The game plan (version 2)

A fresh, focused plan for the game, rebuilt from an interview on 2026-09-13.

This is a high-level outline, not the old 25-file bible. It captures the current direction and the decisions made in this rebuild. The previous plan is preserved untouched in ../plan-version-1/ as the archive — including the eight-volume book series, the deep world lore, and the original design bible. We pull from it as reference; we are not bound by it.

What changed from v1: v1 was the full six-act, multiplayer, seasons + PvP vision — a team-sized project. v2 narrows the first release to one polished act, single-player, so there is something real and finishable to build and ship. The big vision still lives in v1 and remains the long-term destination.

1. The one-line pitch

A 3D Greek-myth action-RPG in the spirit of Diablo (dark, loot-driven) with the presentation and feel of World of Warcraft, set in a real, mythic Greece. You play Myrra, and the world — the twelve gods, the land, the story — is the hook.

Core hook (decided): the Greek myth world itself. Systems can be familiar and proven; what makes this game this game is the setting — real Greece, the twelve gods, the shipped lore. Every design choice should serve making that world feel alive and worth being in.

2. Scope — the first release

Target: one full act (Act 1), polished, single-player, start to finish.

This is deliberately the smallest thing that is a real game rather than a demo: a complete, satisfying loop with a beginning and an end, that you could sell or put in front of players. Everything beyond Act 1 is a later release, not this one.

⚠ Honest scope note — read this

WoW-style 3D + three fully-built classes + skill trees + full loot + difficulty tiers is ambitious for a solo first act. The "short, 1–2 hour" length is what makes it survivable: a small amount of world to build, so your time goes into the three classes, the loot systems, and polish instead of into square kilometres of content. Even so, this is a multi-year solo project. The roadmap below front- loads a prototype gate so you prove the fun before committing to all three classes and the full act.

3. The game, section by section

3.1 Setting & story (Act 1)

• Where: the Rarian Plain / Eleusis — Myrra's home ground (fixed by canon). The game opens where she is from. Region I of the wider world.
• Threat: a local, human-scale crisis — a blight, a failed harvest, a cult at the Mysteries, something grounded — that slowly reveals a divine cause underneath. This is the "real quests" principle: the fantasy explains why things go wrong; the quest is a real human problem (a debt, a missing person, a broken rite).
• Protagonist: Myrra, a woman of the Rarian Plain. She never speaks; everything about her is conveyed by what others say to her. Her class is the player's choice; her name, gender, and origin are fixed.
  • ⚠ Script rule from canon: nobody on the Rarian Plain is her family — no NPC line may name her parents. This is set up for a later reveal.
• Tone visual arc: consider opening sunlit and pastoral, darkening as the divine threat grows — using the Diablo-dark palette as the crisis deepens.

→ Fully fleshed out in act-1-story.md (2026-09-13): the Eleusis chapter as the shippable slice — hub + two zones + boss (the Hungering Steward), a five-chapter quest chain grounded in real problems, with the divine cause only ever glimpsed. Delphi/Python deferred to the next content.

3.2 Classes

Three, forming the classic trinity, all cleanly Greek:

Each gets a skill tree with meaningful build choices. Replay comes partly from playing the other two classes, so they must feel genuinely distinct — not reskins.

→ Fully designed node-by-node in skill-trees.md (2026-09-13), built on the canon class fiction: Hoplite (Valor; Phalanx/Doru/Andreia), Huntress (Favor; Toxon/Agra/Thiasos — canon's richer take on "Archer"), Oracle (Favor; Helios/Manteia/Loimos). Two ultimates per class; tuned for tab-target, single-player, and the Act 1 slice.

3.3 Combat

• Tab-target, WoW-style: select a target, abilities resolve on it, a global cooldown paces a rotation of cooldown-gated skills.
• Chosen for solo-feasibility: it puts your effort into class identity and content rather than the hit-detection/animation-feel problems that make action combat so hard to ship well alone.
• Fallback path: if Act 1 feels too static in testing, layer in a manual dodge and a few aimed abilities. That's an addition, not a rewrite.

3.4 The gods — light version for Act 1

• Gods are world flavor in Act 1 (temples, story, quests, the Mysteries), with one light mechanical touch: shrines that grant a temporary blessing/boon.
• The full divine favour system (patron gods, earned favour, permanent boons) is deferred to later acts. This keeps the hook present without a big new system in the first release.

3.5 Loot & itemization (the Diablo half)

• Full ARPG loot loop: rarity tiers (normal / magic / rare), random affixes rolled on drops, and a few uniques and sets to chase.
• This is the dopamine engine of the genre and central to replay — worth the work even in a short act.
• Difficulty tiers feed it: better loot at higher difficulty is the cheapest content multiplier in the genre.

3.6 Progression & replay

• Leveling + per-class skill trees. Build choices matter; respec via Lethe Water. First-release level cap: 30 (1 skill point/level; attributes auto-scale) — Mortal clear ≈ L15 (one ultimate), Divine clear ≈ L30. Content-gated below the canon ~100 full-game cap. See skill-trees.md.
• Difficulty tiers (Normal / Nightmare / Hell-style): replay the same act at higher difficulty for better rewards.
• Class variety as the second replay axis — three distinct classes to master.

3.7 Art & audio direction

• Diablo darkness, Greek. Moody, shadowed, gritty — Diablo 2's coloring applied to sun-bleached ruins and temples turning ominous.
• Solo production reality: lean on a stylized-but-atmospheric look that hides asset limitations, rather than photoreal (which punishes solo art budgets).
• Audio: an existing track list lives in v1; reuse/adapt as reference.

3.8 Naming rule (added 2026-09-13)

Every boss, miniboss, and NPC must have a Greek name. No generic labels ("the Steward"), no invented pseudo-Greek — real Ancient Greek names or authentic epithets, consistent with canon's rule that the world's language is Ancient Greek (one transliteration scheme, applied everywhere — never mix Kirke with Circe).

• Applies to: all bosses, minibosses, and every speaking or service NPC.
• How it interacts with canon: the books deliberately give some figures role-titles, not names (the Elder, the Argonaut, the general, the Pythia). Those role-titles stay as canon in the books; in the game each such figure also gets a Greek proper name that NPCs use aloud (the game already does this for Myrra, whom the books never named). Role-title in prose, Greek name in play.
• Monsters/creatures already carry mythic Greek names in the bestiary (Keres, Dryad Blights, Skeletal Hoplites) — keep those. This rule is about named characters and bosses, which must read as Greek people/beings, never as a function.
• A boss that is a mechanism (e.g. Act 1's Intake) still takes a Greek name or epithet, not an English noun.

4. Production plan (the solo-dev reality)

4.1 Art asset strategy — the biggest bottleneck

A WoW-style 3D game needs animated characters, and that is the hard part solo. The strategy is a hybrid pipeline:

Bottom line on "can ChatGPT make them?": yes for 2D (concept, textures, UI, promo, book), partly for static 3D props (needs cleanup), no for the rigged, animated characters and monsters — buy those. Budget for a handful of asset-store packs; that spend is what makes a solo 3D ARPG possible at all.

The full art list, AI prompts and tracker (demo, first release, marketing) live in the shared drive (set up 2026-10-15): start at its README.md.

4.2 Engine & tech

• Unity (C#). Reuse the existing PHP→C#/Unity primer in ../plan-version-1/csharp-and-unity-for-a-php-developer.md.
• Single-player only for this release → no netcode. (v1's server-authoritative / FishNet requirement is dropped for now; revisit only if/when co-op returns.)
• Project location: keep the Unity project out of OneDrive — OneDrive fights Unity's Library/ and has already produced conflict copies. (Carried over from v1.)
• Keep narrative strings separate from systems strings from day one (quest text, dialogue, item flavor vs menus/buttons/errors) — cheap now, enables localization and the Greek-narrative toggle later.

4.3 Roadmap — phases with gates

→ The full step-by-step build plan is in development-plan.md (2026-09-13): every phase broken into numbered, ordered steps, the cross-cutting systems, the working cadence, a realistic timeline, and a risk register. The summary below is the shape; that document is the execution.

Each phase must pass its gate before the next begins.

• Phase 0 — Prove the loop (grey-box, no Greek theme, no art). Move + attack, one enemy that dies, one item that drops, is picked up and equipped so a stat visibly changes, and one affix that rolls randomly. Gate: Is move → kill → loot → equip actually fun? Everything waits on yes.

• Phase 1 — One class, vertical slice. Pick one class (suggest Hoplite — simplest to feel good). Build its core skill tree, tab-target combat, the full loot loop, and one small themed zone of Eleusis. Gate: a 15–20 min slice that feels like the real game.

• Phase 2 — The full Act 1 for one class. All Act 1 zones, the grounded-crisis story, the boss, shrines/blessings, difficulty tier 1. Gate: playable start-to-finish, 1–2 hours, fun throughout.

• Phase 3 — The other two classes. Archer and Oracle: their trees, abilities, animations, balance pass. Gate: all three classes viable and distinct through Act 1.

• Phase 4 — Polish, difficulty tiers, release. Nightmare/Hell tiers, uniques/sets, audio, UI polish, options, save system. The hype book finalized. Gate: shippable.

4.4 Books & audience

• A standalone "hype" companion book — written to excite readers and pull them into playing the game. Separate from the eight-volume canon series (which stays in v1 as deep reference and shared canon).
• Shared canon between book and game; the first game release stands alone and requires no reading.
• Bundling / physical + digital sales: decide later, closer to release.

5. Open questions (to resolve as we go)

• Act 1's specific story beats: zones, quest chain, the boss, the divine reveal. Done — see act-1-story.md.
• Skill-tree shape: how many nodes/tiers per class, respec rules.
• Which asset-store packs actually fit the Greek-dark look (needs a survey).
• Save system & difficulty structure specifics.
• Distribution for the first release (Steam? demo-first? free vs paid?).
• The hype book's story and how tightly it ties to Act 1.

6. External / non-code (carried from v1, still open)

• embersofolympus.com registered (2026-09-09). Still to claim handles (X, Instagram, TikTok reported free; YouTube unconfirmed).
• Attorney trademark clearance — Class 9 (software) + Class 41 (entertainment).

7. Immediate next steps

1. Install Unity Hub + Unity, create the project outside OneDrive.
2. Skim ../plan-version-1/csharp-and-unity-for-a-php-developer.md.
3. Build Phase 0 — the grey-box loop — and answer the only question that matters first: is it fun?
4. In parallel, we can flesh out Act 1's story beats and do the asset-pack survey so Phase 1 has art to use.

> **Status:** Plan v2 drafted 2026-09-13 from interview. Next design work: expand
> §3.1 (Act 1 story) and §4.1 (asset survey). Nothing is built yet.