expand, so I have to be careful.

Where can we inject new things?
- Through the initial state. Different map features. The update I mentioned before added walls and teleporters generated at random. I'm not adding any more map features, there are enough.
- New game modes. Started with just one mode in the Kompendium version, added AI and explicit support for turn-based play for the iOS version. Added the 6-hp mode in an update.
- Action cards. In addition to the eight standard actions (forward, back, left, right, fire, jump, mine, replicate), there are an undisclosed number of rare actions. The most common of these flips all positions on both axes; the others you'll have to find on your own. They're not a major feature - you might never see them at all - but they add a bit of occasional variety, and enrich the "glitch" theme. I've added at least one more action, which I'll explain when the new patch goes live.
- Random events? It probably wouldn't suit the game very well to have events happening independently of player actions, but it's a possibility.
- Events triggered by an obscure game event, an unusual corner case in the rules - something that players have control over but wouldn't normally do. Like the glitch that happens when both players die simultaneously.
- Metagame structure: a tournament mode, deck customisation, character progression. Might not suit the game, but worth thinking about.

I'm very happy with the things that have been added, they're definite improvements. There's a real value to leaving a design to simmer for a while, you can come back to it with fresh ideas and deeper understanding. We've been playing this game for almost a year now, so I know it quite well.

      
The first Race for the Galaxy expansion came with some blank cards making it convenient for players to mod the game. I spent some time a few years back experimenting with these, trying to make cards that did things that hadn't been done. Modding someone else's game is a similar kind of design to expanding your own.

The phase-selection system affects the rest of the game but nothing directly affects it, so I tried to make things interact with it - letting you restrict your opponent's options, see their choices, restrict your own options or reveal your choices as a cost. None of the cards I came up with worked very well, but I'm quite pleased with a tournament variant. Apparently the designer had some similar ideas - the third expansion contains Psi-Crystal World. (Part of the reason this works better than my attempts is because it's on a world rather than a development - the effect is balanced by opportunity cost, not just by numeric cost.)

A lot of the game revolves around building engines that produce and consume resources. Different powers on different cards form engine components - producing goods on one world with a power from somewhere else, triggering a card draw from another power, then converting them to points on a different world. Resources come in four different flavours, with different powers synergising with each. I had the idea to have a "wild-card" resource, which would count as any flavour and allow you to connect up powers that wouldn't usually interact. Through a competition, this actually ended up in the official game as Alien Oort Cloud Refinery, which I'm still totally stoked about.
(If you know the RftG rules: the reason it's a windfall world is to allow more powers to interact with it, the reason it can't be traded from even though that would allow even more interactions is that otherwise it would have to cost so much that you couldn't afford to take advantage of all the interactions anyway.)