Devlog #8
Discovering a golden feature is hard, especially when you are building a genre from the ground up, and there’s not a lot to reference.
Basic Weekly Updates
Happy Saturday everyone! We did get stockpiles in, but they kind of opened an unforeseen skill-tree if you know what I mean. To hear more about that issue, head over to the next section. We’ve been working a lot on media-kit type stuff, and doing some art stuff as well - making some moves, trying to ramp up our model production a bit, blah blah. Basically we’re still on track! Now let’s dig in.
Automata Conundum
We say that our dev cycle is longer than it really is - in reality we are always playing our game, iterating on recently built features, and tweaking along the way. Sometimes this is minimal (in the cases where our first guess is pretty good), other times, a mechanics nightmare starts to unfold we would have never expected. So here’s the situation we had.
In your normal RTS, city sim thing, all your AI’s have some centralized register of where this or that resource is stored and how to get it. Using this they can easily route themselves around the town, efficiently using and restocking resources. This doesn’t work in Avalon - you need global and shared automation-spaces. AI’s need to automate inter-player trade, what’s more, one player may have 2 groups of AI’s in different parts of the world - you wouldn’t want your settlers walking all the way back to your capital for a piece of wood. So we needed an intuitive analog to shared resource indexing that worked inter-player and in a decentralized way.
Our first stab at this was that every AI simply looked in its immediate surroundings for stockpiles every time it needed something. It was simple, seemingly could be emergent - what could possibly go wrong??? It was a nightmare. You could place a construction site down outside of the range of your wood stockpile - wood is right over there, but now you need to automate a who supply chain just to build the building. You could make these survey areas larger, but at scale, this was gunna cost a lot. We had to make an efficiency decision to avoid massive, frequent queries to spatial.
Where we ended with feels natural for Avalon, and it actually solves other mechanics issues that it wasn’t contrived for, so I’m really excited about it.
Town-Tile Claiming
So in the new system, to automate anything permanent (that can persist and automate while you are offline) you need to place your buildings in a town/district/settlement/whatever you want to call it. You can found a town/settlement by building a settlement/town center building, which automatically claims a certain amount of tiles around it. From here you can only build in the claimed tiles, each building you build further increases land claim. For every building in this district, a local index of storage, etc. exists and AI’s can automate within them. Using stockpiles, you can set AI’s to transfer resources between these districts. Additionally, players within the same town/district can setup automated purchases between each other.
At the same time, this system begins a conception of war and empire building mechanics. The town centers control town automation and basically the AI in them, so capturing a town center is a natural mechanic for conquest. We’ll talk more on these implications later as we unfold exactly how to use them.
So What’s Next
So we’re currently working on implementing this tile-claiming system, it’s a good 65% done I’d say. We’re polishing up a lot of UI stuff as well. Once this town system is ready, we can throw in repopulation and food pretty easily. From here the only thing standing between us and stage 1 is well... a wall. Yep we need walls and defensive buildings. But that’s it! then we’re gunna stop to play a bit, confirm stage 1 of ACDF and start rolling into politics, banners, chat, and of course, formal conquest.