Earth Evo Devlog #1: Camps, caves and nine playtests
Earth Evo is a Civ-style 4X where every conspiracy theory is true. The full game is a long arc from prehistory onward. Right now I'm building only the first act, to answer a single question:
Is it fun to build a civilisation, lose most of it to the meteor, and then dig up your own past?
Everything in the Act 1 prototype exists to test that. Tactical combat, aliens, a real-Earth map and multiplayer are all out until the hook is proven.
The first milestone, M1, has now passed its gate. It took nine playtests, and each one ended with a one-line note that became a tested rule. The hook itself is still unproven, and this devlog covers both.
What M1 set out to do
M1 had to make the first 20 turns playable: a map, camps, a food loop, and real input from the player. Its gate was blunt: "20 turns and it's not annoying." It passed on 26 October 2026, after nine playtests. My verdict on the last one: nothing annoyed me, but it felt a bit empty. More on that below.
Before that, M0 set up the meteor rule and the naming system: 70 Proto-Indo-European roots for your people and 30 Proto-Uralic ones for the Neanderthals. Neanderthal names borrow Proto-Uralic roots as a flavour choice. Nobody knows what Neanderthals spoke, and the game isn't claiming they spoke this.
What got built
You don't start with a city. You start as a band of people walking an unknown land under fog of war, and the first job is to find somewhere worth staying.
- Making camp is a choice. Your first camp becomes your capital, so you scout for food, fresh water, hills and caves before you commit. Making camp costs one move.
- Urgency. A band carries food that runs out, and only rivers are drinkable; the coast is salt water. You can't wander forever.
- Camps grow and send people out. Each person works one hex. Camps send out lone scouts to explore, and bands of settlers who carry food and materials to found new camps or help another camp build a wonder.
- Other peoples. A Neanderthal tribe and rival camps keep patrols on their land. Fights start from Freeciv/Civ rules: you attack by moving in, terrain defends and zones of control matter. On top of that, palisades defend camps and the winner takes the loser's food.
- Caves are the oil. They are the best shelter from the meteor and the only place to build the Great Cave. They're rare, but there's always one within a few turns' walk of your start, and finding it is news.
- Your log shows only what your people witness. No omniscient ticker: a rival camp settling across the map stays unknown until someone sees it.
What playtesting changed
The first version was a plain city builder, and my verdict was harsh. Every playtest ended with a one-line note, and each note became a rule in the simulation, with a test, before the next build (the last one feeds M2).
| # | What I said | What changed |
|---|---|---|
| 9 | "found a cave" is an important event | Caves become a survival objective in the goals panel (M2); no turn ends with nobody researching |
| 8 | "when fighting I should take their food" | The winner of a fight takes the loser's food; making camp costs one move; the log shows only what you witness |
| 7 | "units should start from the camp, not another spot" | Scouts and settlers leave from inside the camp; settlers cost people and materials |
| 6 | "check I don't have units that haven't moved" | End Turn warns about idle bands and shows you which one |
| 5 | "the danger is the enemy units" | Danger became units: rival patrols, attack by moving in, terrain defence, zones of control |
| 4 | "M1 should have urgency to camp — food will finish, or water" | Bands carry food that runs out, only rivers are drinkable, a river starts 2–4 hexes away |
| 3 | "settling should require a unit" | Things happen on the map only where your people are |
| 2 | "research once you have a camp" | A wandering band can't study |
| 1 | "seems too simple, no exploration of the map, no armies or dangers… very linear" | Rebuilt the start: you begin as a band in the fog |
Under the hood
The game is a pure C# simulation with no engine code in it. Godot 4 only draws the state and passes the player's orders back to the simulation. That split is what makes AI-assisted development work for a solo project: every rule has a test (165 so far), and a headless runner plays hundreds of games to check the balance numbers.
Here's what the runner says at the end of M1, over hundreds of simulated games:
- Pacing is on target: civilisations found camps at the rate the design asks for.
- The meteor rule matches its design table on every row.
- Shelter matters: camps in caves almost always survive the impact; camps on open plains mostly don't.
- The hook's mechanic fires: every game includes several digs into your own ancestors' ruins.
The rules start from Freeciv mechanics, with Civ as the fallback, and every change from them is written down.
The honest bits
It feels a bit empty. The loop works and nothing annoys, but the land has nothing in it yet. That's expected: the Age of Giants has no giants yet. The giants (dinosaurs) arrive in M2. If it still feels empty after M2, that will be the real warning.
The other half of the answer is choice. Each camp should face a real decision: improve itself and the land around it, raise warriors, or send people off to settle somewhere new.
Preparation doesn't pay enough yet. In the runner, the bots that prepare hardest come through the meteor barely better than the careless ones. Caves almost always save a camp, but they're rare, and the bots barely prepare yet. If getting ready barely changes the outcome, the meteor feels arbitrary and the hook fails. It's the biggest risk on the board, and a later milestone is built around it.
No goals on screen. New players ask, "What is the objective?" The goals panel arrives in M2, and finding and holding a cave will be one of the first.
The road to the hook
- M0 Meteor rule and namesdoneGate: runner matches the design table
- M1 Map, camps, food, player inputdoneGate: 20 turns and not annoying
- M2 Dinosaurs, hunting, goalsnextGate: placement feels like a choice
- More milestonesunder wrapsEach with its own gate
Each milestone ends with a gate that can kill or redirect the project. The real test comes later: outside players finding the ruins of their own past.
What's next
M2 fills the land. It brings four dinosaur species, each with readable signs (tracks, bones near water, claw-marked trees), so placing a camp feels like a choice rather than random damage. Hunting brings food, bones and hides. The goals panel arrives with it. I've given M2 six to eight weeks; if it runs past that, I stop and re-plan.
Next devlog: the first giant on the map, and whether the land still feels empty. If you've built or playtested a 4X of your own, I'm keen to hear what your first one-line note was. Email [email protected].
