As a modder or a player, what would you expect from a (semi-)automated scenario generator ?

[Gedemon]

By "scenario generator" I mean some code that will automatically place cities and infrastructures (improvments, roads, districts, ...) on any map using rules and some global settings (set on the setup screen) or civ-specific settings (set in XML/SQL files)

I'm currently developing a test version, my goal is to have the ability to quickly create a game (or a base for a hand-made scenario using the WB) with a pre-set world, either with small settlements and still space to expand (thinking Anno Domini for @Rob (R8XFT) or @SeelingCat maybe ?) or a full modern "what if" world (using the "Relive World Wars" by @Allein, the "Modern Basic Civilizations", or other modern world mods), or, why not, fantasy mods on related maps (GoT,...)

It should be able to work on fully random maps (expanding from Starting Positions), on YnAMP's Terra Map (from random positions or using relative "TSL"), on any classic TSL map (expanding from the TSL, or using a city map when it exists by cross-referencing the City Names list for a Civilization), or any fixed map (with TSL or not) using one of the previous methods with an additional "import" method (from a placement table with coordinates)

Here are some of the variables that I plan to use (either globally from the setup screen or civ-specific from a XML or SQL file) and are mostly coded in:

- City placement type: "generated" (ie based on plots "fertility", weighted with distance from capital), "city map" (for maps that support real position naming), "import" (for maps that have a placement table with coordinates) or "mixed" ("import" first if available, then use "city map" then use "generated" until there is no positions available or the max number of cities is reached)
- Max number of cities ("max" because you may get less depending on the available space on the map), including City-States (a "City-state" may control multiple cities this way, but you'll have to give it a city name list...)
- Capital size
- Other Cities size (with an option to decrement it so that first placed cities are bigger than the last placed)
- Borders type: "none" (only first ring around cities is owned), "expand" (looping on all cities, from city center, one ring further per loop, until a max distance is reached or when there is no free territory left on the last tested ring)

I do plan to add similar options for infrastructures (improvments, roads, railroads, buildings, disctricts per cities, ...) either using fixed values or calculated values in relation to a city size.

The way I'd like to code it, I should be able to create a single scenario mod with just a config row and and a city list for each Civilizations (without coordinates) that the scenario generator could use on multiple YnAMP maps (Terra, Giant Earth, Largest Earth Custom or cropped regions of the last two)

I plan to allow multiple rows for Civilization settings (num cities, etc...), one for each Era, so that such a scenario data would generate more or less developped Civilizations depending on the era you start in (think of a Historical Spawn Date mod without the need to wait 350 turns before playing as the USA)

So while I'm still on the design/early coding, I'd like to know if there is an interest in such a tool (or if I'll be the only one using it), and what you'd like to see in it.

 

[Zyxpsilon]

I'd surely like to experiment with such a tool and eventually.. tackle everything it has to offer (potentially with even more aspects later in iterated versions).

Although complex if not outright chaotic in terms of ruleset mechanics (basic or otherwise presumed balanced enough to sustain challenging conditions), i'd recommend keeping an open-ended model (and fully customized in some pre-deterministic steps) as far beyond the Firaxis design(s) as possible.
It really should leave plenty of room for highly variables features & context as they both can generate interesting AI/Human situations on top of regular tactical/strategic systems **we** always seek out for either consciously or within the RNG factors inherent to the Victory cycles & other major aspects (yeah -- Diplomacy, i'm looking at you!).
Knowing that.. here's a short list of supplemental concerns based on personal opinions or actual gameplay experience (low or extensive in many degrees).

1) Fill out the conditional range of city grids (tall or wide) with parametric polygrams that "extend" the static Hexagonal/Shaping (3+ tiles limits up to 37 total, etc) blueprint of valid development per Citizen total/spots. Triangular, Trapezoid, Diagonal.. you name it. NOT changing the rules directly.. but rather letting the expansion potentials reveal themselves in a visual manner that indicates where optimal choices can or should be made.

2) Bring up an interactive component that can allocate "urban planning" templates (global or one-city based) in various principles; Scientific, Defensive, Productivity, Religious & so on. No matter how we prefer to play or inevitably expand our empire.. there seems to be a constant process where familiar conditions are simply reproduced (Capital and regional hubs included). This sort of coordinated design has great impacts on success or defeat even at the point(s) of selecting a core Victory target. Reaction and/or prediction are two different but important concepts. IMHO.

3) Leave enough tiling spreads for specific map styles. Example; Continent vs Archipelago --- very particular development steps need to be calibrated on Navy and/or a few more logistic plans. Thus, a minimal (but generic) set of custom rules has to be grafted on any given models (as defined by #2 above) while providing enough freedom of expansion towards whichever happens to be necessary. Yup.. here's my React/Predict motto, again.

4) IMHO, "Scenario Generating" is **NOT** really about throwing chaotic gameplay conditions at us.. more like offering a puzzle to solve in a little-less random environment/map, AFAIC. Sooooo -- the chance is there to purposely create a worthy challenge with cool intellectual conditions i'm sure everyone will start posting about below.

Secondly...

I've already discussed something with Zur13 (author of Radial-Measuring-Tool.. as an example) about City-Grid placements of Wonders/Districts (plenty of online details about that!). Pet-peeve of mine as i'm also sure a dynamic/interactive tool could be created to somewhat help in such tasks. Try considering that idea too within the scope of your upcoming generator. I'll explain further if you should ever.

In the meantime, good luck with this & don't ruin the opportunity, please!

 

[Jkchart]

I'd honestly really like it, especially because I think it could make some interesting "alt-history" set-ups to go with your Earth maps. I was a fan of Rhye's mods and yours (especially with this idea) is the next closest setting, so I'm all for it.

I do wonder how it would work with things such as religions. In any case, I'd be excited.

EDIT: If you want someone to ever test run a game as well with it when you get more comfortable with your generator I'd love to try it out.

 

[Gedemon]

Thanks, latest version of YnAMP (github and Steam) include the first options for cities and road placement (international roads not working yet)

By religion, you means giving faith on start or directly adopt a religion?

@Zyxpsilon, you mean predefined geometric placement? And not following the map?

 

[Zyxpsilon]

Yep.. something like that.
I just jumped into your beta sampling phase only to realize how complex this project should become over time -- for good & certainly ambitious reasons!

-- The feeling is really strange to start up on a "randomized grid" of variable cities. As if the Civ ruleset model has been completely reworked with wacky conditions. We just need to adapt to even more RNG patterns and throw the familiar pacing of normal "Capital" loneliness out the window.

-- Many options to test out for sure... as there's no direct feedback predictions possible in such context(s), we'll have to compound evidence of specific results in as many variations as "wished for". Again, that's a weird collision with the new ruleset impacts.. our minds (with multiple years of gameplay) need to shift the Victory planning steps on a chaotic set of paths deliberately spoiled at the start.

-- First suggestion == try modeling the default Era start(s) systems with their enhanced field assets (units & buildings, etc) as the basic grid/map situations are getting created within the scenario cycles. Ideally, based on city counts (if not pre-defined by hybrid gameplay situations such as Scientific goals or Religious campaigns, etc). Since the settlers phase is discarded.. we'll (instinctively) aim to represent a logical pseudo-progress *IN* our ergonomic but still shuffled up empire. ((EDIT .. Oooopppsss, you've already mentionned this sort of feature in your initial comment!!)

I like what i'm seeing so far. AIs react almost as expected, btw.

 

[chrisgatt7]

Any update ?

 

[Gedemon]

On hold ATM as I've switched back to my overhaul mod now that a development-stopper bug as been fixed in the last patch.

But still planned to be expanded, as I will need it for said overhaul when I'll start working on real spawn dates.

 

[Jkchart]

On hold ATM as I've switched back to my overhaul mod now that a development-stopper bug as been fixed in the last patch.

But still planned to be expanded, as I will need it for said overhaul when I'll start working on real spawn dates.

Fascinating, what bug was that?

 

[Gedemon]

this one