This post is my first attempt to write down my ideas for a game in a genre that has no name. Explanation of the genre with links to example games are in my previous post (https://sparr.dreamwidth.org/88186.html), and that is necessary context for most things below to make sense.
My primary inspirations here are Overlord (http://atrey.karlin.mff.cuni.cz/~shade/academy/), modern asynchronous board game platforms, and modern game mod systems.
Most previous games in this genre have existed in the form of a single instance that runs for a long duration. A few have had large enough player bases that the codebase of the game continues development and after one game has ended another begins with bigger and better rules, or some creative new twists. Only one, Atlantis, has directly spawned additional games based mostly on the same code that ran at the same time. I don't want to build a game that only gets played once (even if that "once" might last for 5-20 years if I am as successful as Atlantis or Eressea or Olympia), or that is the same every time it gets played. I want to build a game that is easy to modify, with a modern approach to scripted game mods that can alter fundamental aspects of the game without having to write detailed engine code to support the changes. Further, I want to build a platform that can run multiple instances of the game, and provide all the necessary tools for users to administer their own games with whatever settings they want.
The game will have a web interface and probably a mobile app. Every game in this genre has started out in a purely textual format, but then community efforts have produced GUI clients that make the game situation far easier to comprehend and interact with. Those have all been based on custom written parsers for the plain text output of the original game. There are some planned efforts underway to create a web client for Atlantis v5, but they will be pasted on. My game will be designed to produce machine readable data from the start, with transformation to pretty text or interfaces as a layer on top of that. Instructions from the player will also be in machine readable format, possibly with a parser for plain text instructions if some players want to play the old fashioned way.
In addition to the general grid of world map locations, my game will embrace the idea of smaller and nested locations that are fully distinct. Cities will exist separately from the surrounding terrain; it will be possible to guard a city to prevent entry without needing as many men as it would take to guard the whole region. Conversely, it will be possible to guard a region without entering the city it contains, which will enable laying siege to a guarded city. Caves and tunnels and portals will exist, leading to other locations in ways both realistic and fantastic, and the interior of at least some of those locations will be places units can exist without interacting with units outside.
The economy of a location and larger areas will be driven by systems that produce fluctuations over time even without player involvement. Cities will grow or decline on their own. Trade goods will become available and sell out over time, rather than being produced at entirely predictable rates.
Turns will be subdivided into discrete time slices, probably days. Many games in this genre simply handle whole turns/months at once, leading to silly situations where a unit that is otherwise capable of moving four regions per month can't make a two-region round trip in a single month if it needs to accomplish anything at the midpoint of the journey. I want a unit to be able to spend 5 days moving, pick something up, spend 5 days moving, buy something, spend 5 days moving, attack someone, spend 5 days moving, reach a destination, all in about 23 days.
Instructions given to units will have constructs to enable conditional orders so that units can behave more intelligently than just doing one specific thing or doing nothing. A scout will be able to travel until it hits an obstacle, then travel in a different direction or do something else in the obstructed location. A courier will be able to do something productive until someone hands them an item, then immediately depart in a direction that depends on the item they were given. An army will be able to decide whether to move based on the outcome of a battle, or whether to engage in a battle based on some criteria applied to the environment. Any unit performing an action with an unpredictable outcome, such as a random amount of production or a random probability of success or a random duration of effort, will be able to base their next action on the previous outcome without waiting a whole month for additional instructions.
Most games in this genre have a syntax checker for submitted orders, to tell you if you've submitted something that can't be parsed at all. Very few have any amount of order simulation and warnings, to tell you when you're doing something that may not or probably won't work, such as attempting to study a skill that a unit does not have the prerequisites for, or spending money they don't have. None can simulate orders to the point of knowing that the money you're trying to spend but don't have is money being earned or given/traded from another unit earlier in the turn. I want to simulate orders at least to that degree, including showing the player what the state of the world would look like if their orders were enacted without any other player giving orders. A stretch goal would be allowing a player to submit theoretical orders for another faction in order to see what the world would look like if those orders took place alongside their own.
Those are my thoughts for today. My next effort on this front will probably be to write a very rough draft of what a rules/intro document might look like, with a lot of placeholders for more complex content in the future, but nailing down some numbers about various game mechanics as well as perhaps describing/designing what the interfaces might look like.