RPG Active Renfield

[KleptoLizard]

You must be registered to see the links

After all this time I'm sure people are tired of waiting.
So I'll start releasing the test clients in whatever condition they are in.
Be warned that they are still extremely unfinished and buggy.
There's still a very long to-do list but I'll accept suggestions on what should be prioritized.

At the moment the new animations are extremely janky and weird, but that's due to the various overlapping IK systems fighting each other.
Resolving these issues will be a matter of trial-and-error. It's necessary to do it this way to accommodate models of different heights/bone lengths.

Let me know if anyone has questions about the client or the game project in general.
This has been a long, grueling process with numerous restarts and constant real-life interruptions.
However I plan to persevere no matter how long it takes.

EDIT: I'll give mini-updates in edits like this so I don't spam the topic
12/12/2022 - Worked on code for the big 3 magic crafting systems: Alchemy, Enchantment & Spellcrafting
12/13/2022 - Finished organizing the different schools of magic spells / effects, wrapping up the UI code for spellcrafting
- Got the fantasy name generator working, let me know if anyone wants a sample name for fun. Specify M/F
12/14/2022 - Finished the palace for the capital city. The guy who commissioned it seemed happy.
12/15/2022 - Worked on the core minigame. It'll allow for gameplay similar to Way or Haro by Ponidog.
based it loosely off minigames like Triple Triad from FF8, Gwent from Witcher3 and Caravan from Fallout New Vegas.
12/16/2022 - Installed the palace into the city. Checked all the city's walls to ensure collision was working properly.
- Various adjustments and corrections to the buildings.

 

[KleptoLizard]

You must be registered to see the links

Here's the new walk-the-map test client.
While walking the map, press L to see your coordinates.
Let me know if anyone has feedback or would like to reserve some land.

Also pictured is an early version of the multipurpose minigame.
If anyone has played Haro by Ponidog, you'll remember how the mahjong solitaire accompanied the H-scenes.
This will be similar but competitive, especially in multiplayer mode.

EDIT: More daily mini-updates:
12/17 - Sorted out and configured textures and materials for the hair and clothing. 40 items of each.
12/18 - More clothing configuration, texture and material assignment. Exported addressables for asynchronous loading.
12/19 - Worked on the overworld map. Placing various locations and generating landscapes.
12/20 - Progress on player home client. Interior decorating assets for player homes processed.
12/21 - Organized weapon models. Worked on multiplayer code.
12/22 - Implemented PNG card-based save and load system for mods. Added in-game readable/writable books.
12/23 - Worked on a short intro cutscene and an accompanying soundtrack.
12/24 - Coding practice all day

 

[KleptoLizard]

You must be registered to see the links

Here's a preview of the card-saving system.
Many will recognize this technique from IllusionSoft games.
While they use it for characters, clothing and scenery, we'll be using them for any kind of modded content.
People, places, quests, gameplay changes, rebalancing, weapons, armor, spells, animations, anything.

Download it and try it out.

12/26 - Cinematic camera code, NPC procedural generation code
12/27 - Making promotional renders
12/28 - Coding tutorials with a focus on rpg/strategy/sim game architecture
12/29 - Finished tutorials. Confident I could construct a Mount & Blade-like game.
12/30 - Sorting, configuring weapon models.
12/31 - Rebuilding integrations between core plugins (they updated the character controller suite and broke everything)

 

[KleptoLizard]

Working on a class selection system today.
Also enabling different race presets.
Lastly, implementing a birth sign system that combines your birth month (like Western astrology) and your birth year (like Eastern astrology).



1/2 - Created asynchronously loading skin material assets for drow elves, beast races. Generated weighted genetic trait charts for making custom races.
1/3 - Worked on multiplayer mode.
1/4 - Worked on volumetric occlusion shaders and NPCs in multiplayer mode.
1/5 - Worked on terrain generation. Backers get a plot of land to name and customize.
1/6 - Terrain generation improving. Continental preview map should be ready by the end of the day.

 

[KleptoLizard]

1/7 - Terrain code working. Now we begin shaping the land.

You must be registered to see the links

1/9 - Fixing some bugs with the terrain generation
1/10 - Finally fixed issues with terrain code. Trees and rocks spawning / despawning properly.
1/11 - Primordial Overworld test client should be ready later today.
[Previously mentioned issues resurfaced; test client delayed at least another day]
1/12 - Going to do whatever it takes to fix these terrain crashes.
1/13 - Hopefully the crashing has been resolved. Now shaping the overworld. Again.

 

[KleptoLizard]

Primordial overworld map:

You must be registered to see the links

Please note that this is just to see if it runs on people's systems without crashing or excessive bugginess.
It represents only the first pass of the procedural algorithm.
There's no point in proceeding unless it's stable.

Working on the overworld map. The system I'm writing for procedural landscapes and environments will use a "demi-fractal" algorithm. In other words it will start in broad strokes and work its way down but will allow injection of different seeds during the downward/inward loops.
To the center-right is an image of 256 Skyrims overlaid on Daggerfall, all to scale.
This represents the total walkable space in Renfield.
It's about the size of one of the big 3 areas in DF centered around Daggerfall City, Sentinel and Wayrest.
(this approximation excludes the sea, the mountain ranges in the corners and some of the interior regions of High Rock and Hammerfell)

Ultimately I'd like to capture the essence of the overworld map in Kenshi.
In my humble opinion, it is the finest map of all time with an ideal balance of points of interest and wilderness between.

1/16 & 17 - Working on creatures
1/18 - Back to work on the overworld map
1/19 - Hopefully finishing up the new overworld today.
1/20 - Previous map solution is unworkable. Teleportation / fast travel is working. Spending the day refactoring.
1/21 - Editing audio files. Setting up a custom audio solution for the H-scenes. Finishing first pass of new map solution.
1/22 - Finished labeling the overworld location beacons.
Finished basic timing system for H-audio.
Worked on dynamic H-text generation.
More work on processing audio files.

 

[KleptoLizard]

1/23 - Editing audio files. This might take a while since I have hours of moaning to process into clips.
1/24 - More audio editing.
1/25 - Exporting audio for the H-sequencer demo.
1/26 - More audio exporting. Localizing text for H-sequencer into other languages.
1/27 - Post-export audio editing and volume balancing. Gathering and sorting visual and audio effects for magic spells.

 

[KleptoLizard]

New Overworld Test Client
Fixed various issues, a great deal of refactoring.
Now has fast travel.

You must be registered to see the links

1/28 - Fixing various issues with the map. Expanding functionality of card saving system.
1/29 - Tweaking animations for the H-demo. More terrain bug fixing. Practicing Cangjie keyboard input for quicker Chinese localization.
1/30 - More overworld work. Fixing procedural animation issues.
1/31 - Sorting and editing sound files for H-demo.
Tomorrow on the 1st we'll be streaming on

You must be registered to see the links

2/2 - Still editing the sound files. 144 left.

 

[KleptoLizard]

Today I rendered a warrior who is unsteady in heels and a ballroom gown.
Clothes like shirts and pants are easy to get working.
The dress needs to rescale at runtime to avoid clipping.

2/3 - Finishing the sound file editing. Slight refactoring of audio sequencer code.
2/4 - Sound editing taking longer than projected.
2/5 - Finally wrapped up sound files. Audio editing is one of the worst asset management-related tasks. It requires your ears for the quality and content, requires your eyes to study the peaks. No listening to an audiobook in the background to ease the boredom.
2/6 - Tweaking IK targets for H-animations.
2/7 - Finishing animation tweaking. Practicing 2D art techniques including gif creation.

 

[KleptoLizard]

This game will come with its own original setting and story.
But underneath that, it is intended to be akin to a universal RPG system.
Think of how many paper-and-pencil RPGs run on D20/GURPS/Palladium/Year Zero etc.
Not to mention the flexibility of AD&D itself. (the recent dark clouds gathering over the OGL notwithstanding)

A great opportunity awaits would-be writers who've wanted to create their own setting,
but lacked programming ability to make their own interactive experience, lacked the time to write out a full novel,
or lacked the artistic hand to make sourcebooks which showed the realms they could see in their mind's eye.
I'll be actively working with modders to meet them halfway with tools that make overhauls and original setting mods much more achievable.

Any blend of fantasy and science-fiction is possible.
Any category of "-punk".
Westerns.
Make a Brütal Legend-style surrealist world around a genre of music.
Formulate parody crossover realms between your favorite IPs.
Call of Cthulhu-style investigation adventures involving sanity mechanics.

Renfield's beginnings will be humble, but I envision it being an outlet for anyone who went on a forum saying:
"You know what game they should really make?" "Wouldn't it be great if you had X, Y & Z all in the same game?"
Obviously making these expansive visions will be a lot of work no matter how you approach it.
But even after you filter out people who are purely daydreaming, the lack of a solid framework is the key obstacle.
Barring that, virtually any game you'll ever mod has at least some things that are hard-coded and difficult to work around.

However we're building this game from scratch.
We can allow flexibility in terms of stats, skills, degree of randomness vs determinative mechanics, all combat and other formulae...
Let's divorce ourselves from the limitations that have been placed on us by the industry.



P.S. The hair on this fairy looks terrible. I don't think I'll ever use this particular wig again.

 

[KleptoLizard]

I think it needs some adjusting, almost seems to be floating above the head.

It looks like the hairstyles from both Dumb & Dumber characters magically fused.
I'll spend more time on the scalp fits in the future.

2/8 - Calibrating jiggle physics.
2/9 - Merging subsystems for the H-demo into a single project. Hoping to have the demo ready in the next few days.
2/10 - Last minute tweaking of the H-demo. Going to release it tomorrow whatever state it's in.
2/12 - Adding facial expressions. Fixing long list of issues with sample client.

 

[KleptoLizard]

Dresses like these can have full or partial cloth physics simulations.
However for this render I just introduced a bit of noise to simulate movement.

Speaking of physics simulations...
Someone on Twitch asked me about tentacles.
These are actually really easy. Much easier conventional sex positions.
This is because you are just using a rope physics simulation with IK/FK.
If you want a monster to have sex but its physiology would require new animations, you could just opt to have it use a tentacle.
Essentially what they did in Urotsukidoji.

Cats, dogs, horses, dragons and a few other creatures will be based off highly morphable base models.
All sorts of different breeds of each are possible.



2/13 - Finishing facial expression system. Fixing bug with jiggle physics. Studying Python language.
2/14 - Finishing Python study. We'll be using Ren'Py as a 3rd party program to make quest mods. More fixes for the H-demo.
2/15 - More Python study.
2/16 - More Python study. Rendering.

 

[KleptoLizard]

Some of you might be wondering why I'm studying Python in the middle of everything else.
Let me explain briefly.
Once the H-scene client is cleaned up a bit more users will be able to do animation playback with a few options.
But of course the entire goal of the project is to wrap an adult scene playback core with actual gameplay.

The quickest way to open up possibilities with Renfield's interactivity is to fast track user-created content.
Since Ren'Py seems hands-down the most popular engine for making small scale H-games in visual novel form,
it makes sense to either:

1) Use it as a 3rd party tool to make scenarios.
2) Use it as a template for a custom mod editor we make ourselves.
3) Some mix of 1 & 2.

Therefore I'm going to be spending a while learning the ins and outs of Ren'Py to understand how
to get experienced visual novel writers to feel at home with RF's modding framework.
The first step will be to link the script exporting feature Ren'Py has with RF's dialogue system.
You'll be able to give the various needed commands by designating a character that executes them.

This Technical Narrator will be a character that allows you to fit commands into the script itself.
The exported script would look something like this:

Gloucester "Drive, thoughts, down to my soul: here Clarence comes."
TechNarrator "Enter CLARENCE, Enter BRAKENBURY"
Gloucester "Brother, good day; what means this armed guard that waits upon your grace?"
Clarence "His majesty tendering my person's safety, hath appointed..."

You can call the Technical Narrator whatever you want.
Its quotes won't appear as dialogue, so it's different from Ren'Py's normal disembodied narration.
Instead it acts like an escape character in a String literal.
The quote will be parsed and treated like a command.

Syntax for the commands will be based closely on simple Ren'Py functions.
However some of the defines such as characters and custom UI elements might be better handled
with Renfield's character creator and other tools we make ourselves.

2/25 - Modeling software tutorials
2/28 - Sorry for the updates coming less frequently.
I've been slammed by a lot more work lately.
The project is in a good position at this point but it's been a struggle to find time for it.
As always, let me know if anyone wants more info about the game or about their chosen territory.
3/21 - Alright. The initial glut of additional work has eased a bit. Getting back to daily updates.
3/22 - Tweaking hair models and race presets.
3/23 - Working on spells and spell effects.
3/24 - Patching some data structs, working on some IOStream-based scripts for loading information from all these txt files I've accumulated and parsing it all into something usable.
3/25 - Finishing up the combat formulas
3/26 - Environmental assets. Modular village buildings.
3/27 - Fixing some issues with broken dependencies from yesterday. More environmental work.
3/28 - Refactoring and cleanup
3/29 - Processing spell and activated ability data. 432 individual entries so this'll take a minute.
4/3 - Working on unique NPC data
4/4 - Working on shops and economy
4/5 - Working on armor and clothing
4/10 - Clothing, skin, bodies, hair
4/11 Clothing, jewelry, items. Data entry.

 

[KleptoLizard]

Well the H-client's core functionality is all working. ...and I'm entirely unhappy with it.
The different subsystems have very poor synergy and rhythm with each other.
It reminds me of those MUGEN mods with how jarring the sound and animation can be.
Yet with time and fine tuning it will get much better.

More importantly, I'm nearing the mid-point of a day-job contract which upon
its conclusion should give me the resources I need to start hiring help.
Having more people on the Renfield project will result in a drastic change of progress rate.
I appreciate everyone's patience; we're nearing a point where we'll have something playable both in terms of H-playback and as a proper game.

For a long time it was like playing Starcraft with only one SCV.
If all goes well I'll be able to recruit some talented people I've met over the years doing freelance work in the indie scene.
The project will move faster and I'll feel a lot less burned out.

4/12 & 4/13 - Collecting, sorting and modding Mixamo animations. Practicing animation blending/weighting/masking/layering techniques.
4/14 - Expanding monster race presets and morph targets.
4/17 - Locations, especially ruins
4/18 - Save system & modding framework
4/19 - Weapon models
4/20 - Reconfiguring metadata for asset library
4/21 - Rendering some banner and UI art
4/22 - Graphic design practice and tutorials
4/24 - Coding the Ren'Py file parser
4/25 - Studying C# memory API
4/27 - Installing female beast race morphs. They'd previously been corrupted but have now been repaired.
4/29 - Working on world items, minor terrain features.
5/6 - Over the last week I've fixed various issues with the sound system and voice files, made tweaks to scenes, various other fixes and improvements.
I've also been examining the Mount & Blade: Warband modding system.
The modules are very convenient when installing other people's mods, but actually making mods can be quite convoluted.
5/8 - Typing out a unified namespace for all the various needed data classes, listing all the needed properties.
5/11 - Coding the Renpy parser is going well.

 

[KleptoLizard]

As part of the process of developing a modding kit for Renfield I wanted to take
a fresh look at some of the more successful tools and systems in the industry.
The two main contenders were the Creation Kit used for Skyrim / Fallout 4
and the module system used in Mount & Blade titles, specifically Warband.

What I found played out like a debate between two compsci professors on the
strengths and weaknesses of the coding patterns Composition & Inheritance.

The Creation Kit and the Plugins (ESPs) are a case study in puritanical Inheritance.
A typical bandit encountered in the world is often the result of 10 steps of filtered
"Swiss cheese" selective overwriting and at least one leveled list.
Then the plugins go in sequence overwriting each other allowing for a wide variety
of mods to be blended together but with massive, looming stability / incompatibility implications.

M&B's modules are highly convenient for installing overhauls and switching between them.
Just like an actual mod manager, however the ease of combining smaller mods is lost.
Should you tweak the compiled txt files (which everyone does) you'll run into complications
if another change requires you to recompile the Python scripts.
Furthermore so many values in M&B's scripts are indexes to arrays formed by the txt files.
This is tantamount to a hard-coded spaghetti throwback to the bad old days of BASIC.

Clearly a system can be envisioned that combines the strengths of both these extremes.
However the main weakness that both modding systems have is their lack of null-checking.
Each system could avoid the vast majority of CTDs if the code looked before it leapt.
It's the difference between a mesh not loading but substituting a question mark
as opposed to simply crashing without warning or explanation.

This is especially evident in Skyrim where you could be walking through an area and
some content will attempt to load in a place where the player cannot even see.
Every get() that is replaced with a tryget() in the code will boost the overall stability
tremendously and speed up any debugging processes, making modding much less painful.

5/15 - Progress with the behavior tree system
5/16 - More work on modding tools.
Unlike many in-house mod kits available in recent years which have pop-up/pop-out windows, RF's kit will feature a layout with multiple docked windows which talk to each other.
Now it'll be much simpler and more ergonomic to configure new actors, equipment, factions etc and weave together their relationships via drag-and-drop.
5/20 - Experimenting with OpenBRF & NifSkope
5/21 - Refactoring plugins to avoid needless dependency chains
5/22 - Modeling, texturing, rigging
5/24 - Fixed a really persistent graphical glitch. More refactoring.

 

[KleptoLizard]

On occasion I'll get feedback asking "Can you make the girls look like this?"
The answer to all of these inquiries/requests simultaneously is yes.
Whatever faces/bodies/styles you want, we can find a way to accommodate.

I've been looking carefully and modding systems for IllusionSoft, Bethesda and Taleworlds titles.
Illusion has in-game character creation and "studio" clients to pose in.
But any extension of these systems involves outright hacking the game.
Illusion is often the butt of jokes vis-a-vis how they could create the ultimate
H-game but never will because it would put them out of business.
[Google "The Illusion Grief Cycle"]
Their message to the players/users is "relax guys; the game will have all the customization you could ever want."
The users then proceed to retroengineer nearly every aspect of software.

Bethesda goes for an almost entirely inheritance-dependent paradigm.
A bandit in Skyrim that you encounter in-game is about 10 levels of inheritance down from a base abstract bandit.
[Bandit Base] becomes [Bandit Subcategory] becomes [Bandit Caster Base] becomes [Bandit Caster Subtype]...
And this system can be convenient for making changes to an entire family of units with a few changes.
But the drawbacks become readily evident when trying to pluck something from one of these trees.
The interface of the Creation Kit showers you with popup menus and lacks many
features for seamless side-by-side or bird's eye viewing of related content.
Meanwhile the strengths and weaknesses of ESM/ESP files could be a book all on their own.

Mount & Blade by Taleworlds has a really clever, highly flexible... and incredibly obtuse modding system.
They seemed to want the best possible experience for "modders" (i.e. people who install and play mods)
while also holding an undying hatred for the other type of "modders" (i.e. people who make mods).
Theirs is a wholly composition-based paradigm that compiles content via a Python process into a series of txt files with arrays.
It also includes modules in separate folders which provides natively the function of many mod organizers.
Many talented people have created amazing, in-depth overhauls of the various Mount & Blade titles.
(Keep in mind that this is all based on Classic, Warband, Fire & Sword etc., not Bannerlord which has its own kit with which I'm not yet familiar)
Yet splicing content from one module to another can be a very delicate operation.

After examining these different systems I've come up with some modding tools for RF that will take into account all these lessons.
Among these tools will be a node-tree system for getting your characters to look the way you want.
Whenever I would make new characters or download character cards for an Illusion game,
I'd make the same handful of tweaks to give them more of an hourglass figure.
But with this node-based system I could take a character card and drop it into the folder which populates the game world
and the game will automatically make those adjustments to every female character.
Textures, materials and shaders will all be accessible.
Body types can go through Add/Subtract/Average/Lerp/Floor/Ceiling/Clamp/Multiply/Divide
operations so the distinct variations will be present but they'll all be "drawn in your style" so to speak.



6/11 - Reorganizing data fields, deepcopying data in various places, eliminating redundancies.
6/15 - Reassigning clothing surfaces. Often a clothing item has parts that should really have a separate shader. Like metal or gems on a mostly cloth item.
6/16 - Worked on user-accessible decal system. Useful for tattoos, tabards, flags and scars.
6/22 - Converting outfits with a focus on shoes/boots.

 

[KleptoLizard]

One of the key tools I'm working on is an in-game image editor. You'll be able to mod textures/materials without any 3rd party programmed preview it in real time. Looking at this above example, you'll see the hair has these tails that I wanted to turn off. On a typical wig these would be a submesh that you could just switch off with one click.
But for some clothing and hair there are situations where you'll want to trim it in a specific way. By trial and error it the areas on the opacity texture can be verified and those tails made invisible. The new coder (who will join at the end of July in a long-awaited and very welcome addition) will be able to make a convenient system of clicking on the mesh and getting the UV coordinate.
Also, users can now click on the face/body itself to call up blendshape handles in a manner familiar to players of Fallout 4 or Black Desert. This is a great improvement over the character creator I'd previously released that just had lists of morphs across several menus.
Remember; don't be shy about asking questions.

6/28 - Finishing sets of basic animations for the adventure mode. (Running/Attacking/Dying/Idles etc)
7/4 - Focusing on animations and armor sets.
7/7 - Experimenting with shader and post-proc combinations. Saving presets.
7/10 - Working on buildings.
7/12 - Working on character presets.
Project backers, even former backers, are invited to make requests for custom NPCs in their assigned domain.
7/13 - Outfit coordination and implementing backer feedback.

 

[KleptoLizard]

A few things about game development are so cumbersome in terms of
time investment that to implement it all on one's own is not a great idea.
Environmental creation is just one example.

I've spent years making overworld maps/cities/interiors/dungeons
for the singleplayer mode of RF with many reboots and restarts
and while I learned a great deal, I wouldn't wish this on anyone.

So for the modding tools I'm trying to make things as user-friendly as possible.
The assets I'm using exhibit a huge range of possibilities with thousands of individual meshes.
By default these are textured with a very generic medieval theme but they're easily changed.
There's really no limit to what you can make.

Beyond that I'm developing various quick generation tools.
Rather than place each wall and floor you can use brushes to make a structure in seconds.
The system will pick out and place the appropriate meshes and place everything dynamically.
Users familiar with 2D sprite-based terrain drawing will understand what I mean.
Just as many sprite terrain tile sets identify and reconfigure center, edge and corner pieces
the toolset will make generation of structures and rooms relatively painless.

Another method I'm experimenting with is a SimCity style "zoning" system.
By laying out the roads and zones how you like, you can then just watch the buildings sprout up,
trimming and tweaking them between generation steps.

7/19 - Drawing up presentation materials for the new team members
7/21 - Drafting siege engines and mechanics
7/22 - Working on dungeons and the encounter/treasure tables for them
7/25 - Working on a more stylistic overworld map image

 

[KleptoLizard]

I shudder to think about the time and money wasted on the previous audio solution.
But now AI is a thing and the last key technical obstacle to RF's development has fallen.
There's a lot of opposition to AI from people who worry about its impact on talent.
However no matter how many times I tried it simply proved impossible to find
affordable paid volunteers who would turn in consistent quality audio files.

Once the new files come in we just need to clean up the UI and tidy the code.
The procedural animations are functional but very wonky and awkward.
Fixing them will be a matter of tightening / loosening the various competing IK target weights.
Nevertheless we'll have a working demo at long last.

9/16 - Well the initial pass was horrible. Sounds like she's singing in some creole of Cherokee and Tibetan. Will keep fiddling with the settings.
9/17 - Alright after exploring various AI options and getting estimates from voice actor hiring sites I've decided I just can't afford a new voice solution at this time.
We're just going to move forward with the other stuff.
9/18 - Filling in data sheets
9/24 - Finished filling in character, npc and enemy data. Now filling in spell data.
9/28 - Writing some IOStream code to automate file sorting and renaming
10/6 - Finished exporting temporary voice files. More spreadsheet data entry.

 

[KleptoLizard]

While my new teammate finishes up the demo (which we're hoping to have ready by the end of the month)
I'm keeping busy drafting various mid-to-long term design features.
It's no secret that the Civil War in Skyrim left much to be desired.
(Too busy to play Starfield but rumors of it being underwhelming abound)

If we're going to make a poor man's TES game we might as well do it right this time.
Koei games set in China & Japan have provided me with a lot of inspiration for how strategy and roleplaying can synergize.
They produce a lot of button-mashing Warriors games that are very character-driven in their appeal.
Then they have the Tactics and Kessen games which focus on battlefield command.

Finally they have the Romance of the Three Kingdoms / Nobunaga's Ambition grand strategy games.
Each of these 3 layers (personal combat + army command + nation command w/ sandbox rpg elements)
have overlapped in various titles, such as the Warriors: Empires spinoff games.
(special mention to the non-Koei title Kingdom Under Fire which seamlessly combined the first 2 layers)

But what's missing is a game that combines all of these modes together in a fun, deep, dynamic way.
Renfield will let modders declare factions, battlefields, strongholds and settlements that can change hands.
Various lowpoly versions of character models will populate the battlefields for (hopefully) performant fights.
Meanwhile the flow of the wars will be non-linear (unless you script it otherwise).

10/15 - Working on consumable items & gathering icon art
10/18 - Sorting UI icons
10/19 - Checking reference paths
10/20 - Finished picking out spell icons and finalized magic schools
10/23 - Random world events and location event tables
10/25 - Refining regional event tables and loot tables.