Production context
GDD/07-Production/02-Risk-Register.mdVerified implementation evidence
- Not verified for this documentation pass.
Relationships
Document contract coverage
Use these required Production Register areas during review. The imported GDD below is canonical; items not stated explicitly remain unresolved.
- ReviewScope & Roles
- ReviewWorkflow
- ReviewStage Entry / Exit Criteria
- ReviewReview & Revision
- ReviewFile Handoff
- ReviewSource Control
- ReviewNaming & Paths
- ReviewDefinition of Done
- ReviewExternal / Freelancer Handoff
Canonical GDD content
The biggest risk to Lambeer is scope vs. team capacity: an indie team building a Rust/DayZ-class persistent PvPvE survival game. This register lists the concrete risks, their impact, and mitigation direction. Revisit each phase.
| # | Risk | Likelihood | Impact | Mitigation |
|---|---|---|---|---|
| R1 | Overall scope too large for indie team | High | Critical | Strict phasing (01-Roadmap-and-Milestones.md); ship a small, dense, complete Phase-1 slice before scaling; cut features, not quality. |
| R2 | Networking/persistence backbone underestimated | High | Critical | Treat as Phase-1 first-class work; build server-authoritative from day one; persistence abstraction early; never retrofit authority. |
| R3 | Building-at-scale replication (hundreds of pieces × many bases) | High | High | ISM/HISM visuals + lightweight per-piece records; dormancy + distance relevancy; decide piece representation early (../01-Gameplay/05-Building-System.md). |
| R4 | AI-horde performance on server (~40–60 players + many zombies) | High | High | AI LOD, active caps, relevancy despawn, StateTree efficiency; evaluate Mass only if needed; profile continuously. |
| R5 | MetaHuman cost at scale (memory/perf) | Medium | High | Aggressive LODs; cap full-detail MetaHumans; zombies are NOT MetaHumans (../06-Art-Audio/01-Art-Direction-and-MetaHuman.md). |
| R6 | Cheating (severe in full-loot/raid games) | High | High | Server authority everywhere; lag-comp without trusting client; integrate an anti-cheat before launch ([TBD] solution). |
| R7 | Live-ops & hosting cost for official servers | Medium | High | Player-hosted option offloads cost; plan capacity/budget; phase official fleet to Phase 3. |
| R8 | Vehicle netcode (Chaos physics + passengers + anti-cheat) | Medium | Medium | Defer to Phase 2; one vehicle first; budget smoothing/replication time. |
| R9 | Boss permanent-buff balance (runaway power, unclear rules) | Medium | Medium | Define participation rule, buff catalog, caps/stacking before building (../03-Enemies/02-Boss-Zombies.md). |
| R10 | Offline-raid frustration vs. hardcore tension | Medium | Medium | Decide offline-raid policy deliberately with target audience in mind (../04-Multiplayer/04-PvP-Raiding-and-Trading.md). |
| R11 | Map scope creep (big empty world) | Medium | High | Build small + dense first; expand via Data Layers only after the loop is fun (../02-World/01-World-Design.md). |
| R12 | GAS learning curve (if adopted) | Medium | Medium | Prototype GAS for stats/buffs early; if team velocity suffers, fall back to a focused custom system. |
| R13 | Persistent save-data growth / corruption | Medium | Medium | Versioned save schema + migrations; wipes to bound growth; backups for official servers. |
| R14 | Content treadmill (survival games need ongoing content) | Medium | Medium | Data-driven everything so new gear/zombies/bosses/regions are content work, not engineering. |
Top open decisions to resolve (gating)
These [TBD] items have outsized impact and should be decided before deep implementation: 1. GAS vs. custom attribute/buff system. 2. Persistence backend technology + abstraction. 3. Anti-cheat solution. 4. Offline-raid policy. 5. Boss buff rules (participation, catalog, caps). 6. Map size and region rollout. 7. Wipe cadence/policy.
Guiding mitigation philosophy
The single most effective risk mitigation is phasing and ruthless scope discipline. A small, polished, fun persistent survival slice that ships beats a sprawling one that never does. Prove the core loop on a dedicated server first; everything else is layering.