Roadmap & Milestones

Canonical GDD content

A phased plan that de-risks the ambitious PvPvE-persistent design by proving the core loop before scaling. Phases gate each other — do not start a later phase until the prior phase's exit criteria are met. Calendar dates are intentionally omitted ([TBD] with the team's capacity); these are ordered milestones, not a schedule.

Phase 0 — Foundation

Goal: clean project base ready to build on. - Source control + LFS, UE .gitignore/.gitattributes. - Remove template variants; establish Content/Lambeer/ structure (../05-Technical/05-Project-Setup-Steps.md). - Dedicated server target compiles (Linux); two-client + server test connects. - Decide GAS vs custom (../05-Technical/01-Tech-Stack-and-Engine.md). - Decide persistence abstraction approach (../04-Multiplayer/01-Network-Architecture.md).

Exit: clean, networked, server-authoritative skeleton with player movement replicating.

Phase 1 — Core vertical slice (the make-or-break)

Goal: a small but complete, fun loop on a dedicated server. - TPS MetaHuman player: locomotion, camera, ADS. - Combat: 1–2 guns + melee, server-authoritative damage, hit feedback (including world blood + mesh/HUD blood per ../01-Gameplay/10-Blood-Visual.md). - Rain wetness: server rain, shelter check, replicated WetAmount on pawn/worn-held meshes (../01-Gameplay/11-Wetness-and-Rain.md). - Firearm attach/remove (optic, muzzle, mag, grip, stock, laser), mag-as-container + chamber/bolt handling, replicated attachment meshes, durability/repair. See ../01-Gameplay/09-Weapon-Customization.md. - Zombies: 2–3 types via StateTree + 1 boss. - Survival full-sim stats + HUD. - Inventory + items + full-loot death drop. - Building: minimal piece set, placement, tiers, raid damage. - Crafting basics + boss component → weapon loop + first-kill permanent buff. - Persistence: player + base state survive server restart. - Squad system. - Small dense map (one region) with World Partition.

Exit: players can join a dedicated server, survive, build a base, hunt the boss, lose loot on death, and the base persists/raids work. This is the proof the genre is viable for the team.

Phase 2 — Depth & social

Goal: turn the slice into a game. - Full zombie roster + more bosses + buff catalog. - Clans + territory/build-privilege depth. - Vehicles (Chaos Vehicles). - Raiding tools + raid-economy balancing; offline-raid policy decided. - Learn-by-doing skill progression: Woodcutting, Scavenging, Crafting, Blacksmithing, Engineering, Mechanics, Construction, Endurance, Strength, Melee, Archery, Firearms, Fishing, Cooking, Medical. Effects stay qualitative until tuning. - Weapon customization depth: per-weapon scope rail position, stock length slider (recoil / ADS / mobility). - Map expansion via World Partition Data Layers / more regions/POIs. - Drop-trade flows; loot-economy tuning with source-owned drop-rate tables for all loot sources. - Performance pass (replication, AI LOD, MetaHuman LOD).

Exit: content-complete enough for closed playtesting at target player counts.

Phase 3 — Live, polish & ship

Goal: ship buy-to-play. - Official server fleet + ops/monitoring; player-hosted server tooling + docs. - Anti-cheat integrated and hardened. - Wipe system + admin tools. - Audio/VFX/UX polish; scalability settings; cosmetics pipeline. - Balancing from playtest data; stability/perf hardening.

Exit: stable, performant, defensible-against-cheating live build.

Cross-cutting (every phase)

• Profile early and often (server CPU, client GPU, bandwidth).
• Keep systems data-driven so content scales without code.
• Keep the GDD updated; resolve [TBD] items as they are decided and mark them [CONFIRMED].

Critical-path risks to watch

The networking/persistence backbone, building-at-scale replication, and AI-horde performance are the long poles. They are addressed in 02-Risk-Register.md.